Tue, 24 Jan 2006

GNOME.conf.au

Blogging live from the GNOME mini-conf at Linux.Conf.Au 2006, Dunedin, New Zealand!

Good start to the day. Jeff Waugh gave the 10x10 talk he gave at GUADEC last year — but also incorporated a lot of the feedback and views he’s heard since then which made it interesting even a second time ‘round. [Not to mention it being new to most of the Aussies and Kiwis in the crowd]. It was quite refreshing to hear Jeff being candid as opposed to blindly enthusiastic (because that’s the current party line or whatever).

A good conversation about release cycles got going. One thing that puzzled me at GUADEC and that continues to puzzle me is the recognition (I’m paraphrasing here) that “while the time-based release cycle has led to quick availability of quality software, it’s meant that visionary innovations (an agenda) have perhaps been left behind” … but after discussing it for a while, the conclusion was again “we will therefore stick with the 6 month release cycle”. Huh? If you don’t change the system, why would you expect a different outcome?

The sys-admins in the crowd were fairly unified in expressing their frustration at releases being too frequent, with the result that by the time they got a chance to test a released version set, qualify it for their environment, work out a migration path, and deploy it on their site, GNOME would have moved on at least one release and the hackers themselves would by then be working on the release after that, meaning that bug fixes and polishing for the version our poor sys admin was trying to roll out on their site run would never really happen. (ie when was the last time any of us did a bug fix to GNOME 2.10 branch code, let alone 2.12?).

Lord knows I spend enough time in my professional life going around the world warning people to beware the seductive dangers of incremental change, so it’s difficult for me to be objective about this issue. It strikes me, though, that the core GNOME hackers who gained so much from switching to the 6 month cycle in the first place are having a hard time figuring out how to recapture the initiative without losing their productivity. My impression is that the religious adherence to the 6 month time-based release cycle is a structural issue which is holding GNOME back from making the visionary leaps it wants to make.

Later in the morning Glynn Foster did a live demo of dtrace on OpenSolaris. I’ve been hearing people go on about .d for ages, so it was great to see it in action [when we were chatting over beers in Bangalore at FOSS.IN, Rasmus mentioned that he maintains a Solaris box just so he can use dtrace to profile PHP. Great endorsement!]

Someone I hadn’t met before came up to me at lunchtime and said “hey, I use your bindings”. Cool! And while I quickly set him straight that the java-gnome project is certainly not mine, it’s great to listen to people describe their experiences using software I contribute to.

AfC

2006-GNOME-conf-au, posted by Andrew Cowie in /andrew/conferences/new-zealand at 05:25, 24 Jan 2006.

Category Specific Feeds. Use these links for an RSS or ATOM feed limited to this category and its descendants. Technorati Profile