Yeah, I neglected the history threat to the validity of the comparison. The first was on a Friday and the second on a Tuesday. Maybe there are different browsing patterns.

Or, a conspiracy. I bet ido.1st works for Facebook.

Thanks and good idea on the silent flag. What's <code>sed d1</code> meant to do? On my machine I get an error from sed.

Totally. I tried QSB but it also seems to have lost momentum. (The killer misses for me were 1) no remote host shortcuts and 2) no pivot to paste an email address from a matched contact (issue 605). It also seemed a little slower.

Ah, so everyone already knows this. :) I didn't find anything in a search on "right-click" but "command-click" reveals several prominent pages about this behavior.

Interestingly, the right-click (keyboard-less method) was added in Leopard. And for the one or two people who don't already know, doing the same in Safari will step up the hierarchy of the current URL.

Oh I get where you're coming from now. If you just want them to "the simple logics and ideas to kids" then they shouldn't have to learn English, however little, at the same time.

I would go further though and say they shouldn't have to learn any syntax at all. There are a number of introductory programming environments in which you create instructions by linking graphical symbols from a palette. This makes it impossible to make a syntax error and yet it is Turing-complete. The most successful such environment, to my knowledge, is the Alice project from Carnegie Mellon University.

[in reply to your e-mailed reply. I thought it would be better to keep the discussion open in case anyone else has something to contribute]

Right, there's no machine translation problem in the case in which natural language strings are in Chinese. That's why I wrote that example. I think it's the more realistic status quo, assuming unicode identifiers (which I forgot that Python <3.0 doesn't have).

My point was that once that is the baseline, why not learn the fixed set of language primitives? Especially since they'll need a grasp of English for the rest of the libraries, as you point out.

Thinking about it more, I'm not so sure what to expect. If you're training people to code in Python, then I think getting them into English fastest is best. But if you're training them in the concepts, the language doesn't matter. So it really depends on what your learning goals are.

This is a very interesting blog post. Thanks for writing it.

I am glad to see the possibilities of open-source programming languages be explored through projects like this. I'm sure this will appeal to many Chinese people.

As for learning Python, I expect it would be counter-productive. You've suggested a study which would answer the question, so if someone would do it we would know for sure. Until then, I think theory argues that it would be detrimental if used for more than a week or two. As you point out in Jim Cummins work, English language learners need _some_ English with their domain instruction. The fixed set of primitives in Python may be a fair amount to learn, especially since programming language symbols are mnemonics for computational constructs and not actual language.

It would be bad to expect a Chinese student to write their whole program in English, as you show in the English example above, but that's not what a Chinese programmer would have to do. The following is a more realistic analog of your "Chinese Python" example. Do the non-Chinese terms seem like much of a burden?

# Standard Python by a Chinese programmer

#!/usr/bin/env zhpy
# 檔名: while.py
數字 = 23
運行 = True
while 運行:
猜測 = int(raw_input('輸入一個數字: '))

if 猜測 == 數字:
print '恭喜, 你猜對了.'
運行 = False # 這會讓循環語句結束
elif 猜測 < 數字:
print '錯了, 數字再大一點.'
else:
print '錯了, 數字再小一點.'
else:
print '循環語句結束'
print '結束'