any chance of 2.6.0-test*?
A message that is not in the archive (sent off-list or not yet received): [email protected]
A message that is not in the archive (sent off-list or not yet received): [email protected]
[email protected] (Linus Torvalds) wrote on 12.01.03 in <[email protected]>:
Quoted 3 lines
No, you've been brainwashed by CS people who thought that Niklaus Wirth actually knew what he was talking about. He didn't. He doesn't have a frigging clue.
Well ... he did have some nice ideas. Unfortunately, they usually don't get palatable until someone else has worked on them.
If you look into Wirth's books and see that the example code in there ...
* uses one(!) space indentation
* routinely puts several statements on one line
* typically uses one- or two-character variable names
... then you realize that *every* Wirth source needs rewriting.
---
Niklaus Wirth has lamented that, whereas Europeans pronounce his name correctly (Ni-klows Virt), Americans invariably mangle it into (Nick-les Worth). Which is to say that Europeans call him by name, but Americans call him by value.
---
#if _FP_W_TYPE_SIZE < 32
#error "Here's a nickle kid. Go buy yourself a real computer."
#endif -- linux/arch/sparc64/double.h
---
MfG Kai
Quoted 2 lines · Kai Henningsen
Well ... he did have some nice ideas. Unfortunately, they usually don't get palatable until someone else has worked on them.
I actually *liked* Modula-2... but I haven't used it since the very early 90s. I guess I didn't like it *that* much, huh? ;)
Quoted 4 lines
If you look into Wirth's books and see that the example code in there ...
* uses one(!) space indentation
* routinely puts several statements on one line
I've got a copy of the classic "Algorithms + Data Structures = Programs"; in it, Wirth uses four-space indentation but a proportional font. The style isn't all that bad. Hell, I've got worse stuff in some of *my* books -- oops, shouldn't have typed that... ;)
One-space indents may very well be an artifact of idiot copy editors, and not the author.
Quoted 1 line
* typically uses one- or two-character variable names
There is one instance when one/two-character variable names make sense:
mathematical code that directly implements numericla algorithms from the text. In such a case, short variable names correspond directly to standard notation; using longer names would actually obscure the correspondence between text and code. I also don't see the point of using "array_index" over a plain old traditional "i" in a loop.
Rarely is any coding "rule" absolute. The point is clarity; if a "goto" or one-character identifier make sense, use'em.
For those who growl -- I think this kind of discussion *has* value in the kernel mailing list. Kernel newbies and such can learn a great deal from rational, calm debates among experts; if they learn, their contributions to the kernel will be better.
Of course, note the "rational" and "calm" above, which does not apply to the Stallman debate... ;)
I enjoy the implementation debates; they give me a better idea of where the kernel is going, so I can figure out where to stick my oar in the waters.
--
Scott Robert Ladd
Coyote Gulch Productions (http://www.coyotegulch.com)
Professional programming for science and engineering; Interesting and unusual bits of very free code.
Quoted 7 lines
For those who growl -- I think this kind of discussion *has* value in the kernel mailing list. Kernel newbies and such can learn a great deal from rational, calm debates among experts; if they learn, their contributions to the kernel will be better.
Quoted 6 lines
Of course, note the "rational" and "calm" above, which does not apply to the Stallman debate... ;)
I enjoy the implementation debates; they give me a better idea of where the kernel is going, so I can figure out where to stick my oar in the waters.
Yes, but how many times this discussion already happened ?
I have nothing against the discussion happening ... As long as it doesn't happen a couple of times a year. That's what a list archive is for. The latest generation of Internet users, including some latest generation developers are in general LAZY about looking up archives, they want to post the question now and get an answer, while they could have gotten the answer without bothering anyone. Sure, it takes one a little more effort to search the archives, but that's one person's effort, while posting one message wastes hundreds (perhaps thousands) of people's time. And also, the FAQ for the list states it quite clearly that the people doing usefull work here is only really interested in proof that something is better, performance and/or space wise, so if someone has an opinion, first go prove it on some piece of the kernel, the post their opinions along with the results to the list. That's the way to go. Maintenability is important, after speed, memory usage and portability. That's sound engineering, instead of C.S. dogmas.
Marcelo Pacheco
[email protected] (Scott Robert Ladd) wrote on 15.01.03 in <[email protected]>:
Quoted 6 lines
Quoted 2 lines · Kai Henningsen
Well ... he did have some nice ideas. Unfortunately, they usually don't get palatable until someone else has worked on them.
I actually *liked* Modula-2... but I haven't used it since the very early 90s. I guess I didn't like it *that* much, huh? ;)
Stuck shift key syndrome.
Quoted 5 lines
Quoted 4 lines
If you look into Wirth's books and see that the example code in there ...
* uses one(!) space indentation
* routinely puts several statements on one line
Quoted 7 lines
I've got a copy of the classic "Algorithms + Data Structures = Programs"; in it, Wirth uses four-space indentation but a proportional font. The style isn't all that bad. Hell, I've got worse stuff in some of *my* books -- oops, shouldn't have typed that... ;)
One-space indents may very well be an artifact of idiot copy editors, and not the author.
I think I've actually seen that in more than one Wirth book, but it's been quite a while.
Quoted 8 lines
Quoted 1 line
* typically uses one- or two-character variable names
There is one instance when one/two-character variable names make sense:
mathematical code that directly implements numericla algorithms from the text. In such a case, short variable names correspond directly to standard notation; using longer names would actually obscure the correspondence between text and code. I also don't see the point of using "array_index" over a plain old traditional "i" in a loop.
B-trees is the example I most remember. I once carefully typed his source in, then spent several days reformatting the source and putting in sensible variable names; then when that source was finally readable, I decided that several things in there needed doing differently - I don't remember the details, but IIRC some parameter/result passing stuff was rather poorly thought out.
If *I* were an IT prof and a student handed me code like that, I'd be sorely tempted to give that a "fail". Or at the very least a "just barely passed". Publishing it in a book means the editor fell down on his job.
Quoted 2 lines
Rarely is any coding "rule" absolute. The point is clarity; if a "goto" or one-character identifier make sense, use'em.
And so I do.
Quoted 4 lines
For those who growl -- I think this kind of discussion *has* value in the kernel mailing list. Kernel newbies and such can learn a great deal from rational, calm debates among experts; if they learn, their contributions to the kernel will be better.
That part is fine - what's bad is obvious newbies insisting they know better, two dozen times a day, and the rest of the world not treating them as the obvious kooks they are. Those are the painful parts of this thread.
Quoted 2 lines
Of course, note the "rational" and "calm" above, which does not apply to the Stallman debate... ;)
Let's just say that my ideas about successful public relations (note I'm not talking about programming here) seem to be diametrally opposite to those of rms, and leave it at that. (Ok, I feel I need to add that while I don't personally like esr, his PR efforts seem to have been a lot more successful - and I completely reject the notion that Open Source anything more than a different name for Free Software. A rose by any other name ...)
Quoted 2 lines
I enjoy the implementation debates; they give me a better idea of where the kernel is going, so I can figure out where to stick my oar in the waters.
Similar here.
MfG Kai
Quoted 1 line · Kai Henningsen
* typically uses one- or two-character variable names
At ETH, we jokingly called this WNF ("Wirth's Normal Form". Well, actually "Wirthsche Normalform"): shortening any identifier name would make it conflict with another one.
Isn't it funny how irrelevant those Pascal-family languages become, once you grok C ? :-)
- Werner
Signature · 3 lines
_________________________________________________________________________ / Werner Almesberger, Buenos Aires, Argentina [email protected] / /_http://www.almesberger.net/____________________________________________/