Comments on: Everything is Unix http://www.linux-mag.com/id/7569/ Open Source, Open Standards Fri, 10 May 2013 08:56:11 +0000 hourly 1 http://wordpress.org/?v=3.1 By: wcn00 http://www.linux-mag.com/id/7569/#comment-7139 wcn00 Wed, 30 Nov -0001 00:00:00 +0000 http://www.linux-mag.com/id/7569/#comment-7139 <p>Most of your examples seem to stem from the language in question lacking proper threads support and, in the case of the serialization issue, a proper mutex implementation for threads. I\'m very much in favor using low level facilities when the higher language lacks them, I do it all the time, but the rate at which you have to do it is usually indicative of a weakness in the high level language.<br /> I don\'t think its necessary to resolve those weaknesses because many of these interpreters and compilers were written for specific uses, but if you find yourself stepping outside the language to accomplish crucial activities, then perhaps you\'ve chosen the wrong language in the first place... no?<br /> wcn </p> Most of your examples seem to stem from the language in question lacking proper threads support and, in the case of the serialization issue, a proper mutex implementation for threads. I\’m very much in favor using low level facilities when the higher language lacks them, I do it all the time, but the rate at which you have to do it is usually indicative of a weakness in the high level language.
I don\’t think its necessary to resolve those weaknesses because many of these interpreters and compilers were written for specific uses, but if you find yourself stepping outside the language to accomplish crucial activities, then perhaps you\’ve chosen the wrong language in the first place… no?
wcn

]]> By: lescoke http://www.linux-mag.com/id/7569/#comment-7140 lescoke Wed, 30 Nov -0001 00:00:00 +0000 http://www.linux-mag.com/id/7569/#comment-7140 <p>All of the standard Unix (POSIX) system calls are well supported on most platforms. The only notable exception that comes to mind is Windows where POSIX compatibility is lack luster.</p> <p>I agree the Stevens book should be on every serious programmer\'s shelf.</p> <p>Les </p> All of the standard Unix (POSIX) system calls are well supported on most platforms. The only notable exception that comes to mind is Windows where POSIX compatibility is lack luster.

I agree the Stevens book should be on every serious programmer\’s shelf.

Les

]]> By: ndatta http://www.linux-mag.com/id/7569/#comment-7141 ndatta Wed, 30 Nov -0001 00:00:00 +0000 http://www.linux-mag.com/id/7569/#comment-7141 <p>Nice article, but several grammatical errors.<br /> - ND </p> Nice article, but several grammatical errors.
- ND

]]> By: bridget99 http://www.linux-mag.com/id/7569/#comment-7142 bridget99 Wed, 30 Nov -0001 00:00:00 +0000 http://www.linux-mag.com/id/7569/#comment-7142 <p>I agree wholeheartedly. Also, there are advantages to old-style UNIX multiprocessing (fork() and spawn()) even in languages that do support threads. Newer programmers today tend to think that threads are the only available abstraction for multiprocessing. That\'s not the case- fork() and spawn() are not just older and more widely available than multithreading, they\'re also safer. </p> <p>Assuming one CAN do something with processes (vs. threads), then I think one SHOULD... even if this results in process-spawning code which more up-to-date programmers (erroneously) consider awkward. </p> <p>I work in a pure Win32 environment but make heavy use of things like Cygwin and MinGW just to inject some sanity into my work. Not so long ago I \"fixed\" something by recompiling it using MinGW instead of Visual C++. My \"audience\" probably has very little concept of how much of my work is actually resting on the concepts- and at times the code base - of UNIX. </p> I agree wholeheartedly. Also, there are advantages to old-style UNIX multiprocessing (fork() and spawn()) even in languages that do support threads. Newer programmers today tend to think that threads are the only available abstraction for multiprocessing. That\’s not the case- fork() and spawn() are not just older and more widely available than multithreading, they\’re also safer.

Assuming one CAN do something with processes (vs. threads), then I think one SHOULD… even if this results in process-spawning code which more up-to-date programmers (erroneously) consider awkward.

I work in a pure Win32 environment but make heavy use of things like Cygwin and MinGW just to inject some sanity into my work. Not so long ago I \”fixed\” something by recompiling it using MinGW instead of Visual C++. My \”audience\” probably has very little concept of how much of my work is actually resting on the concepts- and at times the code base – of UNIX.

]]>