Hi Nicholas (and everybody)! We are getting off topic here, but it's still an interesting discussion.
On Tuesday 17 April 2007, Nicholas Clark wrote: > On Mon, Apr 16, 2007 at 08:45:37PM +0300, Shlomi Fish wrote: > > On Monday 16 April 2007, brian d foy wrote: > > > Mac OS X isn't "mostly compatible with UNIX", it is a *nix. > > > > > > Use.perl is missing from the Web Forums link > > > > Thanks for you commentary. I disagree about the Mac OS X one - some > > things about Mac OS X differ from most other Unix-flavours, including the > > absense of most of /etc, their case-sensitivity problems, their extended > > file properties (that have often broken perl Makefile.PL) etc. I hate the > > term *nix because > > eg the use of dylib rather than dlopen (prior to 10.4), Because dlopen is > far less restrictive in what it allows. Given that most open source > software hails from Linux, Solaris or *BSD where dlopen is used, it ends up > accidentally taking advantage of dlopen's laxness. > Interesting. Well, I've started concentrating facts and links against Apple here: http://www.shlomifish.org/open-source/anti/apple/ Email me in private if you have any additions. > > no one understands what it is, while "Unix-compatible, a Unix-clone, a > > Unix-flavour, etc." is much clearer. > > It's a FreeBSD fork atop a Mach microkernel. > > It's pukka Unix, more so than Linux actually is. As I understand it, Linux > doesn't derive code from the original AT&T source tree, whereas FreeBSD > does: > > http://www.freebsd.org/doc/en_US.ISO8859-1/books/handbook/history.html > Yes, I know Linux has all of its code rewritten from scratch. However, it doesn't mean it's not UNIX-compatible. If someone writes a Perl 5-compatible implementation from scratch and it passes all the relevant tests and behaves according to perl*.pod it would still be a Perl implementation. That's just like there are (too) many Scheme implementations that are all mostly R5RS compatible. > Linux neither uses any of the AT&T code base, nor passes the Unix trademark > conformance tests (unlike, say, z/OS) so isn't. Like it or not, nowadays Linux has become the de-facto standard due to its popularity, feature-set, robustness, extensive hardware support (second only to Windows), portability (reportedly even more extensive than NetBSD), etc. etc. It seems that the Linux Standard Base will define how future Linux-compatible operating systems will look like. UNIX as a concept is still very much alive, but Linux or alternatively "*BSD" is what everybody uses and write for. If you know what you're doing, you won't have problem writing code that works on most modern UNIX-like systems, and also potentially Win32/Win64 or even more exotic systems. However, in most cases, you can write a program that will only run on GNU/Linux, and most of your end-users will be happy. > > Whatever term you choose to use, for whatever reason, does not make you > correct when the facts are against you. Right. > > Unix-flavour would be an accurate description, and you are correct that it > is not that similar to many other flavours of Unix, which has practical > implications, which you are wise to note. But clone it is not. > Hmmm... I hate picky semantics' nazism. But I guess I can phrase myself better. When I say something is a UNIX, I don't mean it got the UNIX trademark for the UNIX class of operating systems. I just mean it resembles Linux, the BSDs, Solaris, etc. As you know UNIX flavours come in all shapes and sizes, and have different philosophies and different ways of doing things. But they all share a common paradigm. Regards, Shlomi Fish --------------------------------------------------------------------- Shlomi Fish [EMAIL PROTECTED] Homepage: http://www.shlomifish.org/ If it's not in my E-mail it doesn't happen. And if my E-mail is saying one thing, and everything else says something else - E-mail will conquer. -- An Israeli Linuxer
