http://www.perlmonks.org?node_id=177362

Okay, this is more of a perl culture question than a technical one, but I'm in need of some wisdom nevertheless, so here goes.

I recently posted an article to Usenet asking if anyone would be interested in helping me do a rework of SourceForge.net in Perl. The goal was to take what's good about SF, do it better in Perl, and add other good stuff. The result would be intended to be more easy to install locally (SF is notoriously hard to install on your own server) on an intranet or on the internet (like SF.net is set up now).

See this for the post.

After posting it, though, I got surprisingly little response. I mean, maybe everyone's working on their own projects, but I thought that at least some people would find this interesting enough to talk about. The idea of making an SF-like system that is easier to install and use, AND doing it in Perl, is exciting to me.

What do you think? Am I just easily excited at stupid ideas, or do my Usenet posts have a bad smell?

jpt

edited: Wed Jun 26 15:22:18 2002 by jeffa - shortened link text

Replies are listed 'Best First'.
Re: Why such little interest?
by lachoy (Parson) on Jun 26, 2002 at 13:19 UTC

    There are so many projects out there competing for developer interest and time, you need something unique to hook them in. IME, the hands down absolute best hook is working code, followed closely by easy installation, solves a common problem, documentation with examples and then a little farther back by demonstrated developer commitment and spreading the word. The smartly designed hook is IMO somewhat overrated -- it's important, but once the other hooks sink in, developers feel like they've invested sufficient time to either overcome it or help do it better.

    So the first thing: state what you want to do, then create a really simple working version of it and make it painless for someone to download, install and get running. ("Painless" defined in less than 15 minutes.) Then keep building on what you've got running, always ensuring that there's a miniscule barrier to entry.

    This initial version may be thrown away later, but the central idea is that you need to show rather than tell people your vision.

    Chris
    M-x auto-bs-mode

[reply]
Re: Why such little interest?
by vladb (Vicar) on Jun 26, 2002 at 12:30 UTC
    If you check SF.com you'll notice that out of a godzillion of projects only a handful thrive. It's only natural that over the course of time people loose interest in certain projects. There's a lot of reason for this to happen. Lately, I couldn't find time to work on some of my projects since I'd started accepting consulting contracts. I believe there's a lot of folks in the perl community who share a similar story. The need to earn extra $$ (to support family, that new car thing, and etc) at times overwhelmes selfless desire to volunteer to a good cause.

    However, don't be let down by apparent lack of interest. My advice to you would be to proceed with the project even if you don't get a lot of initial responce. Start out by writing some functional specifications for the final system as you see it. There's a greater chance of attracting people to a project that already has something to play with. It'll also demonstrate a certain level of dedication on your behalf. Being in a position of project initiator, this is a rather important step to make. I find it that in the open source community there's nothing worse than to direct a lot of time and effort to a project that is bound to fail or remain incomplete. Well, a thorough design documentation will serve good to clear those doubts. ;)

    _____________________ 
    # Under Construction

    [download]
[reply]
[d/l]
Re: Why such little interest?
by cjf (Parson) on Jun 26, 2002 at 12:40 UTC
[reply]
Re: Why such little interest?
by perrin (Chancellor) on Jun 26, 2002 at 18:16 UTC
    I suspect most people don't see much value in the SourceForge software. I mean, all of the important stuff is just stock open source software. The real value is that they host it and I don't have to set it up. If I really needed an in-house system, I would install CVS, create a mailing list on the company mail server, and install that Perl script for browsing CVS. Seems pretty straightforward to me, and doesn't require a new development project.
[reply]

      By far, the best CVS viewer I've found is the ViewCVS Python CGI script. It's dead simple to setup and everything works like you'd expect. Since CVS viewing is fairly orthogonal to other project management activities, the fact that it uses a different technology (albeit one accessible by constructing GET requests) has never been a problem for me.

      Chris
      M-x auto-bs-mode

[reply]

      I installed SF internally on a client site, and the developers are going ape over it. They love the community building aspects (ratings/code snippets) and were already needing some sort of project management/bug tracking system. SF beats Bugzilla hands down, in my opinion.

      I think SF does add value by tying all those disparate software packages together into a simple user interface. I want to do the same things, only better and more easily.

      And I want other organizations to be able to host internet-based development just like SF.net does. See savannah.gnu.org, who took SF's source and modified it painstakingly because they needed similar functionality. It shouldn't be so hard!

      jpt

[reply]
        Well, it just sounds like you're in the minority. There are lots of great bug tracking tools out there like RT, and I think Bugzilla is fine although of course you have to customize it for your organization. I'd rather see effort go into imporving those tools than writing new ones.
[reply]

Back to Meditations