Human Information Processing (Perception)

From CS160 User Interfaces Fa06

Jump to: navigation, search

Lecture on Oct 2, 2006

Slides

Contents

Readings

Ramy Ghabrial - Sep 30, 2006 12:05:17 am

This article is obviously a lot more technical than the others we have read thus far. While reading it, though, I could not help remembering Raskin's comment about cognetics, and people attempting to model the brain with digital computer structures: "I am leery of that approach, especially because throughout the ages, thinkers have borrowed the latest technology as models for understanding humans only to abandon or to limit the models when technology is advanced". With that in mind, the approach here, while interesting, seems extremely limited and narrow in scope. I see lots of numbers, definitions, formulas, caveats and other jargon that might be abandoned as soon as we can use more impressive technology as a metaphor for our thinking process, and the only thing I really took away from this reading was a sense of the empirical evidence and examples supporting the given model.

Tak Wong - Sep 30, 2006 03:47:03 am

I think the important points from this article are how fast we can process data into our brain and how to retrieve these information later. How fast we can process data can help us determine how fast to display information to the user if it changes automatically, like an electronic construction sign or an advertisement screen. Knowing that information is never really "forgotten", we should concentrate on giving information using words that is easy for the users can recall later. This means using words that the user can easily understand and can recall while doing everyday activities. It is important in, for example, designing an interface for an advertisement. I think the concepts in this article are really interesting, just that it got a bit too technical.

Tabassum Khan - Sep 30, 2006 12:37:06 pm

I concur with others that this reading is very different from the readings that we have had so far in this class; however, it shares some of its ideas with Raskin's reading about cognitive consious and cognitive unconcious.

The part of memory that fascinates me most is the long term memory which is a more permanent and reliable source of information. This long term memory greatly influences our perceptual system. Our prior knowledge and expectations affect how we percieve and interpret sensory information. This is the place that gives birth to biases and phobias.

Besides chunking, the long term memory can be enforced by repitition of tasks and by recognizing patterns and relations. How fast and how easily a chunk can be retrieved from long term memory depends on how it was encoded and stored there. Some of the many techniques that I have adopted to store information in my darling long term memory are flashcards, pneumonics, pictures, rewriting, and relating things to one another.

David Hoffman - Sep 30, 2006 01:07:56 pm

This reading covered some really interesting aspects to the logistics of human-computer interaction. If people and computers are interacting, they need to communicate on a similar frequency and similar bandwidth. This reading is actually a pretty good resource for determining what are correct frequencies and rates to use with people. It suggests certain types of fundamental limitations in different tasks such as in perception and also motor actions. This really sets some ground rules for what kind of interactions the user is capable of. For example, it is unrealistic to set a computer's sensitivity to want a double click duration to be less than than 50ms, because there is no way to press, realease and repress the button in less than 140ms, according to their button press data. They also cover some of the perceptual problems the brain faces where it can misinterpret a visual stimulus if the display is shown too quickly.

Patti Bao - Sep 30, 2006 03:23:59 pm

I thought the section on motor skill was very pertinent to HCI. Although Fitt's Law is probably something that many of us would immediately associate with clicking on targets while gaming, it clearly has implications for other user interfaces as well. For example, my friend's website likebetter would be much harder to use if you had to click on a small button for either photo instead of the photo itself. Nevertheless, it's also true that it gets easier with practice. For websites that you visit over and over again, you don't even need to read the navigation menu before clicking on the buttons or links, so perhaps it's even possible to hone your motor skills by repeated use of a certain clicking pattern. At any rate, here is a great way to test out Fitt's Law, courtesy of Ken Goldberg: http://www.tele-actor.net/fitts/.

Bowen Li - Sep 30, 2006 06:26:30 pm

I agree with Ramy, in that we don't really understand how the mind works and that using empirical data seems like the best approach at this time. However, as the designer, we shouldn't ignore human limitations. I think the Morse code anecdote is a good example. People's ability to transmit and receive Morse code wasn't limited by technological speed, but instead by human ability to interpret the data. By changing the technology to one that people can use better (beep lengths instead of counting blips), the performance level was able to increase. I think this type of shift will increase as technology advances. With processors running much faster than human ability, what will determine usability and speed is presenting the machine through an interface that people can react quickly to.

Maksim Lirov - Sep 30, 2006 09:30:26 pm

This reading by Card, Moran & Newell attempted to represent the psychology of the human mind in a very scientific form, represented by numerous equations. I can definitely see that this presentation can be useful to optimize a system or application in accordance with the human brain in situations like those presented in the reading - moving picture rate, morse code listening rate, and others. However, like other people have already noted, I thought that the authors tried to exactly define with equations the human mind - something that scientists do not understand fully.

I agree with David's post above in that the main lesson I took away from this reading is that we must take into consideration the rate at which our designed system is feeding information to the user. Laws such as the Power Law of Practice in particular seem helpful in this regard. Jef Raskin's Canon Cat seemed to use some of the techniques this paper presents. In the reading by Raskin, I remember in particular him talking about the Canon Cat quickly loading an image of the user's desktop from the disk and while the user is thinking about what she was doing, the rest of the data is loaded. Tricks like this should be used when available to optimize system response to human input.

Jonathan Yen - Sep 30, 2006 10:58:14 pm

This article really gives a lot of technical details about the human side of HCI. I think that a thorough analysis of cognitive psychology might make task completion quicker and more efficient, but it seems like there are a lot of variables to look at in the long run through this type of approach. To implement what we know from psychology into HCI would probably take a considerable amount of time, given that there are just so many factors to deliberate over and a lot of statistics to take into account. Furthermore, there is a qualitative aspect to take note of in HCI, which is not part of the model. I agree, however, that the model is useful with respect to human performance in HCI.

Jason Shangkuan - Oct 01, 2006 01:12:00 pm

This article tries to establish a couple main points, but what I found most intruiging were using a metaphor for the human brain by symbolizing it as a computer and also applying abstracting concepts that apply to computers to studying the human mind. As mentioned in a previous post, this article truly extends the study of cognition, and as a result, the article can be utilized in two ways: 1) designer can use this information to scientifically target a group to design an interface. For example, by following the chart on visual deterioration, a designer can decide how large to make fonts or pictures. 2) this article can show how it might be a poor design principle to follow the scientific methods and charts. By following the trends, a designer is taking a general and simplified view of humans and what might be expected. This goes against the principle of taking specific personas and scenarios to design to. There has to be a balance between taking the scientific charts and specific personas to make a proper design.

Ming Huang - Oct 01, 2006 03:12:42 pm

The chapter on “The Model Human Processor” does a great deal to generalize and quantify the steps and processes humans undergo when they are receiving, interpreting, and reacting to information from the outside world. Although I wouldn’t go so far as to pin point reaction times and plot trends, I do appreciate these researchers back in the 80s doing the work for us so that we know much better now.

One aspect of the Motor System that the book seemingly overlooked, I think, was the memory of the Motor System. Does it have no memory at all? I frequently hear the term “muscle memory” referring to habitual actions that can be carried out without conscious thinking from the user. It must be true that such actions are remembered or they cannot be replayed at will. Perhaps this memory resides with the Cognitive System, being a part its long term memory, so it was not associated with the Motor System.

The book’s interesting assertion of infinite capacity of the long-term memory begs my question of its reasoning. One can argue that just because someone cannot remember something does not mean it is not in his memory. But the opposite argument is also sensible, because the definition of existence depends partly on the ability to recall. Since there is no reliable way to verify whether someone knows something when he cannot recall from his memory, it is debatable whether we still have any recollection of childhood events not otherwise recorded (on photos or other media).

There seems to be a lot of computer science lingo describing the human process of information storage and retrieval, some aspects of how the human brain works differs from that of computers. The human long-term memory may last longer than data on a computer hard disk, but it decays faster and over time become inaccessible. The data on a computers’ storage is generally available on demand as long as the storage media is intact. The Model Human Processor is only an approximation to the real thing to facilitate our understanding of its operations, it is not perfect, but it works.

Michael Moeng - Oct 01, 2006 04:25:05 pm

This is easily the most technical paper analyzing cognition I've ever encountered. It's pretty impressive how the author manages to not only analyze the numbers, but also shows how these results can be applied. For example, the section on causality, where an event following another event by less than 50ms is perceived as an effect of the first event, but if they are separated by more time than the events might be perceived as being independant. This puts bounds on how much processing is allowable for an interface (a click needs to show some kind of response within the 50ms, or the user will believe something is wrong).

The section on practice -> faster tasks seems especially applicable to computer games, where a game designer can set thresholds for certain tasks: if a user has played X times, they can probably do taskA, 2X-> taskB and so on.

Robert Taylor - Oct 01, 2006 05:27:39 pm

I found this reading interesting but a little too advanced for what we will be doing. Perception rates, Fitt's law, and other measurements are likely too precise for what we are building. Also, the way the author gives treatment to solving his problems using his equations believes one to think his numbers are very exact. Obviously I am no cognitive science researcher, but I feel like people will be different enough and have different enough reactions to not be able to bed describable by a set of constants in a general equation. I suppose they can form good guidelines, though..

Alex Wallisch - Oct 01, 2006 07:17:14 pm

I agree with the general consensus that this article was extremely technical, potentially too much so. Attempting to quantify perception, cognition, and other aspects of the mind seems like a good idea, except that I'm not sure how much science really understands about how the brain works. I'm impressed with the thoroughness in gathering empirical data regarding how humans react to stimuli and how long objects can exist in memory, but isn't it a bit too convenient that human perception - which has existed for eons - should be best described by modeling it on a computer processor - which just happens to be the technology of our current generation?

Patrick Rodriguez - Oct 01, 2006 08:15:48 pm

I, too, found this reading over-technical for the purposes of this class. That's not to say that it wasn't interesting. It's one thing to give a theory about why things are the way they are; it's another thing to provide some raw numbers. The numbers do explain a lot, from why it's hard to remember a long series of items to why it's possible to view a sequence of images as fluid motion. Researchers should definitely continue to research the capacities of the human mind, and maybe even look for ways to expand these capacities. As UI designers, we must always keep in mind the current limitations of our users. Having this numerical information available can help in both the design and testing phases by allowing us to make sure the limits of human cognition are not being overstressed.

Chen Chang - Oct 01, 2006 08:17:52 pm

The Model Human Processor: can be described by processor and memory along with the set of principles of operation. It is divided into three parts known as the perceptual system which consists of sensors and buffer memories, the motor system which carry out responses, and the cognitive system which codes information symbolically and serves as a connection point from the inputs of the perceptual system to the outputs of the motor system. I liked the example presented relating typing, reading, and simultaneous translation to require parallel operation all three systems. Its like the relay of information across three pipelined processors consisting of different stages and then synchronization between all. The example given of the perceptual system is the human eye and how the retina captures information which is the term we deem "vision", then there is the way we perceive moving picture rates which relates to video playback and computer gaming frame rates which either result in smooth playback or lag and flicker. The motor system can be seen with the arm-hand-finger system analogy as the way humans respond by moving their fingers to type, moving their hand to hold something, moving their arm to do a tennis smash. I see the two types of memory and think of them as working memory being like ram with immediate access, whereas long-term memory is like a hard drive and can be stored over a period of time and accessed at a later time (although ironically magnetic hard drives to tend to die over time and human memory also fades over the years). The reading mentioned human reading rates as the limiting factor is not eye movement but rather in cognitive processing, I could not agree more so because the truth is that speed readers do skim to catch only the important pieces and get a summary of ideas rather than the intricate details. A few very interesting examples presented included the spacing of a scientific calculator function key to support more operations as well as the alphabetic keyboard versus standard QWERTY keyboard. I found the results of these examples particularily remarkable as the qwerty keyboard that the whole world is used to only gets an 8% advantage in speed for words per minute as the better arrangement but its by no means a major difference.

Hiroki Terashima - Oct 01, 2006 09:18:14 pm

The first part of this paper describes a model of human mind, saying that concepts like "interconnected memories and processors" and "set of principles of operation" help us "understand, predict, and even to calculate human performance relevant to human-computer interaction". I think this is another good example of how a metaphor is useful in understanding how our world works. By creating a model that we can use to relate to, we can talk about concepts in a way that many people can understand.

Most of the other stuff were complicated analysis of things using graphs and equations that I didn't understand (I tried to understand for a while...). And yes, as people have mentioned, I too believe that these kinds of studies are important in understanding HCI because it adds a scientific, perhaps more objective perspective into the subject.

Yimin Yao - Oct 01, 2006 08:41:01 pm

The article provides a very rational but extremely simplified model of how information flow from perception, then cognition to motor output inside of our brains. The technical data (the numerical analysis of the theoretical calculations and experimental results) is very overwhelming and painstaking to read through, but it does provide some good basis for the model of Human Information Processing.

While mechanics of how the biological brain really works remains as a great mystery due to its extreme complexity, I think as far as this class is concerned, which is to show the limitation in abiliy to process information by human, the articles does a very good job in convincing us that in designing user interface, we can not expect users to work beyond the boundary of their minds.

Melissa Jiang - Oct 01, 2006 09:48:52 pm

Also agree with the general consensus regarding the technicality of this article. Probably the most technical explanation of long term and short term memory I have ever come across. I do also wonder whether the technicalities were necessary. However, the article did have some good points. The article pointed out that many memories are stored within long term memory already. However, storing memory takes a lot longer than retrieving the memory. Therefore, instead of having the user learn and rehearse new commands/function in order to store it in their long term memory, the designer can just design the command/function that will draw on their long term memory. Playing on old function commands (ctrl+some key for example) would ease the learning curve greatly. This idea goes along the same line as the first article stating that generic functions should be including just so users would have familiar and help in the sale of the product.

Roland Carlos - Oct 01, 2006 10:30:21 pm

Is it possible to accurately describe the way the mind works with numbers and formulas? This reading definitely came close. While it was a hard read for me to get around at first, the more I stuck with it, the better it sunk in and actually became more understandable. While I'm hoping we won't need to get exactly precise measurements on reaction time for our users for this project, it's good again to understand the limits that we work with in design. It's too easy to forget when you're working around computers all day. It's almost scary how computer-like we can make people as well.

People are seemingly limited to a great extent by their memories. But then again, it makes sense. It's interesting the game that people play with keeping things in their long term memory and then trying to recall it from those depths. At what point does something become instinct to us? How many reptitions is necessary for people to learn something before they don't need to recall that memory on how to do it anymore?

Tony Yu Tung Lai - Oct 01, 2006 10:11:53 pm

Although this reading is one of the more (and may be overly) technical one, I find this reading to be very interesting. It is important for us, as designers, to learn about the capability of human cognitions, since we will be mostly dealing with humans as users. The part that I like most about this reading is the section about memory retrival. I did not know that we store materials in working memory acoustically. May be that's the reason to the 'song stuck in head' phenonmenon.

Andrew Hao - Oct 01, 2006 11:40:50 pm

Incredible! I didn't realize (but probably should have) that the cognitive sciences were so technical. There must have been much progress and study and advancement with regards to the human brain (interestingly seen as a machine, a "processor"). I am particularly impressed with the way the cognitive scientists are able to link the motor systems to the cognitive processing systems to the memory systems to give us a fuller, deeper (and admittedly, confusing) exploration of the human mind. HCI research directly benefits from this sort of study -- by understanding how the human mind processes information, designers can better engineer interfaces that best suit human tendencies. Just don't ask me how (yet).

Vahe Oughourlian - Oct 01, 2006 10:52:27 pm

There may have been a caveat in there somewhere that I missed, but attempting to quantify the portions of the human brain and the portions of the body is a poor way to teach a designer how to take these features into account. The great pitfall is that designers, being somewhat digital themselves, may take these numbers only at face value, and use them as the end-all, be-all for all their UI interfaces. It is mentioned halfway through that tolerances (a low, a medium, and a high preset should be selected), but little impression is given in the first section that this should be a concern (a very great concern, in my opinion). Physological studies and psycological studies, despite what this paper purports, are not as advanced as the authors want these fields to be. The human brain is not this exact, and even if it were, the variance among users of different skill levels is not quite taken into account. Raskin mentions people falling in love with their metaphors; this paper is only the embodiment of that warning. This paper does have some merits, however. It attempts to condense a large field of study down to the pertinant information that user interface designers may wish to use. However, the language of the paper would make many a designer throw up his or her hands in frustration. Maybe they should have designed their paper using some of their own precepts, eh?

Andrew Tran - Oct 02, 2006 12:32:48 am

Why do we need to know so much about the human mind to design interfaces? Do i really need to know the the speed of the half life of the visual image store is about 200[90~1000]msec? I don't think i really need to know these specific details of the human brain when designing interfaces. However, i do admit some of the things presented within this chapter is quite interesting. I never really knew how memory works. This chapter states people remember things in chunks and there is a spreading activation affect when we recall a chunk of memory from long term memory. I definitely agree with this because this is a topic in cognitive science classes also. In general i believe this article belongs more to a cognitive science class. I think testing your software with users should be enough to make the interface better, not knowing the speed of our brain would really help. If it is too fast for designers to use, then it should already be fast for users.

Kang Chen - Oct 02, 2006 12:44:14 am

First of all, I like to comment what many have already pointed out. This article seems way too technical for what it is describing. Half life? Decay time? Maybe it's true but having a whole bunch of formulas and exact numbers just made it too precise. For one thing, us, the designers, will probably never try to time every transition in our interface just to please the user. That's more for the folks at Hollywood to worry about. In my opinion, the numbers, formulas, and graphs did provide more of an objective measure of human capabilities but is overly precise.

Aside from the overly technical details, there were some interesting concepts brought up by the article. I particularly liked the moving picture rate section where it mentioned the interval in which an image or video must be refreshed to give the illusion of movement. By using the model described, an approximate frame rate was calculated to achieve this task. I am no graphics expert but it seems such a concept would be very helpful in stages of optimization and compression. Further research in this field would be very interesting.

Cheng-Lun Yang - Oct 02, 2006 01:09:47 am

All the statistics presented in the book are not there to overwhelm the readers. They are there to show us that the trend of modeling human mind with current technology. We see from previous reading that the author of that chapter predicts human brain modeled by computer technology. The human mental and physical internal interactions and external interactions with the world are quantified for us to study. Human memory resembles computer memory. So does computer memory created modeling human memory or human memory is modeled by computer memory after its creation? I think the engineers created the computer system with the human system in mind. From the memory to processor to hardwares, the computer system closely resembles the human system.

Antonis Mannaris - Oct 02, 2006 01:48:51 am

I agree with previous postings that extreme details like the cycle times of the three processors of the brain or of the memory are extreneous, and should probably be ignored by designer of user interfaces. Indeed, taking these into account is more likely to take up your precious time and the improvement of your final design will not be significant. Nevertheless, I really enjoyed the article for one main reason. It is a concrete theory that supports claims we saw in previous readings about user interface design. For example, the fact the a user is more likely to remember a piece of information if it is familiar to her can be explained by the paper's theory that data are stored into memory in chunks, and chunks that can be related to already existing memory data are easier to remember (higher probability of moving from working memory into long ter memory). Also, the fact that it takes much more time to record something into memory than to retrieve it, suggests that it is not a good idea to throw out data to the user at a fast rate and expect that she will be able to remember it. Even some of the paper's extreme details hide useful truths. For example, the fact that time increases with distance (in terms of mouse movement). A good user interface should try to be consistent on the location of its option buttons. Also, it is mentioned that the more choices a user has, it takes exponentially longer time to process and decide on an action. In other words, it makes more sense to try to separate mutually exclusive choices and present them to the user one by one.

Randy Hilarbo - Oct 02, 2006 01:44:03 am

It's interesting how the article presented human cognition in terms of a "set of memories and processors." The way the perceptual, cognitive and the motor systems relays information can be a useful model in designing a system. Understanding the concepts presented can also be a good tool in designing a good interface. Designers can present information in chunks so this can easily be remember. They can design features so that these make use of the "power law of practice" concept. Or they can decide an interface so that important task could be readily moved to long-term memory. Speaking of long-term memory, I often wonder why I can't readily write codes using a language I've learned and used heavily before. The discussion in "interference in long-term memory" explained why.

Yang Wang - Oct 02, 2006 02:12:12 am

The article expressed in great detail how the human brain works very much like a computer. Although it tends to overly exaggerate with technical term, but I can see where the author is trying to take with his metaphor. I also agree that his description is very accurate, as I know for a fact that is how I perceive things. However, I can't help but remembering "The Human Interface" chapter 2 from a week ago. As the author from that book points out. Such metaphor helps us in understand, but it undoubtly has its limitations, and it is almost doom to be thrown away in just a couple year, as science progresses and technology advances. The brain is used to be described as a mechanical system; but that idea was thrown away due to the limitations discovered in mechanics. There are serious doubt that how long such a metaphor with computers will last.

Rayhan Lal - Oct 02, 2006 03:26:40 am

I was blown away by how accurate these measures compared with observed values. The study is extraordinarily important because it establishes upper and lower limits for man’s interaction with computers. The author doesn’t purport his measures are anything more than a good rule of thumb. It is much easier to express how two machines interact than how a complex non-deterministic biological system interacts with a machine. I would argue with anyone who claims this article is "too technical" for the class, it contains the most basic parameters for designing an interface. From displaying a smooth animation to providing an efficient input device, these values are essential for two-way communication between computer and user. We cannot design an interface without knowing both the limits of the computer and those of the operator.

Heung Tai - Oct 02, 2006 05:27:01 am

This reading is too technical and doesn't draw a good summary or conclusion. My feeling from this article is that human limitation is hard to measure but still should be under consideration of design. The human can only process information in some specific finite time in addition to the memory limitation of human being, so the display of information from interface should not be like a bomb that throw all information in a screen without giving user a break. Different presentation of information is also critical because it affects human's process time, e.g. analog and digital clock.

Qingyun Tang - Oct 02, 2006 04:17:54 am

The reading reminds me a lot about Raskin's idea of cognitive consciousness and unconsciousness. However, this article is much more technical and harder to understand than Raskin’s. It states that our brain is a fast-read, slow-write system by doing various of professional tests. I was surprised how accurate they are observing and measuring the values. In concerning of the class, what we should learn from this article is that human minds are limited; the user interface we design should not go beyond the ability of human mind.

Huangnankun - Oct 02, 2006 08:38:26 am

This article talks about how the human brain processes information. It attempts to break down the brain’s functions into major systems such as the “cognitive process” and the memory subsystem, the author then talks about the characteristics of these systems in information processing. This information is all extremely useful in interface design.

For example, the half time of letters as images in the user’s memory is one factor made used of by many of interfaces today. One example would be the filter effects in early versions of adobe Photoshop. These effects are applied over the image but often has a “source” parameter which the user can interactively specify by dragging a point unto a proxy of the actual image, however in order to speed up processing, the proxy does not display the actual image. Therefore this mechanism depends on the user’s memory of the image in order for him/her to manipulate the filter effect. In this case, the filter dialogue box must pop up fast enough so that the user has not already forgotten about the image to manipulate it properly.

The cognitive cycle talks about the how our brain realizes process and direct our reactions to external information. The time it takes for this cycle to complete is fairly short. This time is important especially for real time applications, because if the application’s response is significantly longer than the time it takes for the brain to process information, the perception of real-time would be lost. This can be seen in networked first-person-shooter games where “lag” or “latency” can greatly ruin the gaming experience.

Robert Held - Oct 02, 2006 09:25:53 am

After I read the article, I felt the most pertinent point with respect to CS 160 was the speed at which short-term memory fades. When designing an interface, the information can be used to ensure that one never expects the user to perform tasks that are unnecessarily challenging or require feats of memorization. An interface should be built with consideration to the human psychology briefly outlined by the authors. Ideally, the result would be a system that requires minimal effort or frustration on the part of the user, without sacrificing functionality or applicability.

Edward Karuna - Oct 02, 2006 09:54:20 am

What I found interesting was the demonstration that human memory works in a very associative manner, thus creating the problem that when there are many similarities between different records, there is a disproportionally greater chance of error (87%->98%). One would expect an associative-oriented memory structure to be more able to recall many similar records faster and more accurately than a list of very different records. Also interesting is the property that since our memory is associative, learning more similar blocks will cause a greater chance of corruption in older similar blocks, since the two will conflict.

Eric Yoon - Oct 02, 2006 10:19:38 am

I agree with others that it is an interesting article. I have thought that one of the most fascinating aspects of an EECS education is not just to understand how a computer works; but, also, by comparison, to appreciate how the "human computer" performs certain calculations infinitely faster and more fluidly, despite all of our leaps in computing technology. This study takes a stab at analyzing and quantifying some strenths and weaknesses of the "human computer."

On the other hand, I fail to see how much of this data is practically useful to a UI designer. For example, it details, among other things, how fast a person can type, or how easily someone can retrieve a fact from long term memory. But user interface design isn't really concerned with humankind's maximu or minimum capabilities. It is more concerned with what is delightful, clear, convenient or comfortable. All of these numbers and charts only hint at that, but nothing more.

Scott Friedheim - Oct 02, 2006 10:40:31 am

This article was interesting to me because I really didn't know about the level of detail cognitive processing had been examined. I appreciated the analogy of the human computer to a computer engineering prospective of computer hardware. The analogy used frequently throughout the article was the reference to the cycle time(relating to a CPU's clock cycle time). How interesting that a human mind fuses events (images, sounds) that occur within the cycle time; also how cognitive processes need at least a cycle time to compute.

In trying to disprove the idea that there is a cycle time that must be obeyed I went searching on the net for a reflex tester and found one at: http://www.happyhub.com/network/reflex/ Quite disappointingly I did not do as well as I had expected to do. I thought that my skills as a PC gamer would come in handy at this but at best I could only score a 0.234 best and averaged: 0.281. Maybe there is some truth to this cycle time business after all.

Joe Hart - Oct 02, 2006 11:13:39 am

Cognitive processing is vitally important to designing a user interface that is useable by humans. By understanding the various strengths and limitations of the human perceptual system a UI designer can exploit and/or avoid these features to assist the user in achieving their goals. Contrary to what others are saying, I think the article did not go far enough in its understanding of mind. The human brain can be compared to a computer but there are dangers in thinking so. The massive parallel processing, fuzzy boundaries, and cognitive abilities of the human brain should not be fenced in by the analogy of the computer as a mind. Important understandings can be missed and left unused in UI design.

Johnathan Hawley - Oct 02, 2006 11:25:35 am

I think I have had some first hand experiences with 'interference in working memory' since I have been here at Berkeley. It states, "According to the Discrimination Principle, it is more difficult to recall an item if there are other similar items in memory." When the semester first begins, I can remember pretty well what has been taught. As the semester rolls on everything starts to jumble together and I have a hard time recalling anything. I can see how this principle can be applied to human computer interaction and user interface, especially if the interface is really extensive. When we design our applications we should make sure to keep tasks, procedures, buttons, whatever as individual and unique as possible. Otherwise we run the risk of our entire interface meshing together in the user's mind and causing a lot of unnecessary confusion.

Suneet Shah - Oct 02, 2006 11:20:41 am

This article was definitely an interesting read. It is interesting to try and quantify human reasoning and logic. I think the approach they take to do so is a great one because it allows them to show the similarities between the two and analyze how to create better interaction between humans and computers. The article really brings up some great examples: morse code, frame rates of videos, etc. These things were designed with human cognition in mind. Though the article was very technical, it illustrated some of these design considerations one may potentially have to make. I think the general idea that humans and computers will interact well when the flow of information is at a compatible speed for humans was the main take away from the article. Another main point of the article is that by understanding the strengths and weaknesses of the human cognition system, as well as the perceptive power of humans, better and sleeker interfaces can be designed. That being said, it may not be necessary to analyze them to such a detail, but to keep in mind general trends.

Jae Chang - Oct 02, 2006 11:36:39 am

The article is very impressive and amazing. By dividing the Human Information Processor into three interacting systems – the perceptual system, the motor system, the cognitive system, the author describes human mind in computer engineering and psychological terms. The article may seem too technical because of the accurate measurement in the experiment. However, by showing the result of the calculation of human’s physical performance, the author emphasizes that the user interface should be designed very carefully reflecting the human performance. One of the interesting parts in the article was that human long-term memory that can be stored longer than computer hard drive, but human memory decays faster than computer hard drive so that the stored data will be inaccessible eventually. By understanding the fact, engineers may design user interface very intuitive not based on human long-term memory but based on human perception. The article let us to have a new paradigm and approach to the user interface design.

Kimberly Lau - Oct 02, 2006 11:34:56 am

So this is where the picture on the main page comes from!

The first thing that occurred to me in reading this article was how technical its analyzation of our cognition is. I never realized (or thought about) how numerical a breakdown of our perceptions could be. However, these calculations can be helpful for our projects, as it gives us an idea of what thresholds to put our products in for optimal use. For instance, Fitt's Law defines hand movements and the general pace or preferences for clicking, etc, which we can try to manipulate in our systems. Memory storage and its recall are also key for helping users easily navigate the system. It is almost like an affordance, where you want to use a person's inherent associations to a new system's benefit.

Lastly, I thought the reference on acoustical codes was interesting because I often catch myself checking phonenumbers on its tone, rather than with the actual numbers. I never realized I was using acoustical codes.

Bryce Lee - Oct 02, 2006 11:51:40 am

The actual details of the argument made by this chapter seems unconvincing. The ranges given for each of the functions created (justified by the slowman, fastman, middleman arguemnts) are so large that it would seem that a variety of different baseless formulas could justify them. In addition, the reading tries to cast the human mind into a simplier abstraction of a computer processor, something that we have designed. However, the basic motivation is a very important, since limitations to human ability must be recognized. Too often, interfaces are designed to maximize productivity without consideration of how feasible it is to actually accomplish the task. Also, several severe assumptions such as flawless memory leads to further fustrations in production products.

Michael Mai - Oct 02, 2006 11:35:43 am

The Human Information-Processor As neuroscience continues to advance, the depth of knowledge about threshold stimuli and rates of data processing and information storing make the brain more and more comparable to a computer. As a student, the most relevant section of the reading is the Learning and Retrieval section on page 27. As I am typing this comment out, most of the information I have read over the weekend on this article has probably left my brain and is lost until I attempt to reread it again before the midterm for storage. I feel that the Learning section, although very basic in its explanations, can teach students how to study and learn better. It explains the need for verbal mnemonics and how important more associations are for better memorization and recall. This article, although very technical, has a lot of information that students can apply to their lives. And in terms of our UI class, memorization of interfaces is a key aspect of why and how users will take in our designs.

Yen Pai - Oct 02, 2006 12:21:43 pm

Dovetailing on the comments of others, it appears that UI designers should be concerned with efficient use of working memory, building or taking advantage of learned habits (as Raskin suggested), and perhaps with creating opportunities for inserts into long term memory.

I found it interesting that cognitive scientists borrow metaphors from the computer world to describe the human mind. Throughout history, an interchange of terms and metaphors has occurred between the physical sciences and the biological/behavior sciences. The structure and tone of this article read much like Descartes' Treatise of Man.

Leo Chen - Oct 02, 2006 12:25:51 pm

I believe that all this article is trying to point out is how closely the human mind mirrors a computer or vice versa. There is a lot of really really technical stuff that I don't believe is needed to get the point across. It's kind of like reading about the chemical properties of gasoline when all you need to do is start up a car. Although the information in this article could be useful for such things as font size, color contrast in HCL, it could have been presented in a much more accesible form.

Sung Yi - Oct 02, 2006 12:31:44 pm

This article is very technical but this helps us to think about how to modify the user interface design to adopt to the human properties, and thus improve human performance when using the UI. Understanding the aspects of how human brain works is very important in designing UI. For example, the author mentions about how objects sharing similar properties in memory lead to a corruption. Thus, making a shortcut or buttons very intuitive will certainly help the users to greatly improve their performance.

Eric Vacca - Oct 02, 2006 12:25:30 pm

Although modeling fits in with empiricial research, its seperate and quite different. This article treats human processes as a "black box" and works to fit mathematical descriptions to variables different response and cognitive variables. It is a interesting construct, and lends some insight into how we process information, and should serve is a good guideline in UID. Although not all human processes can be simplified to fit a computer process, finding quantities that human and computer processes share (responce time, etc), allows us to develope UIs that keep the human processes in mind. The most relevant section to us seems to be the key stroke section (for obvious reasons) and the causality section because of their direct relatedness to user interfaces.

Julius Cheng - Oct 02, 2006 12:25:22 pm

Hooray for Cognitive Science! As probably the only person who's majoring in it in this class, I very much enjoyed seeing the kinds of things I've studied being represented and put to good use. This reading is indeed the most technical reading we've had so far, though the technical details aren't really true. The formulas were there just to make the accompanying psychological theory more digestable, and I doubt that they were meant to realistically model cognitive behavior. The principles that the reading gave (P1, P2..etc.) are well-established theories derived from countless psychological experiments in the past.

That is my response to the skeptics and those that thought the reading was baloney - In essence, this is really a paper on psychology, and all the formulas were there to make the psychological principles more palatable to the technically-inclined reader. A technical model does not at all insist that its formulas and rules are cold hard fact, but instead gives the designer a more concrete explanation of what the human mind does so he or she can construct better interfaces for it.

Sean Carr - Oct 02, 2006 12:22:51 pm

The Model Human Processor: I think this article would have made more sense if I had more of a technical background or had done some more work with Cognitive Science and Perception. I could have used extensive "basic" introductions before getting the specific technical "facts." It is good to attempt to specify these ambiguous or unclear functions of the human brain. The memory decay part of cognition seems very applicable to HCI and I think it is one of the commonly overlooked aspects of UI design. Since most designers have a very good image of the entire system in their head they forget that the user may not remember what was shown to them 5 screens ago.

Also, the right side of some pages is cut off.

Siu Pang Chu - Oct 02, 2006 12:14:59 pm

I agree that this reading is too technical. It gives lots of equeations that seems not so useful for me. The Model Human Processor divided our human information processing system into three subsystems. 1. the perceptual system is first interaction wiht physical world. That was the signal that being produced by the integrated sensory system. These perceptual memories "hold information coded physically". 2 the motor system is execution system that translate the will into actions and send this singal to the vountary muscles. The cognitive system connect the perceptual system and motor system, like the cable in computer. In the article, it also mention some of the human performance examples, like moving picutre rate, morse code listening rate. It is interesting to see what are parameters in these equations, so that we can design a UI that let the user to make a faster respond.

Tom McClure - Oct 02, 2006 12:49:59 pm

Running down the abilities of the average human as though he/she were a lab rat seems a bit cold and unfeeling. Nevertheless, once you get past that it definitely does seem like a key area of study -- so much of the focus of software design is invested in the computer side of things, we almost never question our assumptions about the human side, all too often falling into the trap of design for ourselves rather than for our target audience. The typical designer does not have this level of in-depth understanding about his target in mind most of the time. He should! So many bad designs could be avoided simply by acknowledging the limits of the cognitive system, the ways that short term and long term memory can fail your target user, etc.

Of course, in following the design process, you end up making changes related to these shortcomings without even knowing specifically about the inner function of why the shortcomings exist. That is, you discover problems and you find ways to make things work better for your target users simply by involving them in the process. A later analysis might reveal why this or that design really failed (related directly or indirectly to the human abilities) but that level of detail can really be skipped for high-level design due to the beauty of the design principles.

Dexter Lau - Oct 02, 2006 12:58:34 pm

Psych2 covers all the same material. People are essentially computers with a cpu, RAM, and a hard drive. In this case they divide the processors to work within the working/shot-term RAM, describing them as motor, cognitive, perceptual. Much of it is common sense, such as remembering an array of characters (visual or audio) becomes harder as the arrays get longer. The interesting part is when people "chunk" their thoughts. Instead of remembering 1492 as a number 1-4-9-2, thinking of it as the year Columbus took a trip is a lot easier. More over the discussion about keyboards is interesting. I thought that they should include the Dvorak keyboard as well... The discussion on repetition is important, especially if the Sholes keyset is standard (might skew results a bit?).

Charles Lee - Oct 02, 2006 10:22:42 pm

Comment 1: In the same way that a good programmer can optimize a bit of code to work best on a machine when he knows the specifications of the machine, a good designer can optimize his UI to make best use of the machine that will process this display of information: the human user. This way, a designer can actually make features naturally available to the user, as opposed to a struggle to learn another new thing.

Comment 2: Oftentimes, the decision to allocate screen real estate is based on a "look and feel." However, the existance of tested variables and quantitative analyses about human users allows for a more discrete and reliable decision-making process. It is much more authoritative to say that the average user can hold 7 things in memory, and so the 8th thing needed for the app should remain on the screen, than to say "things are getting a bit much." Information like pixel distances and movement time required to move to a position can be used with information like attention span of the user to ensure that the user can complete tasks before he gets bored.

Utsav Shah - Oct 08, 2006 06:33:48 pm

I’ve never read a more detailed paper on human mind and performance through the perspectives of cognitive science and psychology. Descartes once said that human body and mind are separate substances. One might argue that we don’t need this much detailing to design an interface. It depends on what kind of an interface one is designing, for example, while designing a website one needs to catch user’s attention immediately and that can be achieved by having attractive and relative images. But one needs to consider the perceptual system, motor system and cognitive system when designing an interactive interface.

The paper brings up an interesting point about the working memory and long-term memory. It seems obvious but it’s important to understand that working memory is made up of active elements of long term memory, thus a subset of long term memory. This is vital for designing a user-oriented interface. Another thing it talks about how human mind recalls things in chunks. Some good examples are given in the reading, another good example is remembering a phone number. It makes it very easy to remember in chunks of 3 digits than a full 10-digit together.

David Eitan Poll - Oct 09, 2006 01:28:08 am

I know it's been said about 100 times in these comments, but there's one thing about this paper that I find disconserting. I don't think the human mind can be narrowed down to a few simple specifications. While I believe very strongly that the mind can be modeled as a highly complex computer, I think this paper vastly oversimplifies what is going on. To me, it's like saying that a modern computer is as simple as a 1 and a 0. Yes, it might be true, and it's a useful simplification in many situations (anything that can be modeled as ones and zeros can be modeled by a computer), but it doesn't even begin to give us the information we need to make informed decisions as developers.

All that said, perhaps the model presented can be enough to allow us to make a variety of useful user interfaces. It is conceivable to me that by modeling the way that people process inputs in a simpler way, such as like a computer, we can be more effective in our interface design. Usability is an important factor, and if we can gain some insight into what makes interfaces more usable by modeling the brain in this way, I'm all for it. I just worry that modeling the mind in this way may keep us from inventing even more effective interfaces because we've oversimplified what the mind is capable of or what's at work.

Simon Tan - Oct 08, 2006 11:12:47 pm

Comment 1: The assumptions made by this article generally centered on one fact that seemed indisputable as I read along, and was confirmed in the closing pages: the notion that human beings have physical limitations in their senses which cannot be transcended. I’m not entirely sure this is true; that is, I’m not sure that we can assume that human beings are forever confined to particular numerical values quantifying their mental processing ability. To me, it is kind of like how the average IQ of a human being goes up over time, and thus the designers of the test have to re-scale the scoring. I think our senses are still evolving; they’re not perfect, and they should get better if the laws of natural selection even apply to us anymore.

Comment 2: Like may others, I noticed a that there was a lot of cognitive science and psychology in this text; I recognize some of the facts and numbers (like the fact that humans tend to memorize things better in chunks, only 7 digits at a time, etc.), and it was interesting to see how computer scientists used these facts to create such a complex, detailed model of one’s brain. Although I felt the facts were correctly researched, I had real doubts that an accurate model was built completely based off of laws and corollaries that stemmed from those facts - I would imagine that the human brain is far too complex and too variable to be described so simply and systematically. Some of the pages in the text look like they could have come out of a math textbook - does the human brain really follow such stringent structure?

Aleksandr (Sasha) Ashpis - Oct 09, 2006 10:47:53 am

1) The simplification of the human brain to “three processors” is in my opinion a bit of an over simplification, but I guess is good enough for a basic raw model to follow, until reseatchers can prove otherwise. I also don’t agree that our “three processors”, the perceptual system, the motor, system, and the cognitive system can independently run in parallel, because without expanding, sometimes we cant fully run our motor system properly, e.g rubbing your stomach and patting your head. Most people can’t do that, does that mean they need an upgrade in processors? And if so, what does that mean, how does a person upgrade?

2) The article claims the frame rate necessary for steady motion is 20 frames per second, then why do the movies use a 24 frames per second camera, why waste the extra film? In addition, I have long heard that if you add a 25 frame per second, that is where subliminal messages come into play. So, two questions: one, why does this extra frame, that we apparently cant consciously see, affect our unconscious? Second, how come video games go as high as 40 frames per second, maybe even higher, why is there no threat of subliminal messages in video games? I always thought I was liked/addicted to video games because they are fun and well made, not because I like it on a unconscious level.

Jonathan Chang - Oct 09, 2006 12:34:02 pm

In response to earlier posts, I don't think this article is really pretending to do anything but make a useful simplification of the human mind, certainly not pretend to be a be-all end-all authority on neural science. I think it's an important simplification, also, in order to make it easier for developers to adapt applications to fit human capabilities. It's very inline with the theme of Anoto, since the what the pen does best is offer up an interface that is the most intuitive and natural of current technological offerings; technology should fit people, and not the other way around.

It's also true that even as a "simplification," this article is very complex and specific compared to what we're attempting to do for this class. We probably wouldn't be taking too much of this information into account, since we have the benefit of an interface that always works at the pace of the user. However, I could definitely see uses in other applications, like games or other highly personalized tasks. Adaptive difficulty and other aspects could use this information to finely tune operation.

Keenahn Jung - Oct 09, 2006 12:48:10 pm

Having taken a few neuro and psych classes, this stuff was perfectly manageable for me to read. As a metaphor, the brain as computer is pretty compelling. Of course, we learn in Philosophy class why this does not strictly hold, but it is useful nonetheless. Understanding the brain is key to understanding what makes good interfaces.

It is possible to design interfaces from the top down, with functionality in mind, and iterate based on user feedback. But, it is also possible to design them from the bottom up, that is, with human psychology central to the design. This paper illustrates some of the points we should keep in mind if we go this route. The numbers and charts help and are good in that they provide grounds for this metaphor.

Also interesting to me was the illustration of the differences between the Mac and Windows toolbar. Personally, I think the Mac is more efficient, and on all my windows machines I move my taskbar to the top of the screen. Of course, when this paper was written, the Mac didn't have a huge row of icons on the bottom of the screen as well.

Siyan Wang - Dec 07, 2006 03:10:11 pm

1.) I found this article to be pretty interesting in that it details a very specific and technical analysis of the human perceptual processor, which in turn helps one develop interfaces to better suit those perceptual processes. I think this is where the HCI research becomes more multi-disciplinary, as this article draws heavily on cognitive science, with which I am very familiar. I am curious about how the auditory and visual store systems interact with each other, for example to correlate the two.

2.) Lots of the memory stuff I find quite interesting, such as the specificity principles and other things for encoding of information. However, it is good practice to allow users to interact with the interface from simply recognizing the interactions, not trying hard to recall them from memory. I suppose encoding specificity comes into play here because certain icons or labels may trigger memories for how to interact with them.

Robin Franco - Dec 15, 2006 12:42:34 pm

Comment 1: This was a very technical reading that was more focused toward cognitive science than user interfaces. Although it was a bit too technical for me, it was nice to read up on the advances on cognitive science and the human perception. I'm not quite sure how useful this would be in helping improve an interface.

Comment 2: I think that as much as you want to think out the perfect interface, it is best to just leave it up to the user. The human mind is a very complicated structure, we can try to simplify as much as possible, we can try to create theorems on these simplifications, but then applying these theorems to the development of user interfaces is a fallacy, in my opinion. Even if theoretically, an interface is perfect based on what we know of cognitive processes, it can be a total failure when applied to the real world. Nothing can beat an old-fashion contextual inquiry, in my opinion.



[add comment]
Personal tools