Laughing Meme

I’m not really a web forums type guy, but the light weight interfaces of bbPress and Vanilla do a lot to redeem the medium. Anyone got a preference from using them?

Also anyone have a mbox import script for importing old list traffic?

Part of the very slow, long running process of re-launching the Magpie site, and breaking my Sourceforge dependency.

( 2004-09-17 10:27:57 – Project CVS Service ) As of 2004-09-17 there is a problem with the CVS host serving anonymous CVS, nightly tarballs and ViewCVS for projects that start with the letters m, n, p, q, t, y and z. We currently do not have an estimate on when they will be back up.

Interestingly projects beginning with m, n, p, q, t, y, and z account for over half of my Sourceforge projects.

Sourceforge is in a death spiral, anyone who lived through late 90s/early 2000s should recognize the symptoms: increasingly cluttered site design, outages, new money making schemes, and an increasing divorce from reality/constituents.

update: oh, and changing your name is another symptom, because you know OSTG is so much better then OSDN. But they’re doing something right to get funded by Omidyar (major Groundspring funder).

I didn’t realize DOAP wasn’t out when Jo was proselytizing it at the Aspiration convergence earlier this month. If I’d know that I’d have jumped on it, you know be one of those early mover cool kids. Now it is released, with a web page, and a OSCON session + Danny O’Brien write up.

My initial feedback is that we’ve got to stop being so damned cute with our naming if this stuff is going to get wide adoption, FOAF, LOAF, DOAP, come on!

Still, Magpie DOAP (hmm, have to see if I can use .htaccess to force proper content-type), courtesy of the DOAP-a-matic. Really Sourceforge should spit this out automatically, in the mean time it would be a trivial hack to write a Sourceforge RSS to DOAP converter. (that was a hint)

I noticed this morning that Magpie just crossed the 20,000 downloads threshold (yay for Sourceforge RSS), including 10,000 downloads in the last 6 months. Even more telling is the real explosion of interesting new services, articles, and projects all using Mapgie. (I’m trying to keep track of some of this Magpie links page, and some it on del.icio.us/tag/magpie which is syndicated to the Magpie blog)

With Steve’s recent work we might finally kill the encoding issues which has plagued Mapgie since day one (anyone want to volunteer to come kill the encoding issues plaguing XML::RSS?).

And perhaps most surprising to me was a copy of Uncommon Grounds showing up at my door this weekend apparently purchased from my Amazon wishlist. (which I had forgotten I had, and to the best of my knowledge have never linked to, and is horribly out of date) Wow. This looks like a great book. Thank you.

And thanks to everyone who has been participating so far.

Sourceforge is apparently still trashing around desperately trying to find a way to monetize what can only be a very resource intensive site to maintain. First they added the Sourceforge managed donation service (wonder how that is working for them?) and then the recent “Premium Sourceforge Subscriptions”.

Both of these services seem to miss the obvious, Sourceforge real customers are project maintainers, we are the one with whom it has a relationship, and it is to us that it is offering real value. Obviously its compelling to try to tap into the millions of users who must hit the site daily rather then the mere 75,000 or so of us who maintain projects, but I think it is misguided.

What They Are Offering

Better search tools for finding projects, better tech support, and better project monitoring, and better downloads. Except for tech support these seem like tools aimed at consumers, not creators.

What I Would Pay For

Better metrics and statistics. For example, Sourceforge hasn’t updated the download numbers since Jan 31st. I’m afraid this is going to be like the 6 month outage that SF.net went through last year (or was that 2 years ago) when they failed to report 95% of page views and downloads. Also I’ve added a little web bug to keep track of web stats for my Sourceforge hosted site, but I might be willing to pay for something that did this for me.

Better CVS server. The CVS server is pretty consistently slammed, and slow. A being hosted on the “premium” CVS server would be very appealing.

Better web hosting. SF offers a pretty complete web hosting, but the one thing I wanted to do (open a socket from PHP) wasn’t possible.

Things I Wouldn’t Pay For But Which Should Be Done

Better mailing list archives. I don’t understand why they don’t just use the mailman’s archiver (though it admittedly has issues), but the Sourceforge mailing list archives are worse then useless.

Hosting a project with Sourceforge is great. They provide bandwidth, CVS with anonymous access (something I never set up on my own boxes), and above all, a certain credibility. But they don’t provide good traffic analysis. There was some recent discussion if people were actually finding Magpie searching for PHP RSS parsing, or if they were simple searching for a picture of a magpie.

So here is a bit of Javascript that will fetch a remote image from your Sourceforge project home page and tack the original referer url to the image.

<script language="javascript">
var img = "http://example.org/bigbrother.gif"
var ref = parent.document.referrer;
document.write("<img src=\""+img+"?"+escape(ref)+"\">");
</script>

I imagine this is what someone like SiteMeter does.

Sourceforge unveiled a major RSS upgrade today. You can get feeds for everything that you would expect, site wide popular, recent, newly released, newly approved, and project specific new releases, new documentation, new news.

They also provide a super cool, thinking-outside-the-box project summary feed, with the project description and various activity monitors(for example, Magpie’s Project Summary Feed) Everything you need for a little remote project display.

Noticeable in its absence is a CVS checkins feed, but that is rather easy to solve.

Also Sourceforge is looking for the ultimate Web guru to move to that hell hole which is Fremont. If you have

5+ years of development experience on high end, high volume websites (3+ million page views a day) [and] as a vast level of knowledge of Internet technologies: PHP, PostgreSQL, MYSQL, DB2, Linux, PERL, Apache, LDAP, Mailman. A flare for design / UI is a bonus. [ed. you can say that again!]

…and are willing to live in Fremont, then you should apply! I want someone I know working there.

I wonder what Sourceforge’s cost per project is? Obviously it varies. I’m sure phpMyAdmin or Gaim which have been in the top 10 most active projects for the last few years, have a much high cost associated them my little project, which in turn is much more resource intensive then the hundreds of projects which died stillborn, a name, and a blurb, and nothing else. Still, I think we should be able to calculate some reasonable average, and I would be curious to know what that is.

Because I’ve been noticing that Sourceforge, particularily the CVS server, has been having problems lately, and I, for one, am worried. I think this is the time when we start looking to new models, and stop relying on rich companies’ philanthropic impulses to support our community.

Would $5/year cover it? I wouldn’t hesitate to pay $5 a year to know that I can count on SourceForge sticking around as I start to build a project’s identity around it. I might go up $10. At $15, or $20 I would have to think if it was worth it to me, as I have the ability to provide everything Sourceforge provides trivially, and I use it as a convience to

• minimize resource consumption on my own boxes
• because I don’t want to enable pserver, and anonymous CVS on my own boxes
• because a Sourceforge URL lends a certain credibility, below www.projectname.org, but definitely above, www.somehost.com/~user/projectname.

The End of Free

“For free” is a sickness. For free is as damaging, dot-com, and capitalistic, as the ridiculously inflated prices we saw during the boom. It doesn’t respect labor, it cordons off the public into powerless consumers rather then invested users, and it doesn’t build sustainable alternatives. Now I don’t think the people behind SourceForge comes at it with much of a critique, but they’ve built an incredibly useful service that I would like to keep around.

Combatting Beginning of the End.

There are 2 problems that anybody who starts charging for a previously free service are going to encounter.

• First, some people get huffy when asked to pay for something they had been taking for granted. Ignore them. Its temporary. They’ll adapt.
• Much more seriously, for those of us who have been living with the web for a long time, moving to a subscription model has the unpleasant sound of a death rattle. Nothing shakes the confidence in a project more then starting to charge. You start thinking,

  “Wow, they must be hurting. If I pay them my $XX and they go out of business in 6 months, then what? I guess I should start looking for alternatives.”

  Salon is a good example of this. This sort of thinking also hurts the ability of projects to solicit donations as well.
  

  “They are down on their knees begging, and I would be happy to pay $XX to make sure they are here tomorrow. But unless XX,XXX faceless other users to the same, I’ve thrown good money down the drain.”

It doesn’t have to be this way, but at this point we’re skittish.

Financial transparency would help

If, for example, Sourceforge sent a note to each project maintainer saying:

• Our annual burn rate is: $XX,XXX
• On operating costs break down like so:
  • bandwidth
  • people
  • hardware
  • support
• We’re doing what we can to make sure we run a tight ship, but
• Each project costs on average: $XX.XX
• If you’re activity profile is about 90% you actually cost more like: $XXX.XX
• We currently have funding to run for 10 months, and if our current business model goes well, we’ll be fine.
• But, in acknowledgement that SF has become an important public resource, we’re asking that our community come together to support us.
• In return, we commit to making sure SF will be here for the long haul.

Some Refinements

It would be up to the folks behind SF to figure out how they need to cover their costs. But I’m inclined to say charge a flat rate per project. More popular projects are clearly serving the community, no reason to punish them for their popularity. Or if that just doesn’t make sense when you look at the numbers, break it into tiers.

Consider charging projects a flat rate for project administration services, but small bandwidth charge for website hosting.

I don’t know. Does that make sense? Am I worried needlessly? (and if so, why does CVS keep timing out?) How much would you pay?