From mboxrd@z Thu Jan 1 00:00:00 1970 X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on polar.synack.me X-Spam-Level: X-Spam-Status: No, score=-1.9 required=5.0 tests=BAYES_00 autolearn=ham autolearn_force=no version=3.4.4 X-Google-Thread: 103376,6511c3dc6e1155c9 X-Google-Attributes: gid103376,public X-Google-Language: ENGLISH,ASCII-7-bit Path: g2news1.google.com!news1.google.com!news.glorb.com!newsfeed1.ip.tiscali.net!tiscali!transit1.news.tiscali.nl!dreader2.news.tiscali.nl!not-for-mail Newsgroups: comp.lang.ada Subject: Re: SourceForge, CVS and revision control systems (was: GWindows and David Botton) References: <2004100312401227544%david@bottoncom> <87zn31njc6.fsf@insalien.org> <10m5opffddc6e4d@news.supernews.com> <87r7odnhv0.fsf@insalien.org> From: Ludovic Brenta Date: Tue, 05 Oct 2004 22:27:01 +0200 Message-ID: <87mzz0oqqy.fsf_-_@insalien.org> User-Agent: Gnus/5.1006 (Gnus v5.10.6) Emacs/21.3 (gnu/linux) Cancel-Lock: sha1:XQLYBR0hvTZzfjZPX7ep2ok9HTw= MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Organization: Tiscali bv NNTP-Posting-Date: 05 Oct 2004 22:30:34 CEST NNTP-Posting-Host: 83.134.237.32 X-Trace: 1097008234 dreader2.news.tiscali.nl 44092 83.134.237.32:37868 X-Complaints-To: abuse@tiscali.nl Xref: g2news1.google.com comp.lang.ada:4753 Date: 2004-10-05T22:30:34+02:00 List-Id: Pascal Obry writes: > Ludovic Brenta writes: > >> Well, I have mixed feelings about this. Despite its known and >> annoying flaws, CVS was popular when SourceForge was established. >> SourceForge chose to support CVS for its simplicity. Now, the >> popularity of SourceForge makes CVS popular, but the deficiencies >> in CVS are still there. > > That's wrong! CVS has always been very popular AFAIK. It has > superceded RCS and SCCS long time ago. I agree that CVS is not the > best version control system around but it is quite good, and it is > good enough for some very large projects ! If you think that the deficiencies in CVS are not serious, think again. No atomic commits, no recording of merges, no versioning of directory structure, no replication, poor scalability (in terms of performance) come to mind. I agree that CVS is "good enough", and I am aware that some very large projects use it succesfully. But that is not my point. My point is twofold: - CVS has known flaws. These are design, not implementation flaws and only a redesign can fix them. There are several projects with the stated purpose to replace CVS with better designs. To illustrate: the name "subversion" is not an accident. - SourceForge helps maintain or even increase the popularity of CVS despite its known deficiencies. The main reasons why CVS is still so popular despite its problems and the concerted efforts by some groups to replace it are the installed base, inertia (especially of large projects, by their very nature), and its simplicity. The large projects that use CVS do so at the cost of extra work on the part of the developers. For example, CVS does not help prevent multiple merges of the same changes into a branch. A skilled release manager is required to perform this function manually, when it could be performed automatically and reliably by the tool. The parallel with the C programming language is quite obvious to me. C replaced assembler a long time ago, and it is still being used massively despite its known problems. The large projects that use C do so at the cost of extra work on the part of developers. For example, C does not help prevent buffer overflows. A skilled programmer is required to perform this function manually, when it could be performed automatically and reliably by the compiler. -- Ludovic Brenta.