Recent blog entries for dlehn

OpenID

A long time ago I had some crazy idea I'd have time to work on a single sign on distributed identity system. Never happened. But I did register openid.net and openlogin.net around June 2001. They are near to expiration and have had a note offering them to anyone for a long time.

As luck would have it, Brad Fitzpatrick, of LiveJournal fame, started up a similar interesting looking project. They were looking for a name and managed to email me about openid.net right before I was going to offer it to them. So I gave it to them for the new and improved OpenID project.

There are lots of similar projects out there with lots of good ideas. None of them have taken hold on more than a few test sites if even that. Maybe the momentum behind LiveJournal will be enough that this system can grow and be actually used by various sites on the net. Hopefully their traditional OpenSource policies keep this stuff as open as possible.

I'll just chalk up giving away the domain as my community service act for the day.

Namespaces

Speaking of distributed id services... it has become seriously difficult to pick user ids on popular sites. It's really just a namespace problem where the entire world has to pick a short, cute, memorable nick at one central spot. See AIM, gmail, Yahoo!, /., IRC, and of course weblogging sites like LiveJournal. Just about any short nick related to your name, initials, or id you use elsewhere is bound to be taken along with 20 variations thereof. Would be so much easier to use email ids everwhere. Then at least the namespace bottleneck is limited to the current domain name namespace issues. Sigh.

I figured I'd register a LJ account just to try the OpenID ideas. Of course just about all the nicks I tried were taken. Oh well. So in the tradition of i18n and l10n names, I got d11n

Uraeus had a problem with his RSS Feed due to an empty <item/> which FEED Validator was nice enough to find. Advogato's XML-RPC Interface along with Python made this easy to fix. Just authenticate and add the missing entry 379. All in the interactive Python shell. Neat. This post will test to make sure the fixed RSS feed also fixes Planet GStreamer. It was always showing the empty entry on top as a empty post with the current date.

Debian

Became a Debian Developer back on 2003-04-13. Only took 815 days to get through the process. I've been maintaining GStreamer related packages. Often takes a whole lot of time to fix small little bugs. Packages take forever to build when all the dependencies are installed. Has been interesting volunteer work.

GStreamer

Project keeps moving forward. Good to see that. I finally got around to releasing gst-python the other day. It's been working for months. Getting motivation to do the initial release cleanups was hard. Decided to make it entertaining and autogenerate all the standard docs, README, AUTHORS, NEWS, etc from DocBook source. In theory this will make future releases trivial. We'll see about that. Writing GStreamer applications with Python seems to work out fairly nice. Various tricky things to be aware of but I think it's easier than using the C API. Need to work on a tutorial.

Did I mention I'm back in grad school as a PhD student at Virginia Tech? I had a enjoyable, yet brief, encounter with the work concept at Intransa, Inc. after getting my MS. Back to the .edu world for now. A quote I heard from one of my professors: "Grad school is the snooze button on the alarm clock of life." True.

I've been busy with Debian things lately. GStreamer is all packaged up and headed towards an archive near you. Managed to package the core, the plugins, and the player and editor too. The core and apps are not so bad but Debian packaging is rough for plugin based frameworks that depend on a million libraries. Didn't take too long to get it done right. The ITP Bug has only been open since Feb 2001. ;-) Perhaps someday I'll actually finish up the new maintainer process and become a true Debian developer. Started that in Jan 2001. Nice and leisurely pace around here.

Also have been working on Python bindings for GStreamer. The PyGTK tools work great! Just some nasty threading issues to solve before the bindings are ready for consumption.

Next up after that is more work on GStreamer DVD support. Then maybe I can actually write a DVD player with GStreamer. Which is why I started working on the project in the first place years ago.

Good luck to wingo during his Peace Corps adventures!

Seems I haven't been using this diary too often. OMS seems to be dead now. Actually, the whole LiViD site vanished into the void. Kind of a shame really. DVD court cases not going well either. And yet, players like Ogle and MPlayer and VideoLAN and Xine are doing just fine. Odd world.

Since my last entry I've mostly been working on GStreamer. Along with omega, wtay, thomasvs, wingo, dobey, Uraeus, and others. Nice project. Nice people. Doing the GStreamer Debian packages is a toss-up between releases. Sometimes it sucks and sometimes takes 20 minutes. 0.4.0 is sucking. Oh well. Hopefully I'll get it done and into Debian proper soon.

I messed up during GStreamer 0.4.0 release... I made thomasvs quite mad cause of last minute commits between final pre-release and release (that's me he's talking about in his diary). I'm sorry! Just trying to squeeze last minute fixes in to imporve the release. We've traditionally only had very long release cycles (months). So I tried to get stuff in this one. Release process is kind of hard in open source world. Sometimes the developers free time to work on the project doesn't match up. Oh well. We'll likely have a stricter release branch policy next time to stop me from being an ass. Sound like a good plan.

I'm missing out on the GNOME Summit in Boston this week. Which sucks. Could have met wingo, thomasvs, dobey, and other GNOME and GStreamer hackers. Always nice to have a beer with fellow coders. (And I could have appologized in person to thomasvs for making him mad.) Pesky work getting in the way. Bah. My priorities are messed up.

Well, I just finished polishing up and releasing the first version of OMS. The LiViD crew has been working for more than a year on this DVD player.

It was harder than I thought to get a release done. We needed to get something out the door at some point. Hit a relatively stable area so we decided to do it. Have new features coming down the pipeline immediately. So the coders want a release done quick. But of course the builds are all broke and docs need work. I had to fix some myself and rally the troops to fix other stuff. Then everyone dissapeared just in time to do all the final polishing work. ;-) Oh well... got it done... just took way too much time. Hopefully future releases will require less work.

lessons:

• handling final release details sucks
• writing docs sucks
• fighting build systems sucks
• bug hunting can be fun compared to other project details ;-)