Human Information Processing (Perception)
From CS 160 Fall 2010
Slides
Readings
The Model Human Processor Chapter 2 of The Psychology of Human-Computer Interaction. Card, Moran & Newell. Only Pages 24-76.
Your Response
Login by using your user name and password, edit the page and at the end of it put a whole line == ~~~~ == (literally) at the beginning of your submitted critique, so the wiki system will index, sign and date your submission automatically.
Alex Aberle 15:26, 9 October 2010 (PDT)
I had never thought about the human performance being quantified like a computer before reading this chapter. It is interesting to think about the limits of our capabilities, especially compared to a real computer. 100msec/bit is quite a few orders of magnitude slower than a modern computer. Obviously, humans can't rely on reaction time or raw calculation ability. It points to some more holistic property that gives rise to human creativity and ingenuity.
The reading stopped short of explaining why this "human processor" model is useful to HCI. Sure, you can use it to calculate roughly how long a simple task might take (with huge variance). But why bother with a rough model when you can just test on a real person? It is very, very easy to create even a high-fidelity mock-up these days, let alone a paper model. I think in practical terms, a good user study is going to give you a lot more insight and feedback than paying an HCI expert to calculate theoretical task efficiencies (or whatever it is that an HCI expert does).
Jonathan Look 17:55, 9 October 2010 (PDT)
The modeling of human performance goes through a technical analysis of limits of human capabilities. In one regard, I found it as uncovering the details of a person's 'reaction time' to an event, breaking it down into cycle times from perception, to cognitive recognition, and the motor response time (like an event, processing, then action).
As far as using this model in relation to HCI, I think it can be applied to features of the UI such as widgets, their relative size, location to other widgets, etc, to gather a more object view towards the design. Of course, testing with real users still needs to be done to get real world results, but both Fitt's, Bloch's, and the Power Law of Practice provide UI designers tools that can help in complicated designs (such as letter placement on a keyboard or buttons on a calculator).
Sean Tai 09:19, 10 October 2010 (CEST)
I have always found Fitt’s Law to be an intriguing because it quantifies the efficiency of an interface in terms of distance and time. It still puzzles why many widely-used, mainstream applications seem ignorant of the law in not placing clickable regions of components flush with a side of the screen. An example is Mozilla Firefox’s tabs, which are not flush with any side of the screen, even in the current 4.0 Beta. Google Chrome had its tabs flush with the top of the screen in its first beta build; this design makes it much easier for users to click tabs by making the size of their clickable region functionally infinite. With tab-selection probably the most common task performed in browsers today, it is hard to understand Firefox’s refusal to adopt Chrome’s superior design.
Melissa Lim 09:27, 10 October 2010 (CEST)
The Model Human Processor gives an interesting perspective when comparing human working and long-term memory to the memory and processors of a computer. However, computers are far from being as complex as the human brain- it is not just about simple input and output. Although there are similarities and trends among people (most of us can only remember up to 7 digits upon first hearing them), there are also many variations regarding what we perceive and interpret from a given stimulus. The environment in which a person is processing information is also important to consider. Beyond the physical environment, this also includes aspects such as previous experience, knowledge, and state of mind. For HCI, it is definitely important to incorporate human trends in information processing, but designers must also consider the context in which the user is working. The model serves as a useful starting point, but I still believe testing with human participants is essential in usability evaluation.
James Yu 10:05, 10 October 2010 (CEST)
The reading on Model Human Processor gave me some insight on how to estimate user reaction time to better design interfaces. It was interesting how human behavoir could be modeled like a computer. Overall the extensive examples given seem to show that reasonable estimates can be gathered from the Model Human Processor, even though during the early parts, it seem like there was too much variance for the results to be useful. The author handled the variance by coming up with Fast/Middle/Slowman. There were also a couple of interesting laws useful for analysis, for example fitt's law for positioning buttons, and also the power law of practice, which can be useful to spot differences between expert and novice users.
Calvin Wang 21:03, 10 October 2010 (CEST)
I find the reading fascinating and extremely informative. This class is about Human Computer Interaction, but it's not until halfway through the semester that we learn a model for humans. The three-component model of the human mind, although vastly simplified, offers an effective way to evaluate and predict human performance in various perceptual, cognitive, and motor tasks. Albeit useful, it is by no means a replacement for the user studies done in UX research. Rather, the power of this model is that it serves as a rough rule-of-thumb that guides design decisions before they must be committed to a prototype for user evaluation.
Frank Chew 21:34, 10 October 2010 (CEST)
The Model Human Processor is a neat yet oversimplified model of the human mind. Taking an MRI of the human brain would not reveal the same processors as this model. Yet this model is necessary in order to come up with many equations that predict behavior. Do the equations on page 27 really come about as a result of this model? Or were they just grouped together because they both model human behavior? What user-computer interactions can really be predicted accurately from the memory/processors model? Their equation (derived from this model) is Reaction Time = Perception Time + Cognition Time + Motor Time. Yet how do they actually verify the Perception, Cognition, and Motor times with real users?
Raymond Williams 01:48, 11 October 2010 (CEST)
This reading was interesting because it looks at functionality. Math is everything, and everything is math - but it gets even more simplistic. 0 and 1. The way I see it, everything can be broken down to 0's and 1's. Even our own brains are binary; each neuron is either firing [1] or it is not [0]. So it only makes sense that we would model our computer systems after this binary model. The problem is that the brain can handle many [very many] signals at the same time while computers are limited to the number of processors they have. Perhaps, eventually, we will fully understand the human system and be able to design an interface to interact directly with the brain. "Oh no, I need to give a presentation but the audience only speaks a language that I don't know." says someone 40 years from now. "Guess I'd better plug in my USB Brain chip and upload the language to my brain."
Derrick Tao 02:23, 11 October 2010 (CEST)
The Model Human Processor is really interesting in that it simulates the human brain. One of the topics in this article is the one about memory. The way that our brains encode memory can be broken down into very simple steps. Another part that was interesting was how layout can be calculated to have a good balance for the user and how they see the application. In any case, trying to make a model of the human brain is very difficult but it looks like that it is getting there. Maybe one of these days we will be able to see a robot interact very much like humans.
Chris Song 05:22, 11 October 2010 (CEST)
Model Human Processor gave me a much more technical view of the concepts that were already very familiar to me from other sources. One of the interesting topics was on memory. Some of the ways the text discusses about storing things in long term memory are the same things they teach you when you are trying to study for a certain exam that require memorization. I have also heard that people with good memorization skills are the same people who have very good word association skills. The reading made it more concrete to understand why that is the case. One of the confusing topics to me was the topic of maximum frame rate for fusion. I understood the concept behind the frame rate required to give a illusion of animation, but it wasn't very clear why movies are limited to 24 frames/second and computer games can go up as high as performance of your video card. What is the benefit then of having 100fps in a video game? It certainly “feels” better to me, when I compare it to the same game that's running at less than 60fps. So is this real perception? Or just a imaginary “feeling” I'm getting due to my knowledge of the frame rate numbers?
Andy Lin 05:27, 11 October 2010 (CEST)
After reading this article, I was surprise how we can design the layout of an interface based on calculations. One law that I believe is useful when we design the layout is Fitts’s law. As the calculator example shown in the article, users can save a lot of time on using devices if the designers can design the layout of device properly based on the distance and the relation of buttons. Another useful one is power law of practice. The example shown is the different layouts design of keyboard, which are much carefully designed than I thought. I really think that a great layout is not that easy to build, and we should take this issue into account when we do the project.
Christine Lu 08:43, 11 October 2010 (CEST)
At first, I thought this article was very odd as it tried to quantify the human mind, but as I read more, I saw that there were a lot of useful properties to be observed from motor skills, reading rates, etc. that are very useful in programming. Interestingly, in programming classes before this, we don't usually take into account these aspects of the human mind and revise our programs for it, but the article does bring up a good point in terms of user interface and making programs seem seamless when in effect they are only just enough to trick the user's mind.
Richard Nguyen 09:59, 11 October 2010 (CEST)
THe model human processor Is an interesting look at the limitations of the human mind and body. What I found interesting was the example of the simulation of causation in programs. The example that they used was the collision of billiard balls. If we're to simulate a collision, the calculation and subsequent movement of the balls have to occur within a specific time frame so that the illusion of reality is not broken. I had not thought about time limitations of calculations and responses from the computer in response to user interaction greatly affecting the user experience. But it's clear that in order to get the proper response from the program and the user, the programmers/designers must take into account the limitations of the human mind when deciding how to implement certain features. Also the calculator example was very interesting as well. By associating keys that are used often, and clumping them in one spot, the designer can greatly improve the performance of the user. Also, using the calculator example we can also see that the calculations that are proposed in this chapter aren't as accurate as we'd like but are still very useful as estimators for what we need to know.
Jeremy Sasson 10:10, 11 October 2010 (CEST)
I think the Model Human Processor is a useful for evaluating an interface design with a specific purpose. In general, I think that testing your product with real users would be the most effective way of fine-tuning your device. However, when it comes to nitpicky things like making the product execute certain actions in the shortest time possible, the quantitative data that was in the reading could prove quite useful. While a real user cannot tell you (to an accurate enough degree) if jumping between two different views could have been made faster, the data that is compiled can.
Daniel Yoo 02:09, 11 October 2010 (CEST)
It is amazing how people are thinking about modeling a human processor. It can be divided into three interacting subsystems: the perceptual system, the motor system, and the cognitive system. When I took artificial intelligent class, there were many examples where robot cannot imitate a human behavior. Human brain was far more complex than the computers simple inputs and outputs. One of the examples was a calculator where programmers must have taken a consideration where mathematical and user interface is important.
Brian Maissy 11:10, 11 October 2010 (CEST)
This is amazing! I'm not sure quite how useful it would be to interface design, though. Most of the qualitative conclusions that can be derived from these quantitative results could have been derived simply from intuition, such as putting a commonly used key in a central location in an interface. This reading does offer numerical estimates of how much faster it would make the interaction, which could be useful when choosing between different placements. The uncertainty principle seems to be a pretty good proof that simple interfaces are easier to use (which was also obivous). I wonder if anyone has ever tried to expand on the algorithms used by the cognitive processor to make decisions. I know it would be ridiculously complex, but I've always wanted to build a software system to model a human mind. Algorithms for emotions, realistic human decision-making, and personality development would be an amazing breakthough in AI.
Yue Chang Hu 12:23, 11 October 2010 (CEST)
This reading is quite confusing and difficult for me. I was intimidated by the amount of symbols, calculations and graphs that it was showing. Nevertheless, something that i found interesting is the "Model Human Processing". If i am not wrong, this is some sort of model or idea used to describe human user reaction or behavoirs to certain situation. The author broke down this model into 3 categories and they are, perceptual system, motor system and cognitive system. The category that I enjoy the most and acutually think i understand is how the author describe about improving cogntive memories by trimming the perceived input into seperate chunks. For an example, trying to remember a 32 bit binary number is much harder than trying to remember its hexidecimal form. Another thing that i think is kind of cool is how the author models the decay of wokring memories, because this knowledge can be very helpful information for implimenting a good User Interface.
Mark Wei 18:29, 11 October 2010 (CEST)
The section on Fitts's Law of the reading got me thinking on mouse pointer acceleration. Most computer power users will say that it is more accurate when you turn off the pointer acceleration, and Fitts's Law seems to explain this. T_pos is dependent on T_c, the time for one cycle of the Cognitive Processor to decide on each correction. The brain has to do more work when faced with an accelerating pointer, so T_c may increase, thus making it harder to make quick judgement on how much correction to make while moving the pointer.
Alan.choi 18:59, 11 October 2010 (CEST)
I find it absolutely amazing that scientists have experimented and analyzed people to the point where they can quantify their actions. However, it's really hard to create interfaces solely based on these numbers, and we don't necessarily have to because we sort of know what these numbers are just from our own experiences with our own body. The section on keyboards particulurly struck me because it was interesting to see how the qwerty keyboard actually came to be with the actual rates of usage and the results of the experimentation that went on.
Courtney Wang 20:14, 11 October 2010 (CEST)
There are several interesting intricacies of human brain functionality brought up and explored in the article that are critical in designing good interfaces for user applications. Something that's not often considered when creating complicated apps is the fact that any given user's working memory is only 7(+/-2) items. This aspect of the Cognitive Processor reinforces the need for easy navigation and simple menus. The "Less is More" principle is directly connected to the limitations of working memory. Designing applications that account dramatically lessens the learning curve for complex applications. The Power Law of Practice is also important to consider if you want to design a simple to use application, especially if it is one aimed to be used often (which most apps should aim to accomplish). Creating repetitive, quick, and intuitive actions for users means more adoption at a faster rate.
Steven Kisely 11:16, 11 october 2010 (PDT)
In terms of the Model Human Processor we are nothing more than a Turing machine. When a human has to make a decision they look at the resources available, variables, and finally their own history. From this information the human makes a choice. I found the breakdown of how long it takes a human to perceive something and react to be very interesting. I did not know human response / interaction time had been so well captured.
Tsung Han Tsai 20:21, 11 October 2010 (CEST)
The reading seems to be just a bunch of statistics. It would have been better if they also showed how fast it would take maybe a supercomputer to perform the same tasks that they described. I guess it was kinda cool how they separated the brain into perceptual, cognitive, and motor sections, which kind of reminds me of MVC model.
Bernard_Wong 20:22, 11 October 2010 (CEST)
Statistically analysis is often helpful in a good user design. Websites often use user statistic to make it more profitable, by analyzing data likeuser cursor, like a heatmap of where users clic most often, how far do they look down a page, how many times, and where they click on a page. Knowing a user's response time and their memory decay would help in many cases, like in the case of a game developer, know these limits would help produce a game that challenges those limits but not to be gone as far as to exceed possibilities.
Samantha Paras 20:22, 11 October 2010 (CEST)
One thing about "The Model Human Processor" that interested me was how easy it is to compare a human mind with a computer. For example, they write about the difference between long term memory vs short term memory and the speed it takes to recall history. This is very similar to the memory storage on a computer, and the speed differences between RAM and HDs. However, I wonder how accurate all these calculates are. Humans are so different and I feel like they way they think and how they react to things can vary a lot. However, all these calculations can be helpful to us when designing UI. The calculator example was a good example of this.
Aaron Loessberg-Zahl 20:56, 11 October 2010 (CEST)
The concept of the MHP seems to be useful in designing interfaces, but this chapter seems like more of a reference than something to be read and comprehended in a single sitting. The prevalence of numeric figures and equations is what, in my opinion, forces this interpretation of the text. Taking this into account, this chapter is very poorly formatted. If they wanted to convey to us the general concept of the MHP, the excessive use of numbers and equations (especially) should be omitted in favor of limited inline numeric results and occasional diagrams where necessary. If the aim of this chapter was to convey the response, decay, and access times, along with the rest of the seemingly endless list of specifications of the MHP, then an extended table or index should be used so that information is readily retrievable. As it stands, this chapter utilizes neither format, and is therefore very difficult to parse into a coherent theme or idea.
Sui Kun Guan 21:07, 11 October 2010 (CEST)
A computer system is similar to the system of human being processing information. Therefore, the Model Human Processor is constructed, and it has three subsystems: the perceptual system, the motor system, and the cognitive system. In order to support the cognitive system of the human-computer interface, the information processing should include task analysis, calculation, and approximation, which is very similar to the process of designing user interfaces.
Anthony Puccinelli 21:13, 11 October 2010 (CEST)
The article's efforts to map the way a computer works onto the human mind is interesting. It seems to make a rough sort of sense, despite the things I've heard about cognitive scientists disagreeing over how similar the brain is to a computer (though I may not have heard from credible sources, my impression was that cognitive scientists think the brain works completely differently than a computer). That said, a human is an organism that takes stimulus and reacts to it, recalls things in memory in a quite different way depending on how short-term that memory is (this is similar to the different behavior you get recalling cached data and un-cached data with a modern computer), takes a longer time to think through (process) a problem the more complicated and uncertain the situation it, and that reacts via reflexes to urgent and imminent things (like a baseball heading towards one's head) despite whatever is otherwise on its mind, which is similar to how a computer will always try to refresh the screen a certain number of times a second and react to I/O events immediately despite what other kinds of processes it's working on in the background. So a case is certainly well-made for the similarities between modern computers and the human brain, my question would be is this of more interest to computer scientists who want to make more efficient or human-like computers based on this model or to cognitive scientists who want to see how naturally the human mind has replicated itself via the work of computer scientists!
Kyle Gorlick 21:23, 11 October 2010 (CEST)
I was initially unsure how valuable the model human processor would be compared to simply testing on a bunch of users. There are a lot of interesting applications of this model though. For example, determining the frame rate of video and using Fitts's law. I wonder if in video games, Fitts's law and this model are used to determine button combinations and the precision required. It's interesting how this abstract computer model of the human brain can be so useful.
Karthik Jagadeesh 21:26, 11 October 2010 (CEST)
It is interesting to draw an analogy between the human mind and the computer. The brain basically consists of a series of connected memories and processors, and a set of memory code types. Each of the processors has a cycle time of around a tenth of a second. Now using this information it is possible to come up with many approximations on how effective certain visualizations would be for the human brain. In order to determine effectiveness you calculate how much of the material would have transferred over to working memory and long term memory, and how much of the information you can bring back.
The time it takes to move the hand is another important factor to consider when designing a user interface. It seems like this value directly corresponds with how precise the hand needs to be to click the button. By chaining the locations and sizes of buttons you can make the experience for the user much better and save a lot time for the user. It is interesting to see how much of a difference these small things can make.
Chao liu 12:28, 11 October 2010 (PDT)
For this reading, it's useful to have a look at the realtionship between human and the marchine. As the information input in the marchine, if the marchine can think in a way that people can understand. It will be very nice for us to know the interface and also how to get a eaier way to do things.if marchine can think in a human way or give respond as a human, it will be very useful to make marchine to help people in some dangerous condition that human may not be able to deal with. And the marchine will easier to be understand.The interface is the way that marchin talk to people,therefore, a good designer need to make the marchine more human like.
Sung Ma 21:31, 11 October 2010 (CEST)
Once, I remember reading a small passage or an article comparing the digital camera lense to human eye and describing how digital camera lense work similar to human eyes. Similarly this passage compares the human performance and computer. The model human processor gave me many thoughts on so many different things. The model human processor which is used to describe and reproduce human reactions and behaviors is used to test perceptual, motor and cognitive systems. What stunned me most was that I had never thought about human to be restricted to any limits. It is the designers job to understand the limitations of human mind to get the proper response from the program and the user. On top of all this, the fact that all this procedures is happening to create a better program makes me astonished.
Richard Laroue 21:30, 11 October 2010 (CEST)
As a cognitive science major, I have had the opportunity to take some psychology courses here at Berkeley that dealt with similar ideas to those in this weeks reading. The first time I began to think of humans as information processing systems was in reading "The Magical Number Seven, Plus or Minus Two: Some Limits on Our Capacity for Processing Information" by George Miller. This article talked about certain limitations to human memory recollection. I have also read about human abilities regarding response times, attention spans, and computation speed. It is interesting to think about humans in this way. And in terms of user interface design, knows how humans process information can provide useful insight into how technology should function.
Danica Shei 21:29, 11 October 2010 (CEST)
This article definitely introduced me to to the powers of AI and how comparable a computer is to the human mind once divided into the subsystems that this article suggests. It is amazing, as well as alarming to see such non-technical actions such as memory and cognition put into such technical formulas and terms. I realized from the reading how ineffective it would be to design interfaces and programs without knowing the basics of human limitations. This article really puts the design of user interfaces into perspective, as it stresses that technical solutions that programmers create must ultimately be understood by users without any technical background.
Alexander Wong 21:32, 11 October 2010 (CEST)
I've compared humans to computers before in everyday situations, sometimes as part of a nerdy joke or to illustrate a point, but I have never broken down our human processes into system processes. The world of AI seen in movies like the Terminator seems distant, but this reading brings its reality closer. Human beings are in very many ways like computers. Of course the technology gap is huge but the analogy presented in the reading was convincing. I wonder if we could learn to build better systems from studying the way we do associative memory.
Anyways, turning the discussion back to UI. The analogy presented in the reading is helpful for finding parallels in computing and human perception to better our UI designs. Learning for instance that our ability to recall vocal memories better than visual memories, pushes us towards having audible feedback instead of just visual.
Soroosh Izadian 21:38, 11 October 2010 (CEST)
I find the data presented in this reading very interesting. The fact that we can quantify human perception based on time is very important because it shows us the limitations we face when having a person interact with any device. For example knowing the required frame rate to present a smooth animation helps us design our video cameras, video encoding/decoding procedures and appropriate video output devices. This is more important when we're designing an interface that has to take into account the user's reactions and choices and the time they take so we can make the ui as responsive and fast as possible and at the same time not present information faster than the user can process.
Tiago Bandeira 21:40, 11 October 2010 (CEST)
It’s amazing that researchers have managed to break down the clock-speed of the human mind. I had no idea that this was in fact possible and had been achieved prior to reading this article. That being said, it is phenomenal that given the number of “chunks” the human brain can remember (roughly 7), someone was able to memorize 81 separate binary digits through a complex method of combining these digits into chunks that represent larger quantities of data (An example would be hexadecimal but hex still only gets you to roughly 21, or 3 times what is possible). How did this person do it? It would take base 100 or 10 binary digits per chunk to get to 8.1. Each chunk would then have 100 separate possibilities. The difficulty to decode and encode this makes it fairly infeasible and the text stated the guy did it in 1 second per binary digit making it a very impressive accomplishment.
Albert Tseng 21:41, 11 October 2010 (CEST)
I found the Model Human Processor to be an extremely insightful read. It provides an alternative view of how human behavior is formed, using an analogy similar to computer architecture. Putting things in terms of processors and memory stores creates a new perspective of the way different biological aspects of humans interact to produce behavior. It allows the study of user behavior to be quantifiably examined, and allows designs to be made according to formalized properties of user behavior.
Robert Connick 21:54, 11 October 2010 (CEST)
That was a very interesting read and much closer to what I thought we'd be learning in CS160. I was a bit skeptical of the numbers at first (you can really quantify the time it takes an average person to make a decision? Come on!), but the references revealed that the values were based on empirical data. It's still only enough to model very simple tasks (not based on knowledge of the "system environment"), so we'll still have to analyse more complex tasks.
Regarding Chris's question about frame rates, I suspect that movie frame rates can be lower because each frame includes motion blur, assisting in the illusion of motion. I don't know if more modern games have this, but it would take up some processing power to render. Maybe this could be modeled by saying that each successive game frame is "more different" than successive movie frames, although I only base that on intuition.
Benjamin Carpenter 21:55, 11 October 2010 (CEST)
It is interesting to see that most of the visual motion effects on computers depend on being fast enough to look realistic. Cartoon framerates have to have a certain FPS in order to look like continuous motion, and collisions onscreen must react to simulate real collisions. The sam is try for audio, as the sample rate has to be high in order to sound like continuous sound. These psychological tricks apply to video/audio compression as well. Lossy compression must through out some data, but keep enough of the right information in order to seem like the original media.
Arthur Huang 21:55, 11 October 2010 (CEST)
It's not surprising to me that the human brain can be modeled into something like the human model processor. Thinking about how humans interact and think in their daily lives is really helpful for designing applications and user interface tools, as seen with the calculator example. However, this was a rather technical article. I found Fitt's Law to be interesting, as I had participated in a research study here where I had to click multiple circles and triangles that popped up on the screen, as a part of research to improve on Fitt's Law.
Don Arboleda 22:06, 11 October 2010 (CEST)
I never expected that so many things about the way the mind works have been experimented and calculated to the number-value level. I could see motor skills being easily measured and broken down, but language? Memory? I would imagine the time frame these things act upon would be quite difficult to measure. I had never even imagined that people would be attempting to measure such values. In light of this, I would think that people have already measured, calculated, and extrapolated the data for such things as even typing this response.
Bichen Wang 22:09, 11 October 2010 (CEST)
This article really evaluates a human as “just another piece of machinery.” If you look at the specs of this machine, it really is absolutely terrible. What is this huge response time with a clock period in the hundreds of milliseconds? Why can it only remember 3 things for 30 seconds only 40% of the time? I would not want to use a human as a machine! It absolutely fails at processing. It’s amazing how computers can do so much more than humans yet are so mechanical in the process. Although it is good that when asked to find an inverse of a 100x100 matrix, it will not decide that it does not want to and just spew out false answers, computers are not creative at all. It might be interesting to see how much processing power is lost in simply the biological mechanics of thinking.
Karl He 22:17, 11 October 2010 (CEST)
Quantifying the abilities of the human user is a good "abstraction" for the UI designer. By breaking human behavior down to numbers and statistics rather than general feelings like "it's slow" or "this button is too small", you can instead say something like humans only feel slowness if the delay is >1s, or buttons must be 24x24 pixels to be easily pressed, for example. Although it may not end up accurately reflecting humans, it gives a pretty good model that can then be run against actual humans at a later stage.
Avery Gee 22:23, 11 October 2010 (CEST)
I think that it is useful to look at how the human brain handles perception and computation. The brain is very well evolved to do some things really well which are hard for computers to do. This includes face recognition and edge detection. However, there are things that computers do much better (like memory). This related to creating interfaces because it is useful to know what the brain is best at doing and what it is not so good at doing. The goal would be to use both the computer's strengths and the human brain's strengths to make the most effective interface for the user.
Theron Ji 00:40, 12 October 2010 (CEST)
While the article presents an interesting theory on how to model the human brain using the analogy to a computer, I think it's important to remember that the computer is simply a construct designed by the human brain itself, and is not natural, unlike the human brain. In other words, this article seems to be a bit backwards in thinking, since the brain came first, then the computer. However, I understand the value in doing this, as the human brain is far too complex to be fully understood, and using a simplified model (or as simple as a computer can get) to try to understand it is a smart idea.
Simran Chaudhry
I think we can use this metaphor comparing a human to a computer to an extent. Working/long term memory, as well as perceptual and motor processors have fairly analogous counterparts in computing. The most interesting aspects of humans -- the cognitive processor -- is barely discussed in the article. Like a CPU, the cognitive connects perception to motor. But, the majority of what the "cognitive processor" can do -- namely learning/retrieval/creating solution -- has no real counterpart in computing.
