From: David Vallner Date: 2006-10-17T05:26:49+09:00 Subject: Re: What is the reason for this syntax? --------------enig6A5C6F6EFC1D0B42C82AD451 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable Kevin Olemoh wrote: > Who said you had to force anything? You did? > Why do you assume that in all > cases having multiple dialects is beneficial whether we are talking > about operating systems or things we use to communicate with people. > Its a flase assumption in reality, there are languages that exist that > do not have a concept of "past" like there is in eg Romance languages > so it becomes almost impossible to try to express the concept in that > language. > Look, this wild sidetracking only to win a virtual Being Right contest is getting a little tiresome. You've gotten past the stage where you're even remotely connected to reality and turned to just spewing random unrelated anecdotes to try to support your opinion as fact. (Despite the fact there is no logical connection between those anecdotes and your poin= t.) You presented your opinion, people disagreed with it, you went on to gather flimsy "proof" to support it, most of the replies you had to that proof were for the sake of argument rather than because anyone actually got to doubt his opinion and lashed back out in a defensive reflex like you keep doing. So... Take it like an adult already and admit that your opinion is just that, and not a religious maxim that needs fervent upholding, and just drop it already - everything that was to be said about the topic has already been said. (And then some.) But since I went on and started this post: > The situation with binary formats is just plain ridiculous and there > is no way to justify a spat between maintainers creating a situation > where repositories for the same distribution (dialect) are > incompatible with each other! (Fedora had this problem for quite some > time.) Had there already been a standard binary type for Linux as a > whole the problem would have never gotten as bad as it did with Fedora > or has with Linux in general. Its fine to have a bazillion binary > formats for all I care however Linux as a community should have one > that will just work more often than not on any Linux system. Linux has a standard binary format: ELF. Statically-linked ELF binaries will work across distributions on the same CPU architecture just fine. Dynamically-linked ELF binaries will work across distributions on the same CPU architecture, provided their dependent shared libraries are present in the required versions. Now, I presume you have gotten your terms confused, and actually meant package formats. Which are RPM and DEB for the most notorious ones. As I already mentioned before, RPM and DEB packages are distribution-agnostic, they just happen to often be incompatible across distributions. RPMs more so than DEBs, most probably because there are only two major Linux distributions using DEB, and Ubuntu is at its core a debian-testing snapshot with polish, support, and marketing; the RPM distributions are more independent, . This is anecdotal (if frequently seen), unrelated to any inherent quality of the package formats, and any standardisation on the front wouldn't do anything for being able to use a FC package on SuSE. (Mind you, I have used FC RPMs on SuSE, Debian DEBs on Ubuntu, and SuSE RPMs on Ubuntu, Successfully. Your point, whichever it is you're trying to make by repeating the packaging system argument again, is moot whichever way you look at it, but I can't let that stop me, can I?) RPM doesn't have the concept of a "repository". Neither does dpkg (DEBs). So once again, you might be confusing terms, whether deliberately, or out of ignorance. And unsurprisingly, you're wrong. Again. The more prominent "repository" bits on top of RPM are YaST, yum, and urpmi. As far as I know, they are technically independent of the underlying package format, as long as the repository can provide the necessary metadata for resolving dependencies. Repository metadata formats don't map 1:1 to package formats - being completely unrelated to each other might, or might not be why. (Guess which.) There is nothing that says that a software installer tool has to be bound to a given package format. Or heck, to a given repository type. To wit: apt-rpm, to handle RPMs from an APT repository. Or smart, able to handle DEBs, RPMs, and TGZs from 11 types of repositories. Oh, and one more nail in the coffin. The standardisation of package formats is barely seen because both DEB and RPM are fairly new, worse yet for the software installation tools. Oh, and RPM is the standard package format for distributions that follow the Linux Standard Base. See how fun the world is when you know your facts? I'll follow Austin's example now and stop feeding the troll, who, doubtlessly enraged by the odd ad hominem jab, will squirm pathethically and try to "prove" that his point, despite being dissected and demonstrated to be wrong, is still by some circular logic valid. David Vallner Popcorn for the show, anyone? --------------enig6A5C6F6EFC1D0B42C82AD451 Content-Type: application/pgp-signature; name="signature.asc" Content-Description: OpenPGP digital signature Content-Disposition: attachment; filename="signature.asc" -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.5 (MingW32) iD8DBQFFM+sCy6MhrS8astoRAoXQAJ0Ri9TLubLrxt+W+br3dlT5e/aicACfX5a6 tyr2IxEBtpDrgddm3fY5bDM= =+Xsr -----END PGP SIGNATURE----- --------------enig6A5C6F6EFC1D0B42C82AD451--