Lambda the Ultimate

I havn't heard comparison between written/typing speed. But I can tell that with {one Chinese two bytes / one English char one byte}, the same thing in Chinese will need much less space in a computer than that in English.

I heard there is person that can type 160 Chinese char per minute by ordinary keyboard. And I have used Speech Recognition software to input Chinese as early as 1996.

...But I've read here and there that the Japanese have put more effort into speech recognition, owing to the difficulty of entering their text on a keyboard.

Whether or not there has been much effort on Japanese speech recognition, it likely has nothing to do with any difficulty in entering Japanese text on a keyboard. I have used the facilities on both Windows and MacOS to enter Japanese text on a standard American keyboard with no problems. Typically one types the romanized version of Japanese words (phonetically) and these are automatically transformed (either on the fly or at the end of a sentence) into words written with hiragana, katakana and kanji as appropriate (with the assistance of an extensive dictionary).

I am a native American English speaker, so my experience may not represent that of a typical Japanese. All the same, I cannot imagine that native speakers have a *harder* time with text entry than I do.

Statistically, Chinese and Japanese students are good at math, especially yonger ones. I do not know about the speed argument, but this may have something to do with their native writing systems. The Chinese writing system (also borrowed by Japan) is symbolic rather than phonetic. That is, one character=one concept, having no or little connection with pronounciation. And the language of math or programming languages are similar. You write just "5" and not "f-i-v-e", or
"+" and not "p-l-u-s".

I think you're making some assumptions about how the brain works. Specifically, that it works like an electronic computer. Some research suggests that the brain only gets faster the more its used--as if a super-exponential algorithm just materializes more transisters on the processor to speed things up.

In any case, there are too many unknowns about the human brain to know for sure.

I've had this hypothisis for a while: Computers for widespread, everyday use would need to be developed by a culture with a simple alphabet.

The reasoning is that the process of encoding and handling the language for complex alphabets (like Chineese) is so difficult that it's almost easier to get everyone to learn a new language. Unicode, for all its complexity, still doesn't cover 100% of all human written langauge. However, the alphabets for most western European languages fit well inside an 8-bit word, and has fairly simple transformations (toUpper(), toLower(), etc.).

Computers still could have been developed as a tool for, say, calculating pi. But I don't think computers for common people would have been developed.

I sometimes wonder just how much the development of computers in general and programming languages in particular was influenced by its almost completely western setting. What would Computer Science look like, if it had originated in Asia, with Chinese as its dominating language?

But that didn't happen and your question is meaningless. Although you could compose an infinite number of possible alternative histories, it is a waste of time. Of what use is this except to create a false mythology of "what might have been"? Your time would be better spent studying computer science.

It really is a pity that CompSci in China (including Taiwan) today is so slavishly geared towards the US... (Take it from a CS student in Taipei.)

Now we get to the real meat of your thoughts: that the oppressive US has forced it's thought patterns on Chinese for decades, in computer science if not in all fields. Rather than focussing on the innumerable advances that Western technology has brought to China, you have chosen to depict things Western as oppressive and you now romanticize about a lost possible alternative history wherein illustrious Chinese computer scientists sitting by now at StarGate-like computer terminals, perhaps computing (with variable word length computers) the genesis of wormholes and the perils of time travel. Undoubtedly, science would have advanced further were it done by the Chinese, you speculate.

It's sad, but I expect to hear a lot more of this garbage in the next 20 years as China staggers to it's feet, mostly thanks to adoption of largely Western methods. Note that the hole China dug for itself, that it is now climbing out of, was dug by the Chinese, not by the US or the Western world.

The formation of a romantic vision of a society (complete with sham history and religion) has happened in multiple societies within our own lifetime:

• Germany's fascism prior to World War II was a romantic view that actually rejected Western methods and embraced a romantic vision of a master race,
• As Japan rose to modernity a similar anti-Western cultural backlash occurred that resulted in Japan entering World War II,
• The acceptance of "Western" methods and science (particularly in the oilfield) in the Middle East has led to a backlash by Muslim romantics that now embroils that region.

The book Occidentalism by Avishai Margalit & Ian Buruma traces the path of the romantic vision you embrace. Read it to find out where such romanticism eventually leads. Sad to say, the roots of this destructive romanticism are often traceable to (surprise!) the universities, wherein romantic young students fall in love with empty ideas that they then nurse and spread.

Anyone should consider themselves lucky to be introduced to leading-edge ideas regardless of the culture of origin. Believe you me, whether it be an Australian aboriginal or Inuit indian or US chemist who shows me how to open a door to the other side of the universe, I'm going to listen to him intently and do my best to understand, regardless of any culture differences.

Alternative histories are bunk.

...but you have very much misunderstood me. All I did was wondering whether Computer Science, had it originated in China, would be different. I did not conjecture _how_ it might be different, and I did not say that I believe such a Chinese CS would be any better than what we have now. Least of all did I revile the West as oppressive.

What I wrote may be a pointless reverie, but I hardly think it makes me a supporter of possibly dangerous romantic ideas or anything.

It is sad that you should misunderstand me that badly, and I apologize for posting something apparently very misleading.

Right now, I'm going to take your advice and get back to studying. Who knows, maybe some day I'll be able to make some _meaningful_ contribution to LtU... (oops, daydreaming again ;-)

Appropriate seasonal greetings, everyone!

Who knows, maybe some day I'll be able to make some _meaningful_ contribution to LtU... (oops, daydreaming again ;-)

Hands up everybody who thinks they're the least intelligent / well-informed person posting on LtU...

There are good and not-so-good reasons why people tend to find this forum a little intimidating. (I'm afraid I'd class the grandparent post among the not-so-good reasons). The good reasons - that the quality of discussion here is often very high, that people really seem to know what they're talking about - are also good reasons to try to overcome one's intellectual trepidation and participate.

This site is a great resource for people who want to learn.

Hands up everybody who thinks they're the least intelligent / well-informed person posting on LtU...

Your original note was seriously just fine, and not at all misleading. Oy veh.

Of course Chinese does have grammar, but it is decidedly less palpable than with any other language I have encountered sofar (at least as far as I, without any linguistic background whatsoever, can see) -- the same way that Lisp has syntax, but less than most other programming languages.

And of course speculations about a Chinese dominated CS (whether Sapir-Whorfish or otherwise) are just that - speculations.

[BTW, good luck in trying to use 了 as a grammatical particle. It works as long as you stick to your textbook examples, but it it is a different matter in actual conversational Chinese -- but that is widely off-topic, of course.]

Really wished we had more words (like the Eskimos) for that stuff that's falling from the sky.

Hoping to have a White Christmas. :-)

By the way, I highly recommend LanguageLog for daily doses of linguistics, cultural commentary and humor. Not quite the laser-focus of LtU, and no monads, but still worth a look :) .

PS: my apologies to Chris if his smiley was meant to indicate he was aware of this issue.

Scott Johnson is right, this thread is drifting, like that cold white stuff the Eskimos have so many names for. ;)

I can't resist a chance to point out that the "received wisdom" about Eskimo words for different kids of snow is not actually true

It's more complicated than that: you first have to show what the received wisdom is. That Eskimos have many words for snow? A specific number? Most commonly 50, perhaps - but 50 is such a suspiciously round number that it's obviously intended as an approximation. Or, is it that other peoples don't have so many words for it? Your average Colorado ski bum has quite a few words for snow, too.

You could make up a similar statement about something such as, say, how many words the stereotypically car-obssessed USA has for "road": there are the classic English-language names like street, lane, way, etc., along with some borrowed ones like boulevard and avenue, but then there are the variations on "highway", which range from old-fashioned to futuristic: turnpike, parkway, freeway, beltway, thruway, expressway, superhighway, plus some more specific ones, like skyway and interbelt.

In the end, does it really matter if the phrase is about how many words Eskimos have for snow, or how many words Americans have for road, or how many words Germans have for bureaucracy? The phrase communicates the idea that people who have a great deal of exposure to something tend to develop a larger vocabulary to describe it. From my layman's perspective, that seems pretty obvious, and borne out by the evidence. The original version of the phrase is presumably a bit more romantic, unless you're actually an Eskimo, in which case it's probably just mystifying.

Pullum seems to rather miss the point, as specialists commonly do when interpreting a statement made in natural language, treating it as though it were a conjecture made in the jargon of their discipline. Such statements have a rich cultural and metaphorical context, which doesn't survive the translation into a more formal context.

Ultimately, the confusion comes down to the denotation of the word "Eskimo". The Eskimos of our collective cultural dreams have exactly 50 words for snow.

I have heard/read that one of the problems with the claim is the agglutinative nature of Inuit languages. E.g., here. But ...

In the end, does it really matter if the phrase is about how many words Eskimos have for snow, or how many words Americans have for road, or how many words Germans have for bureaucracy?

Indeed, Pullum has been popularizing the term "snowclone" for a generalization of this idea. "If Eskimos have N words for snow, X surely have Y words for Z." -- the generalization being the idea of a common phrase framework, into which different values get plugged in. "On the Internet, no one knows you're a Z."

Barely related: I have a book of Hawaiian folk tales, which has a glossary listing about 75 different names they used for the different winds and rains.

Bonus: Chris Small and Keith Bostic's list includes "fritla -- fried snow". :-)

To try to thread this back into a PL-related topic: are agglutinative natural languages analogous to concatenative programming languages?

the term "Eskimo" is considered offensive by Inuit people. Much like other common terms for North American aboriginals, such as "squaw", etc.

I know this is completely off topic but as an expat Alaskan it always rubs me the wrong way when people say things like that.

Eskimo is not a racial slur, it refers to Yup'ic and Inuit tribes. There are also Aleuts, Tlingits, and Athabaskans.

Depending on where you're from (Alaska or Canada) the PC term is either Native Alaskan or Inuit.

[..], or how many words Germans have for bureaucracy?

Exactly one, it's Bürokratie.

I don't believe that people like to make up words for stuff they don't like.

Russian even has distinct verbs that effectively have the exactly the same meaning but take arguments that are tagged differently

Dutch has this too, so maybe English is special here.

Interestingly in Dutch those words have an extra "be" prefix:

• zaaien: to sow (corn) - bezaaien: to sow (the field)
• klimmen: to climb (no object) - beklimmen: to climb (a mountain)
• een tegel: a tile - betegelen: to tile
• veilig: secure - beveiligen: to secure

Any English verb can be transformed into a passive verb which reverses the subject and object.

• I saw him.
• He was seen by me.

The 'be' prefix in Dutch has nothing to do with the passive voice, however.

For verbs to which this construction apply the 'be'-less form indicates an activity that stands on its own, while the 'be'-full form indicates that the emphasis of the activity lies a little more on the object. But note that this does not mean that the 'be'-full form always requires an object; however, it is always allowed to attach an object to a construction with 'be'.

Examples:

• hij klimt - he climbs (emphasis is on the activity, climbing, itself)
• hij beklimt - he climbs [something] (emphasis is a little more on the (absent) object)
• hij beklimt een berg - he climbs a mountain (emphasis is a little more on the object, a mountain)

like: bejewel, to be decked in many jewels; bedeck - to decorate, sometimes in a showy way. These are probably hold-overs from middle English, which I think is the nearest cousin to Dutch.

According to The Grimm Dictionary, the prefixes "be" and "zu" (dutch "toe") are the distinguishing difference between old Norse and old German.

as whether or not issues that arise in natural languages have parallels in programming languages. For example, most of our resident linquists dismiss the Sapir-Whorf hypothesis, but I believe that it is applicable to PL. A PL that makes OO programming easy, will lead programmers to program and think in an OO style (same for imperative or functional or declarative).

In the current case, the question would be whether have a flexible and possibly terse syntax and/or semantics will lead to more concise and easily understood programs (or will it lead to an obtuse and obfuscated style).

Typically, a moderate Chinese dictionary only has about 6000 characters. As I know, a moderate English dictionary has 50000 - 250000 words. (Character is not primitive of English because it has no meaning)

For a more detailed example, in Chinese, there is only one character "chook" and Chinese use "male chook", "female chook", "little chook" (and "chook meat") (two-character word whose meaning is obvious)instead of "cock", "hen" and "chicken" (to understand English, you have to memorize such literally UNRELATED words).

I was thinking more of the tonal issues and ideographic written representation with respect to the number of primitives.

Can you suggest a good book that would explain how Chinese grammar works? Something in English with a linguistics bent rather than the usual traveler phrasebook would be ideal.

Thanks for the insight though!

so I never read any book that explain Chinese in English. But I can tell you is that Chinese has nearly no grammar, only characters. Maybe others can suggest you a good book.

Last night, I also thought about your ideas. And finally, I conclude that languages differs, both natural and programming. Because different languages opted for different design trade-offs, some becames superior while some not.

My favorite Chinese reference grammar is Li and Thompson "Mandarin Chinese: A Functional Reference Grammar".

It has 673 pages of text; an awful lot if the "Chinese has no grammar" argument held.

I would humbly suggest to our native Chinese colleagues that the reason it seems to them that there is not much grammar in Chinese is that they "just know it"; that's what it means to be a native speaker.

Some of the compounds that are "obvious" to them have to be learned by a foreigner; there are often many "characters" that mean the same thing, and you can't just put them together willy-nilly. Only certain combinations are meaninful Chinese words.

One's own language always seems more "logical" than the alternatives because we don't have to think about the grammar to use it. (It also doesn't hurt to have millenia-old sense of cultural superiority to bolster your position.)

If you don't wish to learn to speak it, "Speaking of Chinese" by Raymond Chang and Margaret Scrogin Chang, does a very nice job of describing the language, its constructs, and some of the culture behind it.

A Grammar of Spoken Chinese, by Y. R. Chao, is one of the earlier works on this topic.

i think your comparison is unfair, because characters in English indeed have meaning - they're a pronunciation guide. Going from spelling to pronunciation in English isn't necessarily trivial (it probably happens more at a syllabic level than character level), but it can be done without memorising a (relatively) great number of rules. I doubt that the same is true for Chinese. Besides which, Chinese has multi-character words too, so comparing the number of Chinese characters to the number of English words is meaningless.

+1 on that

english is light on rules, and a bit positional, you can say "greatest small black brittish chair" , in french you'd have to add conjunctions or follow other rules.

but if english is like lisp, should the analogy with chinese be about abstract combinators ? ^^

Did anyone come across research about these issues?

Ehud,

I grew up in Malaysia, and Malaysians get to learn quite a few languages due to the racial mix. Offering myself as a data point, I don't think it is significant.

I find that knowledge of maths is more helpful, since programming is all about concepts like iteration, arrays, symbolic representation and notation.

I've got plenty of anecdotal evidence. We really need someone to investigate this "scientifically" (i.e., do statistics).

Like, how many languages you knew before starting programming, did it help (compared to other people?), how many you learned after, did it help (compared to the previous experience?), etc.

This is something I've been curious about recently. What there was a language that had NO text representation (not text-named variables, if/else statements, function names...)? What would it look like? Now I've seen a few examples of 'graphic' or 'visual' languages but they just seem to move source code on to a fancy window which allows you to connect text-statements using visual arrows (can't remember where I saw it).

When I try to figure out what such a 'language' might look like, I invariably think in terms of trees, graphs, little icons, etc. I think we can be more imaginative. I suspect a visual language will require better ways of thinking about programming constructs (such as bananas, lenses, etc. paper). Another source of inspiration is Wolfram's New Kind of Science. In that book he seems to define 'algorithms' using a visual representation system (sorry, it has been a while since I scanned the book). Another idea is from 3D animation packages and how they allow 'key-framing.' In ther words, could we define insertion/deletion/update procedures on a B-Tree by 'drawing' the tree how it should be just after the application of a procedure, and let the computer figure out how to get there (obviuosly there has to be a limit to how much a computer can figure out). Any way, some quick thoughts regarding an interesting subject.

By the way, if there already ARE such system, I would love to have them pointed out. Thanks.

Wouter's languages are definitely visual. Aardappel is especially interesting, it's based on concurrent tree rewriting.

http://www.cs.bgu.ac.il/~shemeas/juks/

I personally think many of the so-called "graphical" languages are working at the wrong level of abstraction. At the lowest level, text really is the most economical form of communication. It's dense in a good way. Where graphical languages would be useful is at a slightly higher level. In fact, this is an area where I think that PL design can learn from DB design. Nobody writes a relationship diagram with pictures for tuple attributes. That would be just silly. At the lowest level of detail, it's all text. Where the graphics come in is to display the relationship between the different relations. And placing them spatially is more powerful than having to visualize them textually.

We can see hints of recognition in things like folding text editors. Folding allows us to hide the details that aren't relevant to the task at hand. Where graphics would be useful in a PL is at the level of types and functions. If we could see a view of potential call graphs, or visual subclassing/subtyping relations, that would actually be useful. That's why we have UML diagrams. But if a UML diagram were integrated into the language, where you could double-click on a type or a function to open its textual definition, and you could redefine subtyping relations by dragging arrows around a grid, then you would have something that's more than just a modelling tool. Perhaps you could have a button that shows you all the places that a given variable is referenced. Then you click on a variable, and lines are drawn to all the functions that refer to it.

We usually think in text when we think of details, and we prefer to think visually when looking at "the big picture". Note that our language itself gives us a hint of it, since we don't call it "the big paragraph". And when we talk about details, we talk about crossing our t's and dotting our i's, not closing our circles and straightening our lines. A language that recognized this fact and put it to good use, allowing us to visualize the code at higher levels would certainly capture my attention.

Also: The book mentioned in this thread might be of interest.

Almost no one are at least three posters:

xiemaisi:

I think it's not so much about the voices or the number of available syllables, but rather about the almost complete absence of grammar.

Zhu Chongkai:

But I can tell you is that Chinese has nearly no grammar, only characters.

JCF:

In fact, the English language (just like Lisp) lacks all but the most rudimentary grammar

Sorry if my comment implied that absolutely everyone at LtU is clueless about linguistics, I'm sure that's nowhere near true. However my impression is that the amount of absurd claims made in this thread has been higher than in other threads not touching upon human language.

Well, all those people said that Chinese have very litlle grammar, wich is far from saying that it has none.

That Chinese has significantly less grammar than other languages. Let's consider a classic problem of ambiguity:

man bites dog
dog bites man

In languages with fixed SVO order, the S and O are indicated by their position. In languages with rich case systems, the S and O are indicated by markers on the words themselves (such as Latin). Both are grammatical mechanisms. If a language has neither, then the above two sentences are equivalent, making it impossible to articulate the distinction. My wild guess is that it's possible to express both of the above sentences in Chinese, and that the semantics will be determined either by SVO order or case markers.

Chinese, is an analytic language, much like English. In many ways more so; as Chinese pronouns do not distinguish between subject and object (dative/accusitive) forms; nor do Chinese verbs change form to match (agree) with the subject--no is/are/am, for example.

(Of course, as pointed out by someone else, Chinese is also primarily a topic/comment language rather than a subject/predicate language).

Occasionally, Chinese word order does get "reversed" (from the perspective of English linguistics), if the topic of the sentence is the object rather than the subject. When this is done, it is obvious from context what is meant. Chinese also has an equivalent to the passive voice, which is frequently used in such contexts.

Of course, my Chinese is quite rusty--mostly picked up from some self-study, as well as my wife. :)

However, to say that Chinese (or English) has no "grammar" is, as Anton pointed out, preposterous. (Likewise it is ridiculous to say Lisp has no syntax--it certainly does; just a rather simple one). Grammar is not a property which only the synthetic languages (like Latin and German) possess; nor is it a property which belongs only to programming languages with non-trivial parsers.

In my studies of Mandarin, simple sentences like that are formaed exactly as they are in English - Subject Verb Object . In fact most of the sentence patterns I've learned seem to be rather dependedent on word order, unless you are employing a specific cultural idiom. But a large part of learning college level Chinese is learning such idoms, much like college level English students learn bigger more specific English words.

I think something else that should be pointed out is that formal written Chinese is fairly different from spoken Chinese, both in word choice and phrasing, due to the fact that Characters are dialect agnostic, and ideally written Chinese should be legible to an educated reader regardless of the dialect they speak.

I wrote about symbol processing a few months ago on my blog.

For example, I tried to explain that "salad :: [Lettuce]" to someone making a salad.

The whole thing started on #haskell when I realized that I'd know John Goerzen for months, but it had never crossed my mind to wonder how to speak that name. When someone asked how to speak the name, I realized that I probably process IRC/email differently than that person.

In the end I discovered that only one other person on the channel doesn't 'speak' words in their head while reading emails or IRC.

symbol processing! Perhaps this is exactly the point of the poster: chinese people think faster because they think in terms of their imensely large collection of alphabetic symbols? Do they pronounce them when they think or just visually identify them? If the latter, than yes, it should be very fast. On the other hand, i guess it's kind of a feat to quickly recognize the meaning of each one, given context and all...

I do think that languages tuned for a particular medium would be advantageous.

Spoken programming languages would be handy for programmers who are blind or have RSI. They'll also be handy when the entire computer fits into a single earbud.

My guess is that these modes of thinking are not exclusive -- they're combined in different ways in different people. For instance, the language hinted at in the film Minority Report combined kinesthetic input (Tom Cruise moving windows around with broad gestures) with visual feedback.

Hmm...what if an IDE associated sounds with pieces of code? Like, what if a debugger played a tone of a certain frequency whenever a certain function was called? Then you could literally tell whether a program was "humming along" or just "making noise". You probably wouldn't be able to distinguish normal musical tunes, but if you could associate certain functions with certain notes, perhaps you could tell when the function in question is getting called and you don't think it should be.

A musical editor could "play" functions in the source code using similar rules. The benefit would be that coding patterns would correspond to musical patterns, and by "listening" to a function, one might be able to tell whether certain canonical forms were used. Of course, then obfuscated code contests would be as much about producing certain songs as it would be producing nice visual layouts.

One could assign different patches to different types of code. Comparisons could be violins, arithmetic expressions could be piano, assignment could be percussion, etc. The code would sound like a cacophony, but I suspect it would add a powerful dimension of analysis. Normally we only run code at full speed or stepwise. We usually don't run it in slow-motion because that isn't useful. But if we could recognize the normal-sounding behavior of code, it might be useful to run it at a lower tempo, so we can literally listen for the bug.

An interesting consequence of musicalizing programs with function calls being notes or phrases would be that loops with an odd number of steps would sound a bit "off", and the most pleasing data tonally speaking would be that data that was a better candidate for size compression. About the mapping, I am wondering would datastructure->pitch with a fixed instrument bank work better, or treating the data as a sequence of samples (given the latter, you could listen to a program at a faster execution speed, it would just be a bit noisy). As a composer of computer generated music I am having a huge "why didn't I think of this already" moment (well I kind of did think of it, but not quite... one current project is a program in assembler that will send it's own executable code to the sound card, while also rewriting certain parts of it's own code (the idea is from a program I heard about from back in the day, a guy could not afford a sound card or a computer fast enough to do audio, so he wrote a little piece of self modifying code that sent data to the printer port, the printer then being plugged into a mixer), but I had not thought of using audio output as a debugging tool). It may be fruitful to build gdb with the csound api built in, and maybe use some scheme or perl to turn stack frames and memory acesses into note lists and/or csound instruments. that sounds like it could be done as a weekend hack if you already had some familiarity with c/[scheme|perl]/csound.

I needed to test an audio file API to make sure the locking worked correctly, so I had 16 threads each trying to write different pitch waveform to the same audio file. It is an interesting test, because different versions of the OS sound different.

Oy! this thread is too much fun.

I think Sapir-Worf is complete bunk and I thought that was the current thinking in Cognitive Science (not being a cognitive scientist). Language has to do with expression of thoughts not conceptualizing ideas.

I think personality type (as in myers-brigs) or perhaps a psychological astrology outlook (from which myers-brigs silently draws on) would be much a more interesting examination of adept programmers(referring to ehud's inquiry). I think the real work of programming happens in the design which can be independant of language. Language then of course becomes paramount because you must concretize that design or express it to programmers in your employ or share with collaborators.

Programming languages, concepts like Functional Programming or OO, patterns, representations like UML and use-cases, and application design frameworks obviously effect speed and quality of implementation. But ultimatly they are about communication not conceptualization (between developers or a compiler/interpreter). I admit that there is probably a fine line between the two and that there is continuum of abstraction and expression with words constituting quantifications along the spectrum.

I personaly think I am suited to programming due to having a plethora of planets in Libra (air-intelect, relationships-i.e. systems). This is bolstered by Jupiter (the expander, i.e. generalizations/abstractions) in a fire sign.

Why is astrology useful? Because it is a language or framework for examining the psyche. It assigns terms to abstract concepts and expresses relationships and interactions. I think it is a shame that modern psychology does not embrace astrology as a tool. Sure abandon the old programs written in the laguage, but keep writting new ones and develop the language; port it to a new platform... ;-)

If you think in a (natural) language, you are less of an abstract thinker. However you might be a better programmer who could write a better implementation than the designer who gives you a specification.

My experience with chinese is very limited. However, through studying chinese martial arts, I understand that chinese can be limited in the precision of expression. Context is all important and even native speakers can stumble in interpreation. No wonder fuzzy logic was embraced in the east. While faster, it may be a disadvantage in essentially mathematical verbal expression.

I think two things lead to computer science arising in the west as opposed to the east. Economics and mathematical development. While the taoists invented the first documented binary system over two millenia ago, I am not aware of significant mathematical contributions from the chinese. I don't mean to imply any racial or cultural disadvantage and would say that economic tragetory and education practices may be positioning China to be the center of the next mathematical golden age.

As an armchair philosopher, I am prepared to be lambasted for my generalizations and speculations, so bring it on, I'm listening.
:-)
sorry for the long post, as I said, too much fun...

Firstly, I've seen studies of the various attributes of "good" programmers, which found no significant bias towards any Myers-Briggs type (this is a paper I read almost 3 years ago, so I forget the citation), which is as it should be, Myers-Briggs having no scientific basis.

Secondly, there is a mass of evidence that the programming language constructs in people's first language that they are taught tend to be the ones that they are most comfortable with. So, students who are taught imperative languages and iteration, then introduced to recursion (even in the same language), tend to find the latter more difficult, and tend to prefer the former. This is why people typically find it an effort to move to functional programming, not their ability to correctly form the syntactic elements used to encode such constructs. That looks to me like a conceptualisation issue.

Just because Sapir-Whorf is not the whole truth about the human capability to think, that does not mean that it does not have applicability to the way people think about and learn programming. That said, if there are new insights into the way people learn and think about problems, then the computing profession could do with all the help it can get in teaching people good practice.

As I aluded, myers-briggs is rooted in Jung's work which was influenced by astrology and not strict science. But I only said myers-briggs-like, and the metric may not exist to examine the traits that make a good programmer (as if there were one type of programing...).

Your second statement seems obvious and irrelevant. Anything you learn first in a new domain is going to be stronger. But that does not mean that at some point in one's experience new patterns can overwrite earlier ones. The mind is great at working with incomplete information, ie it works with what it has.

At this point I am reminded of Lakoff and his "frames" which I beleive is standard cognitive science jargon. I think Design Patterns ala gang of four represent frames. The word that represents the pattern is inconsiquential and multiple names do exist for the same name. It is the context and details that form the pattern, not its name. One with a richer vocabulary and understanding of design patterns is going to be at an advantage over someone with more primitive patterns or frames. Is this the sort of applicability of Sapir-Worf that you refer to?

New is of course a subjective term. I beleive that studying music as a child, studying Tai Chi as an adult, and having a personality that associates examining complexity with emotional rewards gives me the tools to learn many new things. Practice practice practice practice. Secondly, instructive "etudes" or "forms". Learning must be accomplished in stages with mastery being acheived before advancement.
It is unfortunate that we don't have more universally known examples like "Hello World!".
What I'm trying to say is that I don't think we need any new insights, but to apply existing methodologies from other disciplines that have been proven over long periods of time. Along with identifying the order in which to best present concepts, ie how one builds upon another.

oh, and I meant to include this link that I found.

It seems to be similar to the study you are talking about if not the very same.

I think the real work of programming happens in the design which can be independant of language. Language then of course becomes paramount because you must concretize that design or express it to programmers in your employ or share with collaborators.

If we allow that designers are the elite, then we'd have to say that the language(s) that they use are far superior than the language(s) used by the implementer(s). Considering that most design documents are incapable of being objectively run on a computer, I'd have to question the veracity of that statement (no static type checking, no unit testing, etc).

You are absolutly correct.
But I would still postulate that given two people with the same level of skill and knowledge. One might be better at abstracting a solution to a problem, while the other is better able to do the work of executing the solution.

Isn't UML and things like it an attempt to move towards implementation robots? It may be centuries before we achieve the unification of design and implementation, but for now we need design documents.

Perhaps I worded my statement poorly. I did not mean to belittle the work of programming, I am a programmer more than a designer after all. There are any number of ways an application can fail at it's task. If it is hemroging errors due to sloppy implementation, the best design possible will not make it satisfactory to the end user.
"When you have a hammer, everything looks like a nail" is one of my favorite reminders to myself. So abstracting the problem is an important step in the devlopment process.

People think in different ways.

In Richard Feynman's book "The Pleasure of Finding Things Out", he relates a time when he discovered that he can count at a rhythmic pace while doing anything other then talking. (Read the book, I might have details a little wrong) When he told this to a friend he was amazed to discover that that friend could talk while counting at a rhythmic pace, but the friend could not look at other things. Turns out Feynman is a verbal thinking, and was saying numbers to himself, while the friend was counting by imagining a number line, and "looking" at each number as it went by.

I suspect there are other ways of thinking, but I don't know what it is - I speak the numbers to myself.

When I'm programing there are words flowing through my brain - the words of whichever language I'm working in. I cannot listen to music when I'm working because that is distracting. (Though I'm good at tuning it out)

After I "taught my fingers" how to count in binary, the counting became somewhat automatic, a kind of kinesthetic thinking. See wikipedia's article on Howard Gardner for links to a few other ways of thinking (aka "dimensions of intelligence").

That is very interesting. I noticed many years ago that I almost subconsciously use my fingers and hands to recall the spelling of complicated words by making the finger motions I would do if I were to type the words on a keyboard. They are very subtle movements, I'm not doing any kind of crazy hands-in-mid-air maniac kinda thing. :)

I remember watching this video once which featured Alan Kay in which he spoke of the psychology of UI design. One thing that stuck out was his mention of the mathematician/physicist Hadamard's informal survey among his eminent mathematician and physicist friends. Many of them (I think something like 30%) reported that their problem solving had a very kinethestic quality to it. (I think Kay mentions that Einstein was reportedly in this group.) Another large group relied principally visualization, and only a very small segment used symbolic manipulation.

In the video Kay used this anecdote to make the point that the way problem solving in the sciences is taught is fundamentally broken; it bears no relationship with the way creative scientific workers solve problems. I think everyone involved in mathematics, computer science or physics recognizes this fact at some level but I thought the anecdote was a very good illustration

I actually visualise monads in Haskell as two... I guess spaces embedded in each other (return as one embedding, computations as haskell values as the other), with one space being 'ordinary haskell' and the other having some additional specific structure that amounts to >>=.

Part of me wants to say there's something weird about the semantics of a programming language being visualised as a space, but really, since when's that true? It's almost a phase space of programs...

Well, getting into the topic again as it was brought up by someone else. Olympic Games in Beijing? ;)

I find myself completely unable to speak as I play the piano. I either play or talk, whenever I talk I loose my grip on the notes and rhythm. Thankfully I'm a classical music player, not into songs... :P

When programming, I feel like I'm engaging in an activity like origami. This is actually a deep topic, which made me start a discussion about the role of cognitive models in the design of computer programming tools.

[edit: Context of message got lost. So I've reworded my post.]