Conceptual Models I
From CS160 User Interfaces Fa06
Lecture on Sep 11, 2006
Readings
- The Psychopathology of Everyday Things. The Design of Everyday Things. Chap 1. Norman.
- Tangible multimodal interfaces for safety-critical applications. CACM. 2004. Cohen & McGee.
Ramy Ghabrial - Sep 06, 2006 06:40:18 pm
POET: The chapter expounds three features of well designed UIs: Visibility (making functions visible to the user), natural mapping (accessing these functions with intuitive operations) and feedback (confirming the success or failure of a function, or giving additional information about it). Taken together, these three features help provide users with an image of the system that: a) matches what the system actually does or can do, and b) does not confuse, mislead or intimidate them.
Tangible multimodal interfaces for safety-critical applications: The main new topic in this reading is multimodal interfaces. TUIs can be seen as a generalization of the Anoto concept of a UI you can touch and physically play around with. MMUIs however have not really been discussed until now, and the "four- to nine-fold faster" statistic is interesting if accurate. Applying this idea to a project would seem to involve speech recognition, as described in all the examples. Are there other ways to incorporate MMUIs into an Anoto project?
Tabassum Khan - Sep 07, 2006 11:02:20 am
The Psychopathology of Everyday Things - Norman: This reading is a very good illustration of all the dos and donts of designing. At all times, designers should be well aware of the fact that they are designing a product for others to use not for themselves and therefore the process of designing a product must be user-centered. The basic things that a designer should provide are : Visibility, Intuitive Mappings, Affordances and Feedback.
Here is an example of an everyday thing that has become one of the greatest annoyances of my life - a touch-free faucet. I consider it a "not-so-good-design" because 1. No Visibility : You cannot know how to turn it on unless you put your hands under it. 2. No localized temperature control 3. It stops dispensing water before you are done washing your hands. You have to take your hands off the sink and repeat.
Tangible multimodal interfaces for safety-critical applications - Cohen & McGee: The reading was very interesting with the introduction of multinodal interfaces. The hybrid between the TUI and MMUI sounds like a great plan. However, I am a little skeptical about the fucntionalities that NISChart claims to provide. Speech recognition is not a trivial task particularly when there are people speaking with varied speed, different accents, dialect, sociolinguistic background, and variability in vocal tract size and shape.
Andrew Hao - Sep 08, 2006 04:37:42 pm
The Psychology of Everyday Things explained the incredible challenges that the designer faces, working under the constraints and pressures of a myriad of requirements (aesthetic, budgetary, functional), and gives me greater appreciation of the beauty of everyday designs (e.g. the door handle example: horizontal to push, vertical to pull). An interesting point that was brought up was the schema mismatch of the fridge cooler controls. A mismatch in schema - although it seemed natural to present the user with the easier schema, made it tougher to use in the long run. It is another tradeoff that the designer must make - how much should we show the user under the hood?
Multimodal Interfaces are explained by Cohen and McGee to be real-world/virtual-world interfaces that better bridge the gap better than pure virtual interfaces. These sorts of interfaces have special applicability to safety-critical arenas where professionals are less likely to trust new, radically different technologies. By extension, it seems that TMMIs could be advantageous in every other arena as well (those that are commercial, those that are not as safety-critical).
Bowen Li - Sep 08, 2006 08:55:32 pm
Everyday Things: I think the point that people make a conceptual model in their heads first is a very important one. I know that's exactly what I do when I look at something. Also, the point that a system's working parts should be visible and obvious is a good one. In today's world of electronics, we don't often think about objects having "visible" functional parts, but they obviously do. It's easy to overlook that small aspect.
I was surprised that he said it usually takes 5-6 iterations for a product to succeed. I wonder if products dying is from the company's lack of persistence, or from the consumer not willing to try a failed product more than once. It's a wonder that so many people are still willing to invent new things.
Safety-Critical Applications: I think an important point the article addresses is that electronics WILL break. And because they do, new technology should not replace the human ability to do work. It's the same concept as to why we still learn arithmatic in an age of calculators. The tools should assist but not replace.
However, in the fields of military and medical applications, I am skeptical as to the benefit to be derived from electronic implementations. These are both fields that have random, rapidly changing situations that require fast thinking and innovative responses. I feel that by locking down the implementation (through a digital pen or post-its, etc) then that may limit the number of solutions possible. Even if the pen still works in the physical world, there is an assumption made that the units (platoons, tanks, etc) can be accurately be represented by a post it, and that whatever strategy formed will "work" on the canvas provided. It does not allow for uses of outside objects, etc. Since it is a safety-critical application, I wonder if that is a prudent choice.
Natalie Nguyen - Sep 08, 2006 11:01:18 pm
Tangible Multimodal Interfaces: The ideas presented in this article are interesting, but I feel like the technology isn't really at the point where it needs to be yet. OCR and speech recognition working together to disambiguate sounds great in theory; however, the handwriting of doctors has long been notorious for general illegibility to "outsiders" that the idea of trying to build an OCR system for that specific task is frightening, let alone trying to tie it with speech recognition, which is in and of itself a complex system to design and deliver. The article suggests that the data input would not be entirely automated (i.e. accuracy of the data sounds like it would be verified), but I can't imagine the interpretted data to be more often useful than not.
POET: An interesting and entertaining read preaching relatively simple concepts that aren't always easy to apply; however, I found the author's example dealing with the loudspeakers to be a little contrived. Yes, the control lacks a natural relationship so that it is not immediately intuitive to the user. However, fiddling with the knob gives immediate feedback (assuming there is music playing), so it is painless enough to figure it out. Is it really so bad on the designers' part to ask the users to spend even a little effort learning what they bought? In the case of the phone system, the effort required is nontrivial; however, figuring out the speakers seems like a very trivial task.
Hiroki Terashima - Sep 08, 2006 11:23:28 pm
POET paper advocates testing new interface before purchasing it or deploying it, because the flaws of the interface aren’t obvious to a designer. The telephone with the R could be simple in design, but unintuitive for the user because the functionality greatly outnumbers the available buttons to press. I thought it was unfortunate for the author’s friend, who got locked in between the beautifully designed doors; I also think that the system should be easy to use before being beautifully designed.
After reading Safety-Critical Applications, I too share a similar concern that Bowen mentioned, and wonder if the people using these TMI's will really be able to accept the new medium despite the wide variety of affordances. All of the TMI’s that the paper mentions are not commonly in use today, it seems. With the safety-critical application, I imagine that testing the new medium requires lots of care and thought, unlike testing a new Calmail interface, for example. I found it interesting that Moore correctly predicted the society’s behavior on accepting these new technologies.
Heung Tai - Sep 09, 2006 05:02:38 am
Tangible multimodal interfaces for safety-critical applications: This paper points out an important issue: mismatch in user interface. When a person is used to something, it takes effort for that person to change because he or she already built confidence on that user interface. The idea of TMM is good but one of the modal system has to be same as or close to that person's original interface to reduce resistance of change. If a person experiences failure from a modal in the past, the resistance of reaccepting the modal will be great, so every technology has to be fully tested before giving it out to user.
The Psychopathology of Everyday Things: The article is very interesting because it demonstrates some stupid design that is done by professional designer. It emphasizes on visibility which I think is very crucial for knowing how to operate on an application. When we look at something, our brain immediately recall similar configuration, that helps us to predict what should be done. So, a good application need to have good visibility, and it should someone consistent with what people have already got used to.
Yimin Yao - Sep 09, 2006 05:28:20 pm
POET: I think "natural mapping" is a very excellent criteria for a well designed UI. As the chapter has suggested, there are so many small UI devices in our daily lives, making it impossible for people to learn and remember all the operations from scratch; it is designer's responsibility to convey enough intuitive information through the interface in order to allow users' accurately determine the device's functionality based on their general knowledge and experience. I enjoyed the reading for many of the bad UI designs mentioned (door knobs or bathroom faucets) were common occurances that I have struggled with personally. They are indeed everywhere: my stove control knobs have no indicator for power vs. direction of turning, and there are no feedbacks (such as light signals) when the stove is on, thus I often find the apt empty with the stove on medium power (when it's not hot red in color). Luckily no one has been hurt yet. Also, my refrigerator has two symmetrical handle pivot point on each side of the door, while it can be only opened on one side; so many of my visitors always had trouble on their first try.
Safety-Critical Applications: The paper recognizes the resistance from professionals due to the mismatch between the technolody / interface with their usual working environment and procedure, and suggests that system designers should adapt to key aspects of the users' practice and employ current physical objects/tools. The ideal is correct, however, as a few students have mentioned above, many of the devices mentioned employ OCR and speech recognition, which are both still very limited in their performance. The significant possibility of errors alone can induce enough resistance to the product. And I am skeptical too about how much can such advices really improve the efficiency of the tasks.
Alex Wallisch - Sep 09, 2006 06:52:44 pm
Safety Critical Applications
TMM's seem to be an attempt to harness both the ease of use of physical interfaces with the efficiency of digital backends. Unfortunately, it appears that in an attempt to get the best of both worlds, you wind up with neither. Consider in particular the Rasa system described in the reading. The interface is, on one hand, largely tangible. However, in an attempt to allow a computer to process input, you lose several nice features. For example, one of the nice features of the paper maps is that they allow for collaboration by having several people use the system at once; however, with the Rasa system you have to give up this feature. On the other side, the advertised purpose of Rasa is to digitize battle planning and allow for collaboration across multiple systems. For this to work, each system must trust the output of every other system. However, with Rasa, a discrepancy between the digital data and the physical representation of it generally results in rolling back to what the physical representation says. If you are going to rely on the physical more than the digital, then it makes almost as much sense to not bother with the digital at all.
Psychopathology of Everyday Things
I agree with the author's assessment of the importance of feedback. When I read this article, the first thing that came to mind is a game I have been playing recently (the QT port of nethack, if anybody is interested). The game contains a number of options that can be set to true or false, and the interface to set these looks like this:
The problem is that the printed "true" or "false" values do not reflect any changes the user has made; in the picture, I have actually set the "ascii_map" option to true (hence it is highlighted in green), yet the text next to it claims that it is still false. This design seems like it could be improved by showing the user feedback of exactly what he has done.
Tak Wong - Sep 09, 2006 07:38:04 pm
Psychopathology: I believe that a lot of the problems in the designs the author mentioned can be eliminated easily by observing the users for a day or two. Once the designers spot the problems, they can at least attempt the solve the easy issues such as the glass doors or the projector screen. Otherwise, they have many ideas about what they should improve in the next phase. I agree that the natural mapping is important. Noone hardly reads the manual or instructions, even if they are short and easy to understand. We are tempted to try it out a few times first, fail, then finally left with no choice but the manual.
Safety Critical App: Even if the information is not safety critical, many people still refuse to work without paper because they don't want to learn how to use an unfamiliar application. They are afraid of what will happen if their data is not entered correctly or the database crashes. But we as the CS people know that there are backups stored in secured places if something bad happens. We can communicate with other parties involved more efficiently with electronic data. The hard part is making these people believe that electronic data are safe and make it easy to enter them. In the examples in the article, the redunancy of having records on paper secure the users.
Jason Shangkuan - Sep 09, 2006 07:34:44 pm
POET: This article is interesting because the author points out how we interface with so many devices in our daily lives, yet every single one has its own affordances. The challenge is to clearly understand the affordance of each item, which lies not only by the user to recall prior knowledge and experieence but also in the one who designs the object. Everyday objects such as doors can be confused in the way they operate. As the author mentioned the example of the row of doors, it was a clean design that looked simple yet was difficult to figure out. In my own experience, I have a stove that I did not know how to operate until I made a mistake. The oven has a light labelled "oven", which I assumed once the oven was at proper temperature, the light would turn on. However, when I used the oven and turned the knob to the correct temperature, the light came on, but I did not know if this meant it was working or how I would know it was at temperature. I put in my food and then after awhile I saw the light was off, which meant to me that the oven was off. This confused me until I realized when the light turned off, the oven was at temperature.
Tangible Multimodal Interfaces: The concept of TUIs and MMUIs and connecting the two ideas has something I have always thought of and never realized simple implementations that exist especially using the Anoto pen. I think it would be very interesting to incorporate a network of sensors, which have temperature sensing and positional reading abilities, with the direct interaction of human voice and commands. The network would sense body temperature and where they are in the room and adjust the environment to a suitable level, however, it can be overridden or adjusted with the user says they want it differently. I think that the bridge of TUIs and MMUIs would improve many aspects of our lives, but the correct implementation is difficult to achieve.
Jonathan Yen - Sep 09, 2006 08:09:08 pm
POET: What I find most intriguing about this reading is that a lot of the comments that Norman has regarding good and bad user interfaces in daily life are completely anecdotal. There is no formula for determining what exactly constitutes good or bad design, as it is all qualitative. I think this reflects much of the difficulty in trying to improve user interfaces, as there are many factors to take into consideration, and ease of use may differ from person to person.
Tangible Multimodal Interfaces for Safety-Critical Applications: This article demonstrates many advantages to TMMs. While multimodality seems to have the benefit of collating multiple data sources together into one coherent unit, I think the bigger advantage is the tangibility of such systems. Having a hard copy and a soft copy of information that is always kept in synchronization makes it easier to organize one's data as well as preserving it. As far as I can tell, it would take some time for the technology to be able to suit professionals' needs.
Scott Friedheim - Sep 09, 2006 09:18:54 pm
Tangible multimodal interfaces for safety-critical applications The most important idea I took away from this article is the emphasis on designing for the user. This same idea is reflected in the ButterflyNet article in which the whole design was build around the researchers being comfortable and natural in the use of the equipment. If the design does is not build around the normal methods that a job is done then the design will be foreign and naturally frowned upon. Especially when it comes to safety critical operations in which there must be redundancy, and a strong comfort level in the technology in use and in dependability.
The Psychopathology of Everyday Things. This article was so funny. I relate to everything he says; I too have problems with doors. This usually happens at a Carls Jr...but when you go to the restroom there is an outer door you must first go through and then another door to enter the restroom, similar to the door problem in the paper. However, after going through the outer door you enter what looks like a 4x4 closet with 4 identical doors on every wall. Choosing which door is the right one for the restroom is not that bad but when leaving the restroom it's not easy to remember which door lead out so I'm always stuck checking every door. I really thought that the principle of mapping was a good design idea to follow. Especially nowadays when everything is being miniaturized and more and more multifunction buttons are being created; I personally think it starts to add complication to the product esp. when you now need to navigate 20 sub menu's in from the main menu to change some setting.
Maksim Lirov - Sep 09, 2006 11:24:29 pm
POET: This reading presented examples of products with poor design of the user interfaces, with explanations of why they are bad designs. Although some design interfaces are indeed more intuitive than others, I believe that whether any particular design is intuitive to the user depends on the user's past experiences and knowledge. What may not make sense to one person may make lots of sense for another. It is true that as devices become more advanced and complex, the user interfaces will also become more complex to use. The British Telecom phone with the mysterious R button looks simple to use, but it would require reading of the instruction manual to learn when to use the R button. I think that the reason we see so many "poor designs" is due to the fight to be the first to bring an item to market limits the amount of user testing that needs to be done.
Tangible Multimodal Interfaces: I agree with the authors that TMMs have the potential to provide more robust operation because the data is stored in physical and digital form. The main point that any such system must address is that the digital form should complement the functionality of the physical object (such as paper and pen) instead of trying to have the physical object complement the functionality of the digital form. The description of NISChart in this reading sounds promising because it still uses the paper and pen - which the clinicians are already comfortable using - as the method of interacting with the system. The system complements the physical world by also acquiring pen strokes in digital form for backup and easier sharing. The system doesn't attempt to introduce a totally new way of filling out forms but rather complements the task of filling out forms (i.e. storing pen strokes in digital form as backup).
Tony Yu Tung Lai - Sep 10, 2006 01:13:45 am
Norman: This chapter pointed out a lot of important topics that all designers should consider when they are designing a product. I think the most important point that I got from this chapter is the distinction between a beautiful design and a good design. As Norman mentioned, there are a lot of fancy designs that makes no intuitive sense when it comes to using it. It is important to make the product look good, but the priority must always be usability.
Cohen & McGee: The notion of multimodal user interface is very interesting. One thing that might be a turn off to users is the fact that they have to do two inputs almost stimultaniously (say, writing and speaking). I think it is okay for short term users, but it might get a bit annoying if someone has to use this long term. However, it is nonetheless a very interesting and workable idea.
Michael Moeng - Sep 08, 2006 09:48:55 pm
The Psychology of Everyday Things
I somewhat disagreed with a few of the points in this article. While it is very important to design things so they are easy and intuitive to use, not everything is made for an idiot to use. Just about everyone can drive a car, so cars need to be simple and intuitive, especially since they are already so dangerous. But a control panel on a ship might be constrained by space+size, so it might not be able to have labels on all its lights and whatnot. Just because a light was "badly placed" doesn't mean that an engineer whose job it was to know about the control panel should have a good excuse to not see the light. Between the article's standpoint and "a poor worker blames his tools" I think the truth is somewhere in between.
Tangible multimodal interfaces for safety-critical applications:
One big issue I see in the integration of the paper and digital worlds is that, while many of the integration methods provide for failure on the part of the electronic equipment, many (those using the Anoto pens among them) do not back up the data completely. For example, if a pen crashes, the researcher still has the paper copy, but there is no way to recover the digital data (except with a separate scanner). I liked the idea for the post its on the map (Rasa) where in the event of a system failure, the system could fix itself.
Rayhan Lal - Sep 10, 2006 06:36:41 am
The Psychopathology of Everyday Things: The article provides a very useful critique of everyday items. The design principles of good conceptual models and visibility are relevant when functionality is limited, but I am uncertain about if these techniques can scale well. For example, hold and redial buttons can all be put on a phone, but imagine designing something which could have many more functionalities – a keyboard. You have to decide if the minimal key set is sufficient to provide access to the hundreds of applications/functionalities the computer may have.
Tangible multimodal interfaces for safety-critical applications: If the Anoto system is to live up to the hype we saw in earlier articles, it seems that these are the applications to make it happen. The uses discussed necessitate pen and paper but would be aided by the features afforded by computers. Truth be told, the NISChart was the first use of the Anoto system that came to me when I heard about it (sans the speech input). However, the article was written in 2004 and a quick search revealed no recent developments. Natural Interaction Systems[1] seems more focused on their military projects at the moment.
Eric Yoon - Sep 10, 2006 12:57:12 pm
Tangible Multimodal Interfaces. I certainly agree with a core premise of the article, that it is often futile to force people to adopt unfamiliar user interfaces and tools, particularly in mission-critical applcations, such as in medicine or the military. In fact, the more critical or life-risking the transaction in question is, I think the more hesitant practitioners are to touch anything but that which is tried and true. One key feature of pen and paper is not only its portability or high-resolution, but also its enormous reliability -- when you write something, you know it'll be there today and tomorrow and many days therafter. With computers, ALL of us have had situations where that hasn't proven true. The article raises a good point in noting what the system did for military planners when it failed (it was able to go down and then reclaim lost ground when it came back up again.) For many groups, this feature will be of extreme importance, by giving them a feeling of failsafe reliability that they need.
POET. The three concepts outlined in the article -- visibility, mapping, feedback -- are good ones and help to frame the nature of the challenge faced by the UI Designers. I agree with Michael Moeng's comment, though, such statements are minimally helpful in many situations -- after all, some devices just do not owe themselves to great visibility, a great mapping between action and result, or a great feedback experience. This is particularly true in environments where you are running through data on a computer in some fashion -- whether it be on a PDA, cellphone or computer. Such programs invariably involve a lot of different options, all of which are triggered (flip down menu? buttons?) in rather generic, monotonous and indistinguishable ways. But in environments where quick reactions are required (cars, video games) and functionalities are few, visibility, mapping and feedback effects should be sought and become extremely important.
Patti Bao - Sep 10, 2006 1:14:49 pm
POET: I enjoyed reading this book, and after I read it the first time, I found myself pointing out bad design after bad design in everyday things I encountered. But when I discussed these designs with other people, some invariably disagreed. What I thought was poorly designed, someone else might have thought was well-designed, and I think Norman may have glossed over the fact that user opinions can differ so much. I like the idea of natural mapping, and I agree that it makes sense to make good use of affordances, but as Norman said, doing so depends not only on physical analogies and actual properties, but on cultural standards and perceived properties of an object too. Does something that works well in one region fail completely in another? I wonder if there is such a thing as good universal design, or perhaps I am just being picky. At any rate, the designer's job is definitely a hard one, and Norman conveys that well in this first chapter.
TMMs: I think the authors make a valid point in favor of creating digital systems that continue to make use of existing physical objects. However, I am not sure whether catering to user habits is enough to justify implementing TMMs for safety-critical systems. No matter how well-designed a system is, if it is digital, it has all the flaws of being digital - security, reliability, etc. Like Rayhan, I also considered something like NISChart for the Anoto pen before reading this paper, but I wasn't convinced that doctors would be willing to have their medical records computerized at all. Before TMMs can ever cross Moore's chasm, more research needs to be done on why digital systems have failed in the past - if the only reason is because users are accustomed to using certain physical objects in their practices, then perhaps TMMs have a chance.
David Hoffman - Sep 10, 2006 01:27:26 pm
Donald Norman has some really interesting insights about the conflicting demands on a designer. It seems that the big thing that he really wants designers to avoid is trying to satisfy everyone without actually talking to the people who will actually use the device in the end. Thus desiging million dollar phone systems without trying a prototype system out in an environment of naive users should be avoided. Norman also makes some really good aguments about the importance of mapping. It seems like a really good idea to associate spatial locations of controls and what they control. The example that I thought about when he was describing how easy the car controls were the few counter-examples in cars that I came across and how I resented that they did these modifications. One of the worst was having the controls for setting the rear seats flat located in the trunk. I really agree with him that if it needs a label to make it clear what it does, it is probably a poor design.
The Cohen and McGee article makes some good points about utilizing people's skills with proven technology. I think the Palm is a good example of a system that really alienated users by shunning their traditional methods of data entry. They could have used a keyboard or tried to learn people's writing sytlyes but instead it asked people to learn their grafitii style which is unrealistic to expect masses to learn a skill which will in all liklihood be made obsolete by a better design. I also like his insight in that computing technolgy, once damaged becomes a rock, wheras paper can still be used. Thus, for critical applications, its a good idea to have a certain amount of failsafe features built in. Paper is good because it is much more robust than most computer systems.
Udam Saini - Sep 10, 2006 01:48:16 pm
Poet: I thought that this excerpt from his book gave quite a bit of insight into what a good and bad design consists of. It is interesting to note that I had always thought it was me that 'certain doors (and always the same doors)', I would struggle to remember whether to push or pull. I realized, that this was due to the lack of visual cues as to the appropriate function of the door, and thus, was not really my fault. It was the designer's fault. A good design needs to have visual cues, good feedback, easy to remember interface, and one to one control to feature functionality. There are obviously other things that make up a good design, but those are very important to almost anything that a human designs to help the user properly use the device.
TMM: It was interesting to note that the authors singled out safety-critical applications for TMM's. I believe this would be useful as well in non safety-critical applications. It would be good to use an interface that the user is already comfortable with to create and use something digitally. There are many people today that are intimidated with technology. However, using a digital desk can help people who are more comfortable with the pen and paper to transfer a digital document over to someone else in the office. The multimodal interfaces are necessary in safety critical applications to ensure correctness. Tangible interfaces, however, would definitely be a good idea in other applications.
Patrick Rodriguez - Sep 10, 2006 02:09:31 pm
POET: I thought this was a very interesting read as well. It clearly conveyed the importance of visibility, mapping, and feedback when it comes to design. I have some doubts about his "Paradox of Technology" however. While it is true that "new features" sometimes means "harder to use," the important part is that we actually get these new features at all. Those that wanted or needed these new features will try and figure out how to use them. Those that are confused can just wait for the next version. And the next versions will probably resolve those problems. For example, the author mentioned adding GUIs to expose the additional functionality that modern phones offer. Most cell phones and office phones now have LCDs and are pretty easy to use. I'd much rather have cool, but hard to use, features now rather than waiting for perfect design. If we took the latter route, we'd still be stuck in the stone age.
Tangible multimodal interfaces for safety-critical applications: Now we have two main reasons for a mixed interface (paper and digital) vs. a purely digital interface: familiarity with paper and safety concerns. I personally can't get excited since neither concern applies to me, but I guess I can understand the reasoning behind this family of UI. I think this sheds some light on the failure of Anoto technology in the mass-market. The technology is only really useful in niche applications like ButterflyNet or the military and medical applications described in this article. User interfaces specific to niche areas are, of course, still very important, but Anoto's initial mass-market strategy was very flawed in retrospect.
Joe Hart - Sep 10, 2006 03:00:41 pm
POET: While I agree with the author about good design following the affordences of an object, he completely misses the aesthetics of good design. Good design can be graded on its ease of use AND on its beauty. While some things are awkward to use people buy them because they are great to look at or give them status or whatever. This article does not address how to balance form with function. However, I do agree with some level keeping things simple. An alarm clock that is also a light and a TV and ... ispossibly getting out of the designers ability to clearly present it. Simple, everyday objects should follow a simple design (wall clock), while objects that people come to expect as complex (computer) should try to pack in features and uses for the customer.
TMI For Safety-Critical Applications: This article was very military centric and I had trouble seeing the benefits of battlefield tracking also being captured by a digitalup link . But, I fully believe in the use of some sort of digital auditing for medical applications. The medical industry can benefit from the instant statistical information and processing. If disease was starting to break out in a community a system that allowed for instant dataretrieval from the front line might allow for a better response from the medical staff, county or state. The auditing of doctor practices along with the multiple backups of realtime data ensures the legal industry will have less problems with malpractice suits along with fraudulent claims from individuals. All of this and still in a paper-pen medium that is comfortable and natural for its users and management. Knowing a few people in the medical field progress toward efficiency is greatly needed.
Roland Carlos - Sep 10, 2006 04:06:02 pm
The Psychopathology of Everyday Things: I found it to be a good read. It was a great insight onto many "simple" things we take for granted and basically lets us know that things are not as simple as they seem. I feel like this reading connects a lot to our previous reading on The Task-Centered Design Process [2] which talked about how we expect certain things (like doors, facuets, etc.) to work a certain way. So if we encounter something that does not work to our expectations, we are usually left wondering what is going on (as in the case with the glass doors that didn't seem to open).
Another point of interest was how bad designs are sometimes bad in the sense that they create false causalities. In this case, it seems like a bad design just creates more problems than just being difficult to use. It feels like a self-propogating error that will never end. And just think, the problems that are created because of false causalities could have been easily resolved if some more work was done in the design planning process.
The reading was also good in that it showed that not just overly complex designs fail (for obvious reasons), but also overly simple ones as well. Take the example of the projector that only had one button to use. It seems simple. But without any prior knowledge (again, going against our "expectations" on how a project should work), we wouldn't be able how to figure out how to move the slides forward (a gentle push) or backwards (a hard push).
Tangible multimodal interfaces for safety-critical applications: Similar to the ButterflyNet system, this is another project that is based around a current need, namely that of digitizing our safety-critical operations (the example used in this paper being the military). Like I said in my comments for the ButterflyNet paper, it's a good idea to establish your project around a current and visible need. However, dealing with safety-critical operations is a lot more difficult than just taking notes and pictures. The paper seems to answer some of the safety questions (for example, backing up live reports when the computer crashes), but I don't feel convinced on the security aspect of things.
Nonetheless, it smartly recognizes that we can't do a complete transition from analog to digital communication in one quick move. This project wants to be part of the slow transition and knows how it will play its part.
Bryant Yu - Sep 10, 2006 04:04:40 pm
POET: It's really interesting and helpful to have all these "stories" that Norman provides. I like his example with all the functionality of cars yet a car is so much easier to use. But to be fair not only do cars have features that are located near the devices they affect(ie radio buttons are near the radio and window controls are near the windows.), but also ppl have been learning how to use a car since they are 4. We see parents, friends,and movie stars operating cars and using their features whereas there is no super hero using the redial conference calling abiliites of the phone.
TMM: I really liked reading of the progression of interfacing. With the ubiquity of computers older models for interfacing need to be upgraded or thrown out. As technology evolves, it's less of the old monitor mouse, and more of what people are used to and gestureing; basically using a computer yet not knowing you are using one. I wished they went through each system more indepth how they reconciled the postit notes and the computer models of troop locations.
Robert Taylor - Sep 10, 2006 05:27:04 pm
POET: I think one of the most important and concise things mentioned in the reading is the three principles of psychology: visibility, appropriate clues, and feedback. It's interesting to think how we incorporate these in present HCI's.
- Visibility: take folder windows. Generally it is easy to find open/close/maximize/minimize buttons on the window, and the icons themselves are easy to see. We lose some visibility in items in menus though, as menus within menus tend to tuck things very far back. Or, in the case of internet explorer 7 and wmplayer11, the menus themselves are hidden. You must right click the player to bring up the menus. Is this good? It makes design look better, sure, but intuitivity is lost.
- Appropriate clues: The best thing I can think of here are small info boxes that popup on mouse rollovers. Greying out things that cannot be used also make for good clues.
- Feedback: The first thought that comes to mind here are error messages. These really depend on how much info is given back in the message, and what options are given in the window itself. Sometimes a message itself isn't necessary - providing the wrong logon credentials in OSX causes the logon window to shake.
TMI: The author brings up an excellent point covering what is necessary to bridge the gap between technology and people and create better HCI. It's impossible to make everyone conform to a keyboard/mouse standard; not only are some not comfortable with it, but their own interface for interacting with tools they use has had plenty of time to evolve, and probably is pretty good. To summarize one of the doctors, new!=better. We must find ways to get interfaces similar to what people already use, but that are digitized. Digital pen and paper is clearly a step in the right direction; more complex tools people use will be harder to digitally implement, but it is only a matter of time I believe.
Andrew Tran - Sep 10, 2006 08:29:26 pm
POET: I found this article quite enjoyable to read. Who knew even a simple machine such as a telephone can drive people nuts. I agree with Norman in how he explains visibility and mapping are some key factors that helps make a design much better. However designers still can mess up in their mapping. For instance, in my dad's 2001 Toyota Tacoma, when you want to unlock the doors you have to push the lock button down, and to lock is the opposite. What sense is it to have to push down on the lock switch to unlock the car? Designers need to test their products more thoroughly, or even use it themselves to see what just does not make sense.
Tangible multimodal interfaces for safety-critical applications: In my opinion, i don't believe Tangible User Interfaces (TUIs) are practical, at least not for the military. MMUIs are much better given the ability that even if the system crashes work can still be carried out. However, i would not incorporate a lot of electrical machinery with MMUIs because i believe it is too immobile. The military often moves a lot and it would take too much time disassembling the machine. The Anoto pen and paper would work perfectly fine.
Leo Chen - Sep 10, 2006 10:25:49 pm
POET :
This was an interesting piece. But quite honestly, a lot of popular products fall victim to the "elegance over ease of use" fault. Take the iPod for example. If you were to hand the iPod to your father, and told him you wanted to lower the volume on it, I would guess that he would have no clue on how to do so. The scroll wheel is not, at first glance, a well marked way of adjusting volume. Additionally, that big button in the middle is unlabled. But, if one is to view the instruction manual, then it all becomes clear. I'm not saying door should have instuction manuals, but I also believe that the author is overreacting in some cases.
Cohen & McGee:
I guess what this article really stressed was the durability of a TMM. If a computer crashes, operations won't halt, things still can progress. I see this as the main advantage of these types of systems. In terms of convenience, I believe an RTS like interface would work just as well as a massive digitizer and paper map. However, in terms of redundancy and durability, the paper implementations trump all.
Michael Udaltsov - Sep 10, 2006 10:55:52 pm
POET: To me, one of the most important sentences is: "When simple things need pictures, labels, or instructions, the design has failed." This shows how design problems occur in many devices and things that we use daily - they rely on labels or instructions to ensure proper operation. For example, a door with push/pull labels is badly designed since the action isn't intuitive to the user. An example of a more successful door is shown with different handles that uniquely support the affordances of pushing or pulling. The ideas also progress into conceptual models and control mappings. Well designed devices have a conceptual model that matches the operation, with controls that are mapped naturally and uniquely to available actions. These seem to be the additional reasons for whether a design is good or bad. When a conceptual model is incorrect, the user is unable to operate the device, no matter how much documentation there is. Mapping of controls seems to be a very relevant topic today - in telephones, computers, video games, and a lot of other devices. It seems that the best designs will balance simplicity, good controls, and the amount of available features. Many systems seem to fail in one or more of those areas - they're either difficult to use, have confusing or multi-purpose controls that are hard to remember, or don't provide the right features to users. Unfortunately there are increasingly more bad examples since technology is getting cheaper and more companies are trying to get into the various markets with badly designed products, since they focus on profits and market share instead of usability.
Tangible Multimodal Interfaces: The article describes the design and evaluation of several Tangible Multimodal Interfaces for military and medical applications. There's an argument that some digital systems still need an input method that users are already familiar with, such as pen and paper or voice recording, otherwise the new systems won't be accepted. In addition to usability, the pen and paper methods also provide backup information in case the digital portion of the system fails, which is very important in safety-critical systems. I think this definitely applies to areas where traditionally non-digital input methods are used. But perhaps as the use and reliability of digital devices increases, they will become more integrated into those areas, or at least serve in combination with other methods (for example the Anoto system used by doctors for immediate data entry).
Ming Huang - Sep 10, 2006 11:21:47 pm
The ACM article, Tangible Multimodal Interfaces for Safety-Critical Applications, highlights much of the pros and cons about the physical and electronic media discussed in class. The physical medium is preferred for two reasons: paper has the affordances of being light-weight, durable, error-free (in terms of interpretation by other humans), and easy to access by multiple people. The electronic medium, on the other hand, is much more computationally efficient, stores much more information, and are easy to organize. Multimodal interfaces, coupled with physical medium, allows the user to communicate with computers in a much more natural, human way, fits into the user's existing usage habits, and has better ability to disambiguate and integrate information expressed through multiple venues. As UI designers we are much more concerned on a high level of ways to let machines better understand natural ways humans use to manifest their thoughts and intentions. The technical difficulties, like existing user habits, influence our design as well. Someday we might be able to communicate by as little as a thought, which requires little to no work. It might seem silly to think about such obvious point, but this may be the only thing that separates UI designers from science fiction writers.
It was not a pleasant feeling I had when I finished reading the first chapter of POET. Everyday things that baffles the minds of some of the most brilliant. Facinating. It is interesting that personally I did not encounter any difficulty using electronic appliances around me, leave alone doors. Maybe I am just lucky. The author blames it on reasons that can be summarized in two ways: lack of feedback and unusual by some common conception. It is true that a design should be intuitive (or "natural" as the author puts it). The book calls upon us to be more observative and adaptive to users, which is a very important point, because the concept of what is natural and what is not is sometimes tricky to get. It might be what one is customed to, or metaphors used repeatedly and commonly: the so-called conceptual models. Labels, lights are the best visual cues, and if designed carefully they can also serve to clarify the usage of products that might not fit into existing conceptual models.
Dexter Lau - Sep 10, 2006 11:11:29 pm
POET Norman discusses the importance of the natural flow for interface design. More specifically, the affordances that lend to a system's use. He gives and excellent example of a seemingly simple door causing a lot of trouble. He also relates interface design to the natural mapping of things. Often this can easily be seen through natural mappings that are more physical in nature, such as a seat adjuster that is shaped as the seat itself to make it nearly fool proof.
Tangible multimodal interfaces for safety-critical applications In mission critical systems, the ability for a system to continue even when fatal crashes occur is very important. In the case of military tracking, Rasa is a good example of a system that uses current standards and takes it a step further by digitizing it. By moving familiar Post-It notes over a map, units can be tracked easily. And even when a crash occurs, reports continue to come and and the Post-Its can continue to be moved. When the system returns to normal, it can simply re-align with the current positions of units. TMMs such as this allow physical objects, gestures, speech, sketches, etc to be used on top of digital systems easily.
Aleksandr (Sasha) Ashpis - Sep 11, 2006 12:09:37 am
~TMI for Safety-Critical Applications The TMI for Safety-Critical Applications brings up several good points, the two most important ones are that if the tech industry wants to reach professionals, we must adapt to them and not them to us, because everybody has there own way of doing things and a veteran of many years is not going to want to change the way they do business because some CTO has made the decision to do so. The second important point is the quote “a computer map with a hole in it is a rock, while a paper map with a hole in it is still a paper map”. The quote sums everything up with why some stay away from technology and also is a great argument on why the Anoto pen can be a great way to evolve paper and at the same time revolutionize the IT industry.
~POET Who hasn’t had trouble figuring out how to change a feature on a computer or a cell phone, with no standards the users are at the whims of the designer, who unfortunately, usually treat the user interface as a non-issue. What struck me the most is that people keep buying bad user interfaces and cope with them by learning the basic functions and not using the rest, while designers think that since there stuff is selling it must be good, and continue to do the same old thing. This article makes me realize how much more importance should be placed to mappings and instructions.
Kang Chen - Sep 10, 2006 10:18:28 pm
POET: I really liked the examples presented in this article. Certain things such as the design of door handles really surprised me since I never noticed such things before. The author also mentioned, "Added functionality generally comes at the price of added complexity", which is true most of the time. However, I would like to point out that these features are introduced mostly because there is a market demand for it. If there is such a demand for it, shouldn't it be safe to assume that users will, to some degree, bare with the added complexity and adapt to it over a period of time? In the article, Norman mentioned the telephone system at the university that automatically reroutes calls to another phone when no one picks up and labeled it as too complicated to use. However, if someone were to work there for some time, s/he would probably remember the code for common routing points or the predefined routing setup. It is complicated but not nearly as impossible to tame as he had suggested given time.
Tangible multimodal interfaces for safety-critical applications: The article presented the an excellent point that users should not be required to change in order to adapt to the design of system designers. Let's take the example of Unix/Linux OS for example. I've tried nearly every method to persuade one of my former floormates to switch to linux. It's free, stable, secure, and can do just about everything other OS could do and arguably better in some areas so why not? It all boiled down to "The interface seems complicated and different. I don't want to have to spend hours learning a new program just to do something simple." While I would of course disagree with his comment, he did point out something very important - evolution vs revolution. Imagine being in a life and death situation, would anyone want to read a few hundred page manual before they could learn of a way to survive? Probably not.
Vahe Oughourlian - Sep 10, 2006 10:44:49 pm
The Psychopathology of Everyday Things
I could not help at being amused that almost every example Norman brings up (the car, the phone, the radio) I've had problems with myself. However dated this source material may be (mid-80s), most of the points still ring true nearly twenty years later. Here the well-meaning designer is sharply ridiculed by the user. This attitude may change somewhat, since, due to the spread of more technologies in the 90s, the designers of this era come from the frustrated users of the previous era. Norman does quantify, though, really what makes a good design; for instance, the mapping, the appropriateness (or affordances) of certain materials (though, in the discussion of glass vs plywood, he doesn't really provide a sensible alternative), the visibility of features, or the conceptual mappings that people go through will all their devices. However, one aspect he does not mention is the idea of accepted conventions, even despite bad design. For example, the Nokia N-gage tried to elminate the experience of rubbing facial oils all over the screen of one's cell phone by modifying the way people talked into the phone; namely, "side-talking". While a reasonable solution (though simply flipping the device over to talk into the back might have been better, in my opinion) to a common problem, expecially considering the gaming nature of the phone, that feature became one of the most ridiculed functions of the device, since it "didn't look right". People were just too used to the common way of talking into cell phones, which led to the adjustment in the Ngage QD to the more accepted standard, though the problem of obscuring the screen did reappear.
Tangible multimodal interfaces for safety-critical applications
This article describes less, really, about the "safety-criticalness" of the applications, and rather about the problem of users' reluctance to "get with the program", so to speak. Repeatedly in these situations the users need the same input methods they had before in addition to the digital robustness and features, which leads designers between Scylla and Charybdis, where they cannot make the proper choice. Though these solutions, such as those with the mapping and the digital pens, do work and cost a bit more per map, they did not seem to take into account the errors that will crop up with folds in the paper (as maps may be folded, though larger maps are usually rolled). Also, it seems that these solutions sometimes want to do too much with the current technology. Take, for example, the medical solution. It calls for robust OCR technology when it has been repeatedly proven that OCR does not hold up to translating human handwriting, especially that of people under stress, which members of the medical community in emergency situations are, reliably.
Julius Cheng - Sep 10, 2006 11:55:27 pm
POET: This reading was very informative as well as interesting with all of the real-world examples provided along with witty commentary. Norman's view that the designer has failed if instructions are needed to do anything is quite controversial, but a very good design principle to live by. Indeed, given that there are tens of thousands of everyday objects in our lives, everyday items should be easy to figure out. However, the more complicated a device is, the more following this rule is not a black-and-white affair. In the example of the digital watch, where there are buttons but no labels or intuitive way to figure out their functions, what is the designer to do? The watch may be too small for readable labels, and losing stopwatch, calculator, or what-have-you functions defeats the purpose of having a digital watch in the first place. What's the solution?
Also, I think one of the most important lessons one can take from this book is to test your design on real people. Most the poor designs featured in the book are from designers thinking, "Wouldn't users appreciate a function to do x?" and "Wouldn't it be easier on users if we implemented y?" without observing how real people operate the device. TEST ON PEOPLE!
TMI: I think the TMI proposals are very good ideas that will cater to the professional the prefers evolution to revolution. Again, instead of changing the way people do things, they enhance them, a common theme throughout this course. I found the idea of having multiple modalities quite interesting, particularly how they can be used to disambiguate one another. Indeed, OCR, voice recognition, and any other way of converting non-digital communication mediums to digital ones are error-prone, and anything that improves accuracy is highly desirable. However, such technology is far away from being useful for safety-critical uses. The durability, ease of manipulation, and graceful degradation of pen-and-paper systems has been proven to be hard to replace. TMIs need to have a near-zero chance of error. A single error, or much worse, a death resulting from machine error cannot be tolerated, and such an occurrence would destroy what little trust professionals have for high-tech interfaces for a long time to come. Unfortunately, OCR, voice recognition, and handwriting recognition is incredibly far from being error-free.
Robert Held - Sep 11, 2006 01:15:18 am
Cohen and McGee: I've noticed in several of the Anoto-related articles presented in class, including this one, that the digital pen is introduced as a transitory gadget. That is, they present Anoto-based products as a way to ease doctors, scientists, and military personnel into the use of computers for their everyday tasks. The articles then move on to list the crucial aspects of pen-and-paper work, thus negating the transitional image. I think articles such as the one by Cohen and McGee should emphasize from the beginning that digital pens offer a long-term solution. They mention that the pens provide paper copies for medical records, which is an excellent example to support the idea of permanent use in the field.
David Norman: Several items in this passage stood out to me. In particular, his mention of artistic form over usability hit a chord. This weekend I was at a reasonably fancy restaurant with salt and pepper shakers that were identical except for the number of holes at the top. They were stylish objects and fit well with the table and general decor of the restaurant. But my wife and I had a brief discussion over which one you would assume to be salt. We concluded the single hole shaker would be salt, since it would be more necessary to restrict the flow of salt, and that pepper typically requires more force to dump it from the container. But we never knew for sure because we never used them. I think this example fits well with Norman's art vs utility concept. One should not sacrifice one for the other.
I also enjoyed Norman's reference to designers being afraid of confusing the user, and as a result producing controls that make no sense and render the object useless. His refrigerator was a great example. I think the scenario leads one to question the initial design. Perhaps the designers and engineers should have been in better communication from the beginning of the project, so that the final project would be easy to control without lengthy instructions.
Randy Hilarbo - Sep 11, 2006 01:35:19 am
I found The Psychopathology of Everyday Things really interesting. I have not really realized user interface in terms of everyday things until it is discussed in this book. I do agree with the author when he said that bad interface wastes a good design. Personally, I don't bother reading manuals about the functionality of "everyday things." I rely heavily on the visibility of the controls of the design to figure out how to use it. I tend to just ignore the many functions a particular "thing" has because the interface is not very obvious. Good conceptual models, natural mapping and feedback are three principles in the article that are very important to keep in mind.
In the Tangible multimodal interfaces for safety-critical applications, I was fascinated by how it validates how automating alters the aspects of what the users do and value and how this makes them reject new designs. This article again points out how existing methods of doing things are so important. I also find the digital pen/paper applications that this article talked about very interesting.
Cheng-Lun Yang - Sep 11, 2006 01:52:56 am
Tangible Multimodal Interfaces: The creation of Tangible Multimodal Interfaces has many causes. The first cause is due to safety and disaster proof problems come along with computer based systems. The second is the educational status of general public and human being’s psychological trend of reluctant to change from old, more familiar technology to new technology. Using pen, paper, and voice recognition system for data entry to computer database is a middle ground to solve the problems. It makes the general public more comfortable by using traditional pen and paper. Also it prevents data loss when the computer system crashes by backing up with the paper copies. I believe Tangible Multimodal Interfaces are necessary and will always be needed with the creation of new technology.
POET:
From this article we learned that a well designed product is not a product with the most hidden features, it is the one that’s the easiest to use. There are too many products on the market that features many features in one thing. However, there are only a few operating buttons or switches. Users usually need to press a combination of buttons to get their desired job done. So engineers should always aim for easy to use products that even a kid can operate even though that means cutting down the fancy features. Simplicity and affordance is the key to all design products.
Utsav Shah - Sep 11, 2006 01:45:56 am
Cohen & Mcgee: Again, these articles convince me first time I read it. As with other digital pen based articles, this one also talks about making the user's job easy. They talk about how instead of making users change their style, designers should adapt to users' work practice. What really amuses me is that even after making these products user-friendly, they're not going anywhere.
Don Norman: This was an interesting read because Mr. Norman talks about things that we take it for granted or not normally think about. After he talked about the phones, I looked at my cell phone and realized that I don't use some of the features simply because it's not clear to me what it does or how should it be used. Another interesting point Don brings up is that it should be visible to user what the object is designed for, trial and error is an example of bad design. I think some of the most badly designed machines are the ones at the gym. It's not clear how to use some of them, it usually takes some trial and error.
Mjiang - Sep 11, 2006 02:14:07 am
POET: I really liked this article for the fact that it made me not feel ashamed to admit that I cannot operate doors as well. The doors of VLSB's main entrace have 3 doors, 2 face each other. If you do not know which side to push, you are stuck. Unlike the hotel design mentioned in the chapter, those doors have a horizontal bar that extends the whole length of the door. I can't even count the number of times I couldn't figure out which side of the darn door to push. I also like this article for pointing out that elaborate things may not always be the best. Designers do want to cover every basis of functionality, but by doing so, causes an opposite effect of not being able to send the message of how to use it easily. Human beings are lazy creatures, I guess sometimes we over estimate our willingness to read manuals explore complicate but well meant products.
Tangible Multimodal Interfaces: This article provided another useful application for the digital pen and paper. Like the ButterflyNet article, this centered upon a very specific type of task for the users. It almost seems as if the digital pen and paper seems to be popular only within specific regions with specifically tailored software for that region. I wonder if this is why the anoto isn't as popular as we thought it would be.
Sean Carr - Sep 11, 2006 02:20:55 am
POET: I cannot say how true this book is about common everday design. So often products are designed a certain way more because of form than function. I think we see it even more with current technological advances. It seems like so many high tech inventions are "revolutionary" in many ways but fail to catch on because of poor usability. One thing I'm surprised they didn't delve into more is standards. Standards are the foundations and glue of many science fields and are extremely important in computer science, but I think they are even more useful when applied to user interfaces. They talked about visual cues and certain visual elements conveying a certain affordance, but unless these things are standardized every designer has to conduct his/her own research to figure out what the best symbolism/visual cue for each aspect of their design is. Since this is not possible, designers usually make up whatever makes sense to them and this is why we have so many of these problems to begin with.
Safety Critical Applications: It seems like this is a move in the right direction in order to find a middle ground between the digital and physical world and use the benefits of each. However, it seems like they are still overlooking some problems with their systems. We've talked about several of these Anoto based collaborative mapping tools and the one thing I can't seem to understand is how they are truely collaborative. Yes the digital aspect makes it easy to transit/share the data but if two groups are working on the map at two different locations how do the changes get merged? The map doesn't update like a computer display because it takes a long time and a lot of money to reprint it. The Rasa system at least goes as far as projecting the current digital map onto the paper map and declares that if the computer crashes no data is lost because the map still has the data. This isn't the case in a collaborative environment unless you retraces everything that is drawn at the other site. The NISMap system doesn't even use the projector idea so how does one get the data from someone elses map onto theirs without printing it all over again? They make these map systems sound like they are great for collaboration but I think they should read POET and think more about the possible use cases before they can be so sure about the capabilities of their system.
Antonis Mannaris - Sep 11, 2006 02:48:03 am
In POET, D.A Norman addresses an extremely valid in my opinion subject about the way technology is headed. The technology paradox as he calls it, where added functionality adds so much complexity that the usability and essentially benefit of a product diminish. Maybe the old cliche that old people cannot operate modern appliances is not their fault after all. On the bright sight however, not all technologies are lacking good design when new features are added. Look at cell phones for example, a technology which in my opinion has been developing as rapidly as any other technology in the world today. Cell phones today are also cameras, music players web browsers and essentially mini-computers. Nevertheless, this added functionality has not made the cell phone unusable for anybody who can read and interprete simple pictures. The idea of a "menu" is essential in this, and the designs of most menus are sufficient for an average human being to use. Cell phones have both visibility and good feedback, because of the screen! Imagine using a cell phone without a screen. Not so easy I bet.
In the second piece about TMIs, I feel that a very important enhancement to the types of User Interface and input studied so far was the combination of more than one such interface. The idea of synchronized writing and speaking has positive effected on almost all applications that the ANOTO pen and other types of technology may be used. It also gives a great tool to field workers like researchers since talking about a find requires no heavy or complex equipment, and everybody is accustomed to it.
David Eitan Poll - Sep 11, 2006 02:51:54 am
In the Psychopathology of Everyday Things, various flaws and common errors in design are discussed. The author makes some particularly astute observations, my favorite of which regards door design (I have found myself in the "trapped" situation many times). I think the important thing to take away from the reading is that it's very easy to be taken in by goals of aesthetics and low-cost, but these often result in a backwards-step in design. There is a good way to add power or features to a product, but it should not be at the expense of usability or understandability. It is these two factors that will decide the user's experience, far moreso than the look or feel of the product.
In Tangible multimodal interfaces for safety-critical applications, examples of effective systems are provided. I think that the notion of implicitly backing up data through physical duplication is highly valuable. The example they gave regarded military planners operating through a system malfunction, where the backup was implicit because physical objects were used to represent the entire state. It required no additional user interaction, and was effective in allowing the required tasks to be completed successfully. This has particularly useful implications in our projects using the Anoto pens, as they too leave a physical backup of all data. Unfortunately, with the Anoto pens, the data cannot be reconstituted automatically by the system if all it has is the "physical backup" (the paper that was written on).
CarrellKillebrew - Sep 11, 2006 02:38:02 am
POET This may seem like a minor point, but the author states that 'there is no concept of more or less in the comparison of different pitches'. This is most definitely not true. I can easily tell you which pitch is higher and lower. Anyway, the major points made are very good ones that I had noticed before, but never stated in such concise terms. I really thought mapping, feedback, and make things visible were the most important points made.
TMMI for safety critical applications I thought the interesting thing about this paper was the discussion of the two mappings sytems, and how a design priority was that the computer failure mode does not render the system useless. So then essentially they provided added functionality on top of the paper version, it seemed to me. They left the input system mostly the same for the user's sake, which is of course good. On the other hand, it seems to me what they did is fairly obvious too. Just upload a digital copy of the physical model the users are working with.
Qingyun Tang - Sep 11, 2006 03:18:26 am
POET: This article points out a lot of small UI in our everyday life. UI is built into small things like door knobs, cars, refrigerators, and even watches. Without proper design of UI, it will waste a lot of time for people to figure out how those things work, and sometimes bad UI design even costs accidents and tragedies. It is very important for us UI designers to consider the intuitive function of the products and thus make the interface reasonable for users. Good UI design will also result a good sales, because people expect to use a product right the way after they purchase it instead of reading a manual for 10 hours.
Tangible multimodal interfaces: This article mainly talks about how hard to make users to get used to a new user interface. In fields like military, electronics require high durability and smooth functionality. Just like other people posted above, OCR is really urgent to be developed in order to improve the user interface. Without OCR, digitalize things does not become so convenient; everything just turns into pictures. However, if we can improve the correctness of voice recognition, it helps with the user interface develop as well.
Eric Vacca - Sep 11, 2006 01:22:30 am
The Psychopathology of Everyday Things: I agree fully that there needs to be a natural mapping of an interface to its function. This occurs naturally in mechanical tools (such as scicsors), but with the digital age the natural mapping does not come natural at all, and so Norman is right on in saying that designers now must take extra care in their design.
Tangible multimodal interfaces for safety-critical applications: I disagree with Eric Yoon's comment that if digitized information is less reliable. Many times i have written something down on a post it note and later overlooked it when i needed it. Furthermore (this is somewhat of a tangent, but interesting nonetheless), I recently read an article hypothesizing that digital information one of the only things (if not the only) in the universe that is not subject to entropy, that is to say, once digital information is orginized it stays that way and does not degrade.
Regaurding possible errors of the post it note TMM interface, the author also did not address ambiguities of using post it notes as opposed to push pins. Push pins are exact and represent a point in space, whereas post it notes leave the exact location of troops unknown and anywhere within the area of the sticky part of the note? Overall i think the concept is sound but the technology needed to incorporate the concepts reliably are still years away.
Chen Chang - Sep 11, 2006 03:59:52 am
The Psychopathology of Everyday Things: I found this to be a really good read as it was very easy to relate to many of the practical albeit embarassing situations presented such as going full speed at a closed door but guessing the wrong side of the horizontal bar handle and getting halted to an immediate stop, ouch! Norman's book does a brilliant job of presenting the bright and gloomy contrasting sides of designing in stating that "well designed objects are easy to interpret and understand" whereas "poorly designed objects can be difficult and frustrating to use". Chapter 1 definitely enchanced my appreciation of everyday objects that look so simple but still have countless hours put into the design process to make it a good design. I felt that the visibility example of invisible door hinges has no better way of depicting an example of an incomprehensible design that leaves many thinking "what was the guy thinking?" Modern electronics and technology can be so intimidating to those that are not tech savvy which explains the "paradox of technology" as there needs to be a balance between bells and whistles features vs the elevated learning curve. An example that comes to mind is part of my observations over time through students that bring laptops to their classes is that many simply use the default settings on their computer like the windows xp blue bliss theme, it leaves me wondering if they are satisfied with it or don't know how to or don't bother want to bother changing it as there are a slew of aftermarket themes available for download on the web.
Besides the importance of making things visible, Norman also stated the need to provide a good conceptual model, a natural mapping, and useful feedback. I totally agree with these points as I have used objects where I intend for one function but instead the opposite of what I want happens, clearly the result of a mapping problem. I think its natural to be accustomed to feedback, take for example a situation where you just moved into a new house and flicked on a light switch but you can't find what light or what device you actually activated because you don't notice any changes in the room you are currently in, aggravating if you ask me.
Tangible multimodal interfaces for safety-critical applications: This article presents critical military and medical applications in which the Anoto digital pen is used. I think the ideas are headed in the right direction, but at the same time I feel there are issues still in development to consider. The main one would be speech recognition as its not easy as simply talking, every person has a different voice with different frequencies and reverberations. With these critical applications, I believe errors must be avoided at all costs so voice recognition must be polished and refined before it hits the big stage.
Suthee Chaidaroon - Sep 11, 2006 04:43:32 am
POET: This reading points out some common errors in a poor product design - the lack of visibility, inappropriate clues, and no user's feedback. He said these principles constitute a form of the psychology of how people interact with things. I like his idea of natural mapping in which taking advantage of physical analogies and cultural standards and leading to immediate understanding. It seems to me that a good interface will require many redesign iterations. And it is true that controllability always come with usability.
COHEN: This paper presents some military applications that take advantage of both physical and digital world. It is very critical for this military application to be super accurate and no errors. Since the system is composed with OCR, symbol recognition, and speech recognizer, how disaster it will be if, for example, the symbol recognition mistakenly interprets a retreat comand as an attack command instead.
Vijay Rudraraju - Sep 11, 2006 09:58:34 am
POET This chapter of Norman's book was an eye-opening experience. I am accustomed to assuming that whenever I buy a new camera or stereo, I must be willing to spend a significant amount of time learning by trial and error how to use the features of the equipment. And I have certainly had experiences where my friends and I engaged in a battle of wits competing to see who could figure out first how to open some poorly designed window. And I don't think I have ever been able to properly operate an electronic thermostat. As much as I like puzzles, it much more interesting to think about what it is about certain objects that make them so easy to use. Among the concepts that Norman outlined, proper feedback mechanisms seem to be the most significant hurdle in a well-designed digital pen/paper application. Now if I could only figure out how close the damn blinds in this room...
Tangible multimodal interfaces for safety-critical applications The authors of this paper present interesting ideas for augmenting the shortcomings of pure pen/paper systems with multimodal input and feedback. The most promising aspect of these systems for military applications is the idea that the dependence on the computer components of the system is not absolute. In the case of Rasa, officers can still plan and plot movements when the broader system is down and update the latest moves on the system when the computer comes back up. The NISMap system can also have this type of resiliency as long as pens are easily replacable and there is enough local storage on the pen in case the database goes down.
Anirudh Vemprala - Sep 11, 2006 10:31:38 am
POET: I liked the fact that the author addressed the problem of having multiple functionalities mapped to a single button/controller. In today's micro-sized gadget crazy world, a number of cases of poor design is a result of problem. I wonder, however, if this idea can be "scaled up" towards devices that attempt to combine a number of functionalities - music-playing, photo-taking, fax-receiving cell phones in particular. At what point do these devices start to deliver diminishing returns to their users? Is there a limit to what a device can 'afford' to their users?
Cohen: I think there has been a lot of stress in the past few weeks on the benefit of commonplace UIs like pen & paper. As programmers, I think we forget that and assume that everyone is happy typing away at a computer. A great supplement to something like the Cohen reading would be actual discussions with Cal military studies instructors who could tell us about their preferred computer interfaces and the challenges they face with existing UIs.
Jonathan Chang - Sep 11, 2006 10:36:53 am
POET: An interesting enumeration of the different considerations involved in a good user interface, and indeed what it means for something to be a good user interface. One problem I have with this article is it doesn't seem to distinguish between a simple and understandable interface, and one that is merely familiar. The example of the telephone hold function, for example. The blinking lights are a good design to be sure, but is the resistance to the tap button truly an inferior design or merely unfamiliar?
Tangible Multimodal whateverstuff: An excellent point; no need to fix what isn't broken. Tangible interfaces immediately suggest themselves towards people who've been oriented in this direction since time immemorial.
Yen Pai - Sep 11, 2006 08:01:08 am
POET: A very informative piece on good design and the competing requirements that a product designer must fulfill. The concept of maximum visibility and high feedback are often opposed by cost measures and a need to accomodate "expert" users - users who know what they're doing and want a shortcut through the menus. One can imagine how the very verbose instructions of office telephone systems and less than helpful interface might be a product of this sort of thinking. If you can activate some telephone feature by dialing * and then some code, so-called expert users are able to reach features quickly and novice users have the manual to fall back on. The company producing the system saves the trouble of producing a better interface and pushes the burden onto the technical writers responsible for the manual - a phenomenon that seems rampant in electronics today. Another issue, illustrated perfectly by comparing cars and telephone systems is that some products lend themselves to better spacial association of controls. For example, there is much more surface area to play with in an automobile and functions like steering lend themselves well to direct association with a movement, while something like "3-way calling" is harder to designate physically. Also, it should be noted that while essential functions like steering, acceleration, braking, and perhaps cruise control are often very intuitive in a car, the same cannot be said for car navigation systems and stereos. Part of the reason for this might be that the best designed features are ones that an automobile company cannot afford to get wrong for safety reasons and therefore, in these cases, there is alignment between the sometimes competing interests of cost, good UI design, and maintainability.
Tangible Multimodal Interfaces: The essence of this article presents the fact that "old habits are hard to break", but it also speaks to the potential disconnect with existing practices and so-called "newer, better" forms of data input and storage. Resistance to new technologies and processes can be magnified when the user group works in field where safety and time are critical and perceives themselves to be short on time (any new action, including the learning of new processes, may be perceived as a burden). Further, users in such professions, like physicians and military officers, follow traditional processes that have been proven and honed in real-life situations over a long period of time. Perhaps the a primary goal of any user interface is to make adoption easy: all the elegance in the world does no good if nobody is willing to use it. Once a user interface is adopted, the primary user group can enter the design cycle, iterating the design under real life situations, a process that not only improves the design but promotes the confidence of the target users in the new processes. A certain level of confidence is required when users have to work around new problems introduced by a solution, for example lost pens and computer crashes.
Jack Yeh - Sep 11, 2006 10:52:42 am
The Psychology of Everyday Things This article points out that feedback should be plentiful, and the response time should be close to immediate. Imagine that each response would come back in 2-3 minutes. At that time, the user could be doing other things and mentally connect this response with the wrong function. Therefore, we try to eliminate extra features with long response time. However, those features could make or break a product because that's what consumers look for in the first place.
If there's a way to get the message "More isn't necessary better" to the general public, designers would save tons of trouble without worrying about interfacing weird features with limited buttons (the telephone)
Tangible multimodal interfaces New system with new interface often alienates users whose main concern is reliability and re-usability(if one component were to fail, does it still work?) Digital system makes the storing and transmitting easier and faster, but at the cost of less-reliable and no protection. Combining two systems (one digital and one traditional) sounds interesting but the syncing issue would be hard to resolve for the years to come.
Johnathan Hawley - Sep 11, 2006 11:27:56 am
The Psychopathology of Everyday Things - I found the paradox of technology something I had never thought too much about before. It is interesting to think that designers before my generation of devices like radios, phones, and watches paid more attention to user feedback than designers do today. Some older technology has advantages because of this. I have a professor who doesn't seem to be up with all the latest technology. To teach his lesson he brought out an old projector and the lesson went on fine. In contrast, most of my other professors use their laptops to lecture from. While there is more functionality with the laptop, I don't know how many times a lecture has been delayed due to some kind of 'technical difficulty' so to speak.
Tangible multimodal interfaces for safety-critical applications - I'm more sold on TUIs than MMUIs. Having to talk to my technology seems a little awkward. It reminds me of when I had to call my insurance company and enter my phone # vocally. I kept having to repeat myself over and over. I still don't think voice recognition is that good. Didn't the previous article say that companies have given up on voice recognition because it has failed too many times?
Suneet Shah - Sep 11, 2006 11:29:25 am
The Psychology of Everday Things: This was a great article that was a great crash course in things you should consider when designing anything. The user-centric focus that this article stresses is critical, and there are many thigns out there today that would have benefitted from their designers reading it. I too was surprised to see that most products take 5-6 iterations to see success. It was interesting to read about things like doorknobs and other everday objects that are poorly designed.
Tangible Multimodal Interfaces: I think this article really highlights people's resistance to change and their lack of trust in newer technologies for things that really are critical. But this is funny to me because there are plenty of things that are extremely critical that rely entirely upon tecnology. For example, air traffic control and planes. They rely so heavily on automation but we have no problems trusting them. I think people like TMM's becuase having a hard copy and a soft copy gives them some sort of comfort in knowing that there is a hard copy backup. However, this does add additional overhead in having to maintain both.
Huangnankun - Sep 11, 2006 11:28:48 am
The psychopathology of everyday things
The swinging door example articles brings up the important debate of functionality vs design, which is an important decision that has to be made in the making of every important product.
In the leitz slide projector example, the author brings up the topic of elegance of design vs the need to appeal to the intuition of the audience.
One product which I found has been excellent in functionality, design as well as appealing to the intuition of the users is the ubiquitous ipod. The scroll wheel on the ipod is a work of genius. The circular shape of the wheel prompts the user to move their fingers over it in a circular fashion. This is the “mapping” principle as stated by the author. The main interface in the ipod also has the scroll bar on the left side; this automatically prompts the user to move his finger down on the left hand side of the wheel, thus turning it in a clockwise fashion.
With the projector example, the author brings up the topic on the importance of visual feedback. Many products designed today does not give this kind of visual feedback, this makes users puzzled about whether the operation was completed.
The phone example brings up the importance that designers have to appeal to the user’s habits existing habits.
One product which I think does great in terms of visual feedback is the Logitech G5 mouse. In a traditional “gaming” mouse, there is usually the option to dynamically adjust the sensitivity of the mouse via the mouse wheel and a button combination. However there is usually no way for the user to know what sensitivity setting he is currently at. The Logitech G5 mo use has a simple LED slider that displays this and it makes the user’s life much easier. The LED slider is also aligned in the direction which the scroll wheel turns to adjust sensitivity, making the operation even more intuitive.
Tangible multimodal interfaces for safety-critical applications
This article brings up the problem that the advance of technology is being halted by the lack of usable and intuitive human to computer interface that allows users to operate modern technologies efficiently.
The author suggests using TUI or tangible user interface, which are physical objects that can be used to interface with a computer. Since users are much more used to manipulating physical objects, these will ease their operation of the system.
The author then suggests the combination of TUI and multi-modal use interface to make interaction with systems more efficient.
The usage of TUIs can be easily seen in modern day jet planes. Older planes use physical connections between the flight stick and the various control surfaces on the plane in order to control it. Nowadays most jet planes are controlled by computer and hydraulics, but the joystick and rudder peddles remain because they are physical objects which the pilots can easily manipulate.
The author then goes on to give example systems where TUI and MMUI can be used in conjunction to take advantage of the strength of both while avoiding the pitfalls of both. These interface systems take advantage of the ease of use provided by TUIs while taking advantage of MMUI for storage and feedback.
Sung Yi - Sep 11, 2006 12:16:02 pm
Norman: Simple devices must be very intuitive in terms of design. For example, the telephone or the QWERTY keyboard could have been much more efficient with different design in terms of performance. Although I agree somewhat with what Donald Norman is saying about how design should be easy and intuitive for the user, some devices have so many features in it that it cannot possibly be more intuitive than simple devices. For example, like in digital camera, how could it be so intuitive to use when there are so many functions and features? In this situation, there’s no way but to educate the user about the device using some kind of manual.
Cohen: This article points out that the lack of human’s reliance on the digital system is due to less-prolific and unintuitive interfaces, and lack of portability (PDA and tablets are still not enough) and heavy-weight. However, I think there must be a major reason behind why people do not depend on computer so much: computer errors. Although human-error must be more frequent than computer-errors, why do people decide to rather take the consequence coming from human-errors?
Siyan Wang - Sep 11, 2006 12:16:43 pm
POET The author brings up some very interesting points that seem so obvious, yet its even more obvious that many designers completely ignore those points. Also, I think in addition to his 3rd stage of technology where people add more functions, making things even more complex, there is a stage where designers believe the technology so well accepted that they create things that stress asthetics over useability, which to me seems completely counterproductive.
Tangible multimodal interfaces for safety-critical applications I think this is a very interesting approach to fuse interfaces that take different modes of user input. However, for these safety-critical applications, I am skeptical of our current level of technology for things like OCR or voice recognition, which the article described that it would eventually incorporate. The examples described in the article seem simple enough, but on the battlefield, a stressed commander barking out the platoon name might be misinterpreted by the system, which could be very problematic.
Tom McClure - Sep 11, 2006 12:00:15 pm
POET
Great observations, a must-read for any product design team. I got a kick out of the "phone with a screen" idea, definitely dates the article. I was disappointed that the examples of bad design were mostly critical, only a couple times did the author actually clearly suggest what a better design would have been and how to apply the "natural mapping" principle to the bad design to make it a good design. For instance, he mentions that the burner controls on a stove are usually placed with two on the right and two on the left (to control the left two burners and the right two burners) and he correctly criticizes the fact that within each set it's hard to tell which dial controls the back burner and which controls the front. Why not mention that placing the back burner controls slightly higher and the front burner controls slightly lower -- so they naturally map visually to the actual burner locations -- would be a natural solution? Then again this is just the first chapter, I'm sure the intent is to frame the problems first and then discuss detailed solutions later on.
Tangible...
My wife is a physician, so I have observed first-hand the resistance of that field to digital charting. Reliable context-specific voice recognition (with deep "understanding" of concepts a la cyc) and handwriting recognition are very hard problems, but they are in my opinion the keys to the ideal user interface for computer-assisted applications. Only when these technologies are robust will it be possible for computers be accepted with minimal disruption (and fear) and maximum user satisfaction. The acknowledgement that paper does not have to be scrapped entirely for certain domains is a step in the right direction. Augment current workflow and respect current practices rather than trying to reinvent the wheel.
Edward Karuna - Sep 11, 2006 12:17:52 pm
Clearly what the ACM paper addresses is the point that many professionals are resistant to change, and so are more amenable to change when the changes made to their tolls are imperceptible, or at most not fundamentally jarring. With the Anoto pen, the fundamental interface of pen and paper are preserved, and the benefits of digitality are added. Thus, we can gain the implementations of the multinodal interfaces.
What I find interesting in POET is that the author write often about how lovely, or how beautiful all these things are, but, ultimately, how that beauty or elegance causes difficulties in their use. The point that I take away from this is that things should, for the best experience, be designed to be used easily first, then elegance or comsmetic beauty be incorporated. Not the other way around, because then, so many of these things simply become interesting paperweights, just a conversation piece, perhaps. "Look, you won't believe how you use this!" comes to mind as a phrase I have used many a time when exhibiting such objects to my friends.
Michael Mai - Sep 11, 2006 12:05:12 pm
POET - As technology continues to advance, many of the things that are simple will become more complicated as people have the urge to want more functionality from fewer things. Because of the complication of more objects, I believe the concept of feedback is becoming fundamentally more important for both the user and the manufacterer. Giving feedback allows for visual information to be displayed and it helps the customer to establish a mental map of how to use the object. If one doesn't look at the instruction, a person usually goes about learning the functions through button pushing and memorizes and the designer should take this into account and make a mild learning curve. It is well worth the cost of a display with two more buttons because the user can understand and use the product better.
TMMI - Patients go to their physicians for personal attention and care. By using technologies that are foreign to both the physician and the patients, the visits become impersonal and like the governmental system and perhaps the school system, you become just a number with no real personal attention. By using the paper and pen technology, companies can introduce their technologies without being invasive to the physicians and patients. Perhaps a way to slowly convert the large number of doctors to more computerized methods would be have their technologies attached to the new generation of doctors that are currently at medical school. As more graduate, more professionals will be accepting of the technology and hopefully the rest will follow.
Jason Lee - Sep 11, 2006 12:11:23 pm
The Psychopathology of Everyday Things
Since my project proposal was based on the idea of videogames, I had them on my mind as I was doing the reading this week. The author of this book's point of the absence of visual clues in many poorly designed interfaces made me think about some early videogames of the NES era, in which it was sometimes impossible to figure out how certain parts of the game work. These days, most videogames come packed with some kind of tutorial within the game itself in order to help the player figure out all the controls, features, and rules of the game. In the early days (mid 80s - 90s) of videogames, games rarely offered any help and often required careful and thorough reading of the manual to figure out all the controls and objectives of the game. And through talking with friends, it was obvious taht not everyone read the manual. Furthermore, even the manual left out a lot of gameplay nuances. An example of this is in the Mega Man games, in which the main character, Mega Man, acquires new weapons every time he defeats a boss robot. However, the operation of some of these weapons were not explained and took some fiddling around to figure out how they work. Most weapons are operated much like the regular weapon in the game (tapping the B button), but some weapons required a different tactic. The Charge Kick weapon from Mega Man 5 required the player to slide into enemies (down + A) to attack people. The Super Arm weapon from Mega Man 1 required the player to be next to a certain type of block in order to pick it up and throw it at people. Starting with Mega Man 6, the game gave a demonstration of each weapon's use as soon as it was acquired, eliminating this problem.
Tangible Model Interfaces
To me, the idea of Tangible Multimodal Models seems a bit confusing, as there are too many ways to record information. It is true that the different methods of recording are optional, but too many available options can lead to confusion in the way that the system is used. The Rasa model is a good example of this, as the user can put down information on a Post-it and also vocally record information about that particular symbol. To me, this is combining too many paradigms at once. However, I do like the idea of the graceful degradation of the system, as the Post-its can still be used as a "hard copy" of the data, similar to the Anoto digital pen and paper. Perhaps i am afraid of change, but too me a system with too many input modes can lead to confusion.
Gene - Sep 11, 2006 12:38:52 pm
In POET, I liked the comparison between the car and the telephone. The comparison well illustrates how complex things can have conceptually easy-to-grasp controls, and the opposite--how something with simpler functionality like the telephone can end up with ambiguous controls.
Before reading the paper on TMMIs, I had not realized that MMIs can reduce error in the way that they do. In the end, though, I think it's a major plus to be able to have different ways to input and access data.
I've also never been aware of the chasm between early adopters and "conservatives" I guess there's a whole spectrum of how mature a product must be before a person is willing to try it. It's the vast span of this spectrum that is slowly becoming more apparent to me. I think this relates to the question posed by the writer of POET, of why we put up with everyday things that don't function desirably. My answer would be the following:
That just as people refuse to adopt early because there may be not ultimate gain in productivity, people put up with everyday crappy things because it takes so much effort to fix them, especially when their attempt may end in fruitless failure.
Siu Pang Chu - Sep 11, 2006 12:11:04 pm
POET is very important reading that mention many commone errors in design that people have made. The most interesting example for me is the desgin about the door. Different size and position of the bar can be good enough to tell the user which way to pull and push. A good design is always like that sending a clear signal to user , so that the user won't be confused by the fancy function. A good design is a like this door , no matter how many function it has, at least it should tell the user how to use(open)it first instead of reading 10 min instruction to figure out how to go out. Other good design ideals that are providing a clear conceptual model, natural mapping, and meaningful feedback.
COHEN I really agree that "computing too often requires professionals to alter their work to fit current technology". A new product should fit into the user instead of asking their customer to adapt the new software. One important issue of the new is to fit into the current envoirnment of their target market. For example, for military, their current environment is maps, commands. Thus, the new system have to have OCR and speech recognizer. Also, since military is a big issue, the software have to be accurate and providing error checkings.
Kimberly Lau - Sep 11, 2006 12:25:44 pm
Tangible Multimodal Interfaces: Tangible multimodal interfaces seems like a good idea -- it aims to incorporate only the most optimal conditions of physical interfaces into digital avenues. Being able to employ both paper maps, post-its, and vocal information for a combined digital effort can significantly simplify inputting such data with manual methods. However, the limitations remain that neither of the fields this technology hopes to juxtapose are very developed. In combining them, the final product may not function as expected and could hinder future prospects of acceptance should these initial tries prove too inconvenient and difficult.
Psychology of Everyday Things: Donald Norman's paper considers the design of everyday things. For me, the design process is critical in creating the best product possible. This includes considering users: their needs, how they react to an object, seeing the overall picture, because a bad design will result in mistakes as users attempt to use the product. There is no one way to guarantee a good design, but his proposed natural mapping idea is a nice way to start.
Simon Tan - Sep 11, 2006 12:43:01 pm
POET: This article was entertaining. I never realized how badly designed the things around me were - I guess we have just put up with them for so long that they suddenly become tolerable. I agree with the comment above that designs are intuitive to us based on the user's past experiences. Again, this is an example of using old designs not because they're better, but because people are more used to them. Which works, but leaves people hard-pressed to find innovation for a toilet, faucet, etc.
Tangible Multimodal Interfaces: One aspect of the concepts introduced within this article stood out to me - the fact that the applications would still work even if the digital system behind them crashed or lost power. When we think of our digital systems today, we are usually confident that everything comes to a halt if the power goes out - however, if everything worked the way suggested in the article, people wouldn't mind using computers and digital systems so much because of their fallback nature.
"Physical objects and computer systems have different failure modes" - this line brought the concept home. We need more failsafe systems in our lives.
Jerry Yu - Sep 11, 2006 12:58:06 pm
It's interesting how several of you have noted that Don Norman seems to suggest that aesthetics and look and feel take a back seat to usability. Many of Norman's critics have also noted this, and his emphasis on user-centered systems design is often used more as a charge against him than as praise. To paraphrase Norman's own paraphrase of his critics, if designers followed Norman's advice, products would be usable but ugly!
POET was first written in the late 80s. Since then, Norman has taken a more accomodating view of aesthetics (or maybe has had this view the whole time) and has written a book titled Emotional Design, where he explores a more "holistic" view of product design, including how aesthetics can (sometimes) overcome less than perfect usability. Straight from the horse's mouth, here's an article on the Nielsen Norman Group website about the book.
Also, for more on what he's doing now, here's an audio recording of a lecture he gave at the School of Information last spring.
Bryce Lee - Sep 11, 2006 12:57:48 pm
POET This article brought an interesting take to UI design. Rather than focusing on a particular issue, the author expands the scope to the state of design in general. If you look at a modern software manual, it is apparent that design has gotten out of hand. Becoming fluent in using a program now takes an investment of time, marginalizing the payoff of using technology for anything other than repetitive tasks.
Safety-Critical Applications The applications mentioned in this article highlight the epitome of usefulness for computer applications. Many times, new advancements actually hinder productivity as mentioned by the Cedars-Sinai Hospital staff; however, they are still adopted in the pursuit of utilizing technology. The military application highlights a seemless integration of technology, an unobtrusive addition which provides real, gracefull-degrading benefit to the user without any learning curve.
asdf
Charles Lee - Sep 17, 2006 03:19:39 am
Psychopathology of Everyday Things 1:
Visibility of features, much less instructions, is a very important part of design that many computer programs neglect. Many programs that provide shortcuts do not provide easily accessible lists of such shortcuts. One personally annoying example is Windows popup dialogs that allow for a keyboard way to select an option (alt-O for OK), but do not show which letter is used for the shortcut. In previous versions of windows, the proper letter would always be underlined, but many dialogs now have no underlined letters by default.
Psychopathology of Everyday Things 2:
It seems that the number of features unrelatable by a natural mapping is the complexity of the application. Natural mappings, when used properly, have already been learned by users, and it would be a waste not to draw upon that learning. Moreover, it is significantly more effort for the user to learn something that they do not already know, especially if that thing is rote memorization. Thus features that cannot be analogized to a natural mapping become unused features.
Tangible multimodal interfaces for safety-critical applications 1:
I find it interesting that military officers and medical practitioners insist on using paper and pen instead of potentially more flexible and useful formats. These people work in matters of life and death, yet they insist on a low tech format. That would imply either that these people are indirectly putting lives at risk for the sake of tradition, or that the known, existing, digital alternatives are indeed worse than the format they seek to replace.
Tangible multimodal interfaces for safety-critical applications 2:
The article mentions that legal reasons require a paper copy of some records. This is an interesting usage of the pen, with requirement of a backup physical copy. The same concept could be used to comply with Sarbanes-Oxley record-keeping requirements, allowing for easier record-keeping, one of paper's basic usages. Storing signatures could decrease the significance of signatures, though.
Jae Chang - Sep 18, 2006 02:49:19 am
POET (1):
Great observation and consideration for designers. The article points out the importance of the interface design. Product design with intuition and visibility are crucial facts for success as well as functionality. By contrasting tradeoffs of design and prices, the user’s usability is emphasized as well as psychology. As engineers focus more on features and technologies, user usability or user interface can be less-weighted. In the author’s point of view, industrial standardization can be made only if both functionality and usability meet users’ needs.
POET (2):
Providing product by solving should make people’s life much easier. However, some products based on innovated technology is very difficult to adopt not because of its functionality but because of its usability; sometimes new technology makes life more complicated. Intuitive understandability and usability from good perceptual model is more important than its affordance. I agree with author’s saying that inventor should make things visible because visibility which creates interaction between users and product provide easy interface to access with. Moreover, feedback phase takes the role as the most important key to success for company and customer’s satisfaction both.
Tangible multimodal interfaces for safety-critical applications (1):
As the usage of computing increases, the needs of easy and fast user interface for computing also increased. An example of such interface can be MMUI which creates connections between physical world and computationally. As the author implied through the article, I believe that the next generation of UI after GUI will be TUI or MMUI because such systems provide physically and computationally in the current existing workplace. The major advantage of TUI and MMUI is that it can be used seamlessly and can be adopted very quickly in our physical world.
Tangible multimodal interfaces for safety-critical applications (2):
In the article, the author stresses people’s resistance not to trust and adopt newer technologies by giving examples in safety-critical arenas. The fact that computer-errors can be great treat in the areas might be true. Digitalization and computation of information can reduce so much error and be more accurate than human operation or computation. What if hospital is on fire without patient record backup system over network because pen and paper is better than computer UI? The author’s suggestion of TUI and MMUI can be the solution for the circumstances.
Yang Wang - Sep 18, 2006 11:05:33 am
The POET really addressed a lot of the issues that I personally have with things in our daily life. I too can identify with the guilt that the author is feeling that he cannot even operate something that is very simple. But it is really the design of the things that have the problem. For it does not show very obvisouly what is the purpose of the tool is.
FOr example, the mighty mouse is very slickly designed. But from the perspective of user-friendliness. It breaks all the rules. First, we have no idea where the button is. Also the scroll button is completely different from the conventional design. A new user will have to try to figure out how to use. The mighty mouse also come with two buttons, but can people tell that. Nope, there is no way to tell that.
If we look at the two different combination lock from above. I have the same problem with the combo lock on the left. I still have to read the instruction every time I use it. I found it to be unnecessarily uncomplex. The one on the right is much better and serve the same purpose
Yang Wang - Sep 18, 2006 11:30:36 am
Another comments I want to make is also about POET, because it is simply amazing and address all the things that I have concerns about.
First, visibility is the key. We want to make all the functionalities visible whenever possible.. For it increases complexities unnecessarily. I always like to keep a couple different devices with me. Because from the experience of playing with different gadgets, I realized that "one device deos everything" is simply a nightmare. Because it is either so hard to use that trying to call someone require 5 or six clicks just to get the dialing started. Otherwise, it falls into the problem of too much visiblity that I get overwhelmed by a million buttons.
Second, feedback from the user is very important. Because from the designer's perspective, everything just simply work together. But it is not until we observe how the user to work with it that we start to get a better idea how it works. Like the door the locked author's friend in. The designer will have no problem entering the building, and probably have no realization that anyone will. For the flash application I worked over the summer, I created a slider tools fthat control the time period. but it is not until that I sent it our for user testing that i realize people cannot figure what the slider does. Even for something that simple, it has to be obvious what it does. But from my point of view, I don't see a problem.
Heng Kuang - Sep 18, 2006 12:01:27 pm
The Psychopathology of Everyday Things – 1: Very similar to the examples given at the beginning of the chapter, it took 4 college students (2 Business major + 1 EECS + 1 Chemical engineer) over an hour to figure out how this rice cooker works.
At a family gathering over the weekend at my cousin’s place, two extremely experienced housewives couldn’t turn on the rice cooker after attempting for 10 minutes and ask their children for help (the four college students mentioned above). Countless attempts failed until someone came up with the brilliant idea of downloading the manual online and reading through it. It turned out that you need to press three buttons “prewashed” – “quickcook” – “cooking/keep warm” in exactly that order to get it started. What are all the other buttons for? I still have no idea. Many buttons are confusingly labeled. Some of them play music when pressed, some produce a “beep”, while some do nothing. BAD UI !!!
The Psychopathology of Everyday Things – 2:
It is true that there are 2000+ things that are designed badly in our everyday life, but why do we put up with it? To save oneself from embarrassment is one of the answers. I was on an AC transit bus the other day. A guy stood next to the rear door of the bus looking very confused. The bus stopped, he pushed the door, waited, pushed again, waited, nothing happened. The bus started to move again, “WAIT !” he yelled. The bus stopped again. “How do you open the door?” he finally asked, in a very low voice. His hand was covering the button, blocking himself from seeing it... Sometimes, we dare not ask when we have no clue because it seems that everyone else has it figured out. The door-opening-button in the bus is not particular badly designed in my opinion, but apparently it is not visible enough to everyone. It confuses people probably because the buttons next to the door look quite different from the buttons at other places in the bus.
Tangible multimodal interfaces for safety-critical applications – 1: Quoted from the article: “Rather than require that users change, system designers could adapt their systems to key aspects of the users’ work practice.” I have to say that this practice is not commonly seen in office software tools nowadays. It is very surprising how uncomfortable many people are computers. Story 1: One of my college professor who had a master degree from Cal types in Word, make diagrams in Excel, and hand cut and paste the two together using glue. Story 2: A director with a PhD from Harvard prints out all her important emails and archives them in a folder. We can not safely assume that computers are as intuitive to everyone as to all the CS/EECS majors here in the class.
Tangible multimodal interfaces for safety-critical applications – 2: The designers combine TUI and MMUI because the each have their own advantages and weaknesses, but when combined they are suitable for the task in hand. This is very interesting because when I try to solve a problem, I usually use the “top-down” approach. (i.e. How can I break this problem into several modules?) The approach taken by the designer of TMI is more “bottom-up”. They identify the objective, and then look at what current technologies currently available can but combined and modified to solve the problem. It is probably a more productive and realistic way of thinking.
Rory Martin - Sep 18, 2006 11:26:26 am
POET(1): A lot has changed since 1988; however, the basic ideas that are presented in this article still hold true. Objects that we use have natural affordances, which can be exploited in order to design a good user interface. I think that we have all run across a lobby door that seems like you should pull it even though you are really supposed to push it. The real problem is the way that doors vary, some doors have handles on both sides, some have a handle on only one side, some don't have any handles. Sometimes it is unclear as to what the actual properties of the door are. Is it that handles have an innate "pull" property, or is it that most doors follow one schema, handle on one side only. This reminds me of the Task Centered Design Process article that we read a few weeks ago, common interfaces need to follow conventions, users will attempt to interact with the device according to their past experience with similar devices.
POET(2): I feel that the best part of this article is when it focuses on the development of new devices. According to the author most products take five or six attempts to get the product right. Most products don't last long enough to become successful, even revolutionary products will be considered failed after the second or third attempt. This shows just how important the initial design cycle is, if the product is flawed from the get-go, then there is little to no chance that it will become successful, even after several iterations.
Cohen(1): I completely understand the need to keep paper backups of everything that a company does. At my old job we had to keep financial records of everything, in fact your required to keep all financial records for 7 years (i think). While financial records may not seem as important as the militaries war documentation, there can be very devastating effects if the information is lost. Computers are not nearly as reliable as they need to be in order to live in a paperless world. It is therefore important that we make these interfaces between paper and the digital world in order to attempt to bridge the gap.
Cohen(2): Cohen places a lot of emphasis for the failure of the paperless office on the physical characteristics of current technology, "these systems need far better resolution, much less weight, and better portability." I think that if you asked a doctor if they would rather carry around a tablet pc that worked with 100% reliability or a clipboard that they would choose the pc. The problem with current technology is that it is unreliable and can oftentimes lead to increased amounts of work that need to be done in order to workaround such limitations. Is the problem that the doctor might have to carry a 4lbs tablet pc? probably not. Is the problem that the pc has to low of resolution? I don't think so (have you ever seen a doctor's handwriting? very messy). It really boils down to the importance of the information that is being taken, the liability of that information being corrupted, and the reliability of the paper and pen model. While Cohen does list system crashes as another reason for why digital offices have failed, I think that his main focus is in the wrong place.
CharlesLeung - Sep 18, 2006 11:20:49 am
Cohen (1) - I thought it was very interesting how the author commented on the two types of people who adopt technology. Although I thought that I was someone who wishes to adopt new technology quickly, I realize that I am very much like the doctors or soldiers who are comfortable with what works now and hestitate to adopt new technology. In fact, I even think about some of my friends who on the surface should be very comfortable with new technology, but are very slow adopters of new gadgets and things.
Cohen (2) - I would agree with the author's conclusion that too many products demand customers to change their work habits in order to adopt new technology. That kind of design decision is a major deterrent for me personally to adopt new technology. Because of the critical nature of war and healthcare, it seems like the barrier to adopt new technology will be realtively high.
POET(1) - I liked the example of the German bus put forth in this paper. Even though some UI's look like they are very complicated and may seem on the surface to be poorly designed because they look complicated and imposing, those UI's could still be good as long as everything is placed in a logical and natural place. As long as things are done in a rational way, it would seem that the difficulty for a user would be greatly decreased.
POET(2) - I also like the point the author makes about having to go through a few different cycles in order to come up with a good design. If something is really going to be revolutionary, it's probably going to be true that it will be hard to come up with the great design right off the bat. The author also points out something very important in that if something is done poorly the first few times, then people will give up on that idea. This is important to point out because that means that designers need to be very careful when releasing products to customers, lest the give up on potentially good technology just because the early editions have bad designs.
Keenahn Jung - Dec 08, 2006 07:31:57 am
Cohen: The battle map seems like a great application of the anoto pen that is superior to using a tablet PC. The author made the point that the paper map degrades gracefully, that it is still largely functional even if it has a hole in it, whereas a tablet PC would be useless with a hole in it. The need for paper backups and familiar interfaces is where Anoto pens really shine. This comes up again for the NISChart application. The doctors can't be bothered with manual data entry, they need the interface to be as painless as writing. The Anoto pen provides this interface, as well as a redundant paper copy (the "definitive" copy) in case of computer failure.
Cohen: The conjunction of MMI's and TUI's provides very good ways of interacting with a system that were not previously possible. The Rasa system, in which officers draw units on post-its and annotate them with voice recordings, does a great job of this. If the computer fails, the collaboration is weakened (they can still communicate via other channels), but they can still update their paper maps using the post-its. Then when the computer recovers, the map updates. This is good because it allows the officers to manipulate physical objects, but also provides the advantages of digitizing them. Again, if the computer dies, they still have the paper map, which is still very useful.
Poet: While the author has trouble opening doors, and his friend appeared to be trapped between two rows of glass doors, I have rarely had this problem. Sure it is annoying, but the elegance and beauty of a completely unmarked, symmetric, glass door is worth the annoyance. The author seems to take the stance that functionalilty takes precedence over aesthetics. In a hotel, you can assume that the target user group will not be in a terrific rush that they require the utmost unambiguous functionality. In fact, it is probable that the aesthetics of the lobby doors in a hotel have a much greater impact on business than do the functionality. Ideally, everyday objects would have both, but everything is a tradeoff, and while the vast majority of UI's out there fail in some respect or another at being functional, I think it is OK to make the tradeoff in favor of aesthetics some of the time.
Poet: Making controls visible is again a tradeoff between aesthetics and functionality. The author prefers the car method, where there are approximately 100 different controls, most of which perform exactly one function. This is great for functionality, but can be pretty ugly. Take for example the ipod. I will admit that the UI is not the best, and that I have been confused by it before (I do not own an iPod, so I was using a friend's). For example, what does it mean to go "back" in each context? Will I go back one song? Back to the previous menu? Up a level in the folder hierarchy? However, the iPod is one of the most quickly adapted pieces of technology in the past decade, and indeed it is iconic of our generation. Would it have been as quickly picked up if it had 100 different, explicitly labelled buttons? Probably not. The "iPod effect" comes from when a person sees it's simplicity and elegance, even if this obscures functionality.
Robin Franco - Dec 15, 2006 10:26:07 am
POET 1: I think that the whole idea of “form over function” is true to a certain extent. Yet there are many other examples around us where there is a much better way to solve a problem, yet the designers stick with the old simply to satisfy users. For example, elevators. A much more efficient way of interfacing with an elevator would be to input your floor number right at the entrance of the elevator. This would allow elevators to create more intelligent stop patterns which would increase efficiently. However, this system is not used because users are too accustomed to the old method. Entering an elevator with no floor buttons becomes a slightly scary situation. This leads manufactures to stick with the old, even now.
POET 2: Its true that technology is becoming more and more complex, but I don't think its necessarily becoming more complicated. There are many good innovation that greatly add new features yet still keep the product simple. One good example is the car. In the old days, the car was very complicated and unreliable. To properly use one, one had to also be part mechanic. Nowadays, one could argue that car are much more complex with new features coming out every year, yet the care is easier to use.
Tangible Multimodal Interfaces 1: The reliability of “mission-critical” applications is something that has been in constant innovation. Anytime someone's safety or life hangs on a balance, its not something to be taken lightly. This reminds me of the famous Therac-25 bug which led to deaths of many patients. When people's lives are at stake, not only is the internal workings of a program important, but the UI as well. One has to assure that the operator has as little impedance in his or her tasks.
Tangible Multimodal Interfaces 2: But what about non “mission-critical” applications. I think they too should work on improving their reliability. If not for the improvement of that one technology, but also for the improvement of technology in general. If all applications and systems start moving toward 100% reliability, then we will see more widespread adoption and use from the general public. I still know many people who are afraid to use credit cards over the Internet simply because they feel it is an unreliable system. If you can break this barrier, many other advances can build upon these innovations.
DavidWallace - Dec 15, 2006 03:53:16 pm
POET 1: It is impossible to make an interface that everybody will understand the first time. Each person is coming from a different background and set of assumptions. While it's fun to complain about confusing interfaces, we should recognize that there's no escaping them completely, and that part of the responsibility lies with the user.
POET 2: Furthermore, there are times when style and appearance are more important than usability. The glass hotel doors in the article are a perfect example. The hotel probaby cares more about being stylish than about saving its guests three seconds the first time they enter. We shold recognize that usability is not always the top priority, and in those cases tradeoffs will be needed. My car's door handle doesn't say PULL HERE TO OPEN DOOR on it.
Cohen 1: Users continue to use their traditional paper interfaces for many reasons. As the paper argues, to replace them with digital alternatives does require matching the user's expected work styles -- but it requires much more. It requires matching the reliability and predictability of paper. I suspect that the technologies this paper relies upon (speech recognition, handwriting recognition...) are not yet mature enough to be reliable. Imagine the user's frustration when they have to write the same word three times for it to be recognized correctly.
Cohen 2: I suspect a better approach would be to make the digital aspect of the system more passive. For example, recordings and photos of the paper workspace could be made on a regular basis. This would prevent them from worrying during their work if any recognition algorithms were working properly, and the digital records would then serve as a backup rather than the primary store of information. Although the system would have fewer features than the one Cohen suggests, I think the users would be more likely to adopt it.
