English version
Accueil     Ŕ propos     Téléchargement     Ressources     Contactez-nous    

Ce site est rarement mis ŕ jour. Pour les informations les plus récentes, rendez-vous sur le nouveau site OCaml ŕ l'adresse ocaml.org.

Browse thread
[Caml-list] Completeness of "Unix" run-time library
[ Home ] [ Index: by date | by threads ]
[ Search: ]

[ Message by date: previous | next ] [ Message in thread: previous | next ] [ Thread: previous | next ]
Date: 2004-03-18 (23:41)
From: Gerd Stolpmann <info@g...>
Subject: Re: OCaml's Cathedral & Bazaar (was Re: [Caml-list] Completeness of "Unix" run-time library)
This whole discussion is interesting because it shows that the users of
ocaml have expectations for a good language environment. There seems to
be some confusion, however, that the core ocaml product is not a
complete environment, and that third-party tools and libraries are
needed. The confusion is that it becomes unclear who can be made
responsible for improving the environment. The Unix module was mentioned
as an example.

For me, I am quite happy that the Cathedral limits itself, and that it
does not try to incorporate more components than needed. There would be
also lots of problems in a world where the Cathedral delivers all
important components, e.g. it would be harder to get bugs fixed. So
please, don't dream of that. It might turn into a nightmare.

In the case of Unix, I do not see any advantages of developing
improvements within the Cathedral, this can really be done on the
Bazaar. Maybe there are some corner cases (e.g. IPv6 sockets) that would
require cooperation, but my experience is that INRIA finally accepts
good proposals (e.g. remember of DLLs, users urged INRIA to implement

Of course, the Bazaar needs a bit more organisation, and as you know, I
wrote software supporting this: Findlib and GODI. Findlib is mature
software, even from a professional point of view, and many people are
already using it. GODI is already usable. So the technical solutions
exist (that means: THERE IS A CPAN FOR OCAML, although it uses a
different implementation for a different language, and does not match
C.*AN), and the remaining question is: Does the community accept it?

I don't mean technical acceptance here, because this is usually not the
barrier keeping people away when the software works, but acceptance in a
social way: GODI would be accepted when it would be the usual way to
install ocaml. Of course, this a quite ambitious goal, but also the
natural one for such a project. I am glad to hear encouraging words from
Xavier Leroy. 

In order to reach this goal, a number of questions should be answered
(best as some kind of community process):

- How can people participate (add packages, fix bugs, improve the base

- How can the quality be ensured?

- How are decisions made?

- How can the platform be kept open?

My vision for GODI is that it limits itself to a platform for software
authors, i.e. some kind of distribution channel, and organises things
only where necessary (the question of policies). I don't want to build a
second Cathedral that fixes the errors of the first. (So there will be
place for an alternate stdlib if people want it, but it won't replace
the official one.)

Expect an announcement about a mailing list for the details of this
discussion soon. Of course, we can already start on caml-list, it's you
who decides.

Gerd Stolpmann * Viktoriastr. 45 * 64293 Darmstadt * Germany 
gerd@gerd-stolpmann.de          http://www.gerd-stolpmann.de

To unsubscribe, mail caml-list-request@inria.fr Archives: http://caml.inria.fr
Bug reports: http://caml.inria.fr/bin/caml-bugs FAQ: http://caml.inria.fr/FAQ/
Beginner's list: http://groups.yahoo.com/group/ocaml_beginners