At Tue, 14 Aug 2007 12:44:19 -0400,
Gabriel Farrell <[log in to unmask]> wrote:
> Will made some good points that I agree with, but the truth is I started
> this conversation because we have some projects that need a home here
> and now and, for a variety of reasons, none of the hosting options out
> there are as exciting as hosting the code ourselves.  We might need to
> think more seriously about administration once we've got a couple dozen
> projects, but I say let's cross that bridge when we come to it.

It is perhaps unfair of me to be discouraging of this. If you need a
place to host code there is certainly nothing wrong with running your
own system. But if you are going to encourage others to place their
code there, I don’t think you are going to have many takers.

Here is my essential problem with this: you will be responsible from
that point on, to people you don’t know, for keeping alive a version
control system, bug tracker, mailing lists, etc. These people will be
frozen in their tracks when the system goes down, and if it goes down
for a few days or weeks, like sf.net did a while back, they will be
unable to get anything done, & they will leave, & give you a bad name.

At Tue, 14 Aug 2007 10:24:39 -0400,
Will Kurt <[log in to unmask]> wrote:
> […]
> Erik, there are several good reasons to build our own rather than
> use space available in other domains. The first and foremost is that
> the library community is big enough and specific enough to warrant
> its own centralized location for these things.

Can you name another community that runs its own code hosting system
(besides the free software community)? Why have other communities not
gone forward with similar projects?

> Another issue is that there are a large range of skills that are
> useful to library application development that simply aren't touched
> on in other areas. There are plenty of people who understand AACR2,
> FRBR, LCSH etc that wouldn't go near a place like sourceforge
> thinking there is no room for them there.

This is a good argument for code4lib, which is definitely a great
thing.

> Simple branding is another very important reason. Google the phrase
> 'library open source' and tell me if the results give you any sense
> that the library community is actively developing open source
> tools/libraries/applications/etc. to meet its needs.

This is an misleading search because of the other meanings that
‘library’ has in software. It is not clear to me how having another
site will help this problem, in any case.

> I've known a fair amount of library-staff who work on little code
> projects in isolation, who if they knew there was a larger project
> they could work on and get involved with they would (this is also
> true for the relatively large number of ex-software developers I've
> met in libraries). Snippets of code and various packages/libraries
> need to be organized and collected, but the larger aim would be to
> create a community of people interested in creating open source
> software applications for libraries.

This is a matter of getting people aware of other projects out there,
a task best served, IMO, by a news site, weblogs, etc.

If you want to build a simple system for a small group of developers
who are close to code4lib to host their projects, great! But if you
are trying to build a google code for libraries, I really want to be
discouraging, because it will be a lot of work, you won’t get what you
want out of it, & you are going to have angry users when it has
downtime.

best,
Erik Hetzner

;; Erik Hetzner, California Digital Library
;; gnupg key id: 1024D/01DB07E3