Robert: ‘All a distributed VCS does is make it easier to return changes to the project’ is an irritating statement to me. It is just part of the story. Yes, good VCS makes getting changes upstream easy (which is good) but, necessarily, it also makes maintaining changes inexpensive. And the biggest impetus right now for ‘large-scale’ forkers to merge is the cost of branch maintenance. If branch maintenance is suddenly easy and cheap, then the biggest single argument for actively and regularly merging goes away. The kernel community likes good distributed VCS exactly because they have this insane maze of branches, all not quite alike, and they (crazily) think that is a feature, not a bug.
Now, this is basically a social problem that was just hidden by our crappy tools- if the cost of branch maintenance is the biggest reason for regularly getting patches upstream, then we willl need to find or create other reasons, instead of doing what we currently do, which is rely on on our crappy tools to create reasons :) And like you say, the cost will be lowered, so we might not need as many reasons as we did in the past. But just because we’re uncovering a problem doesn’t mean it isn’t a problem. So, again, I’m irritated when distributed VCS proponents (who generally are pimping a great product) respond to a legitimate concern by saying ‘oh, it isn’t really a concern.’ It is a concern. It takes away a stick we had (for better or for worse) to get people to merge, so we’re going to have to find some carrots. Don’t get me wrong- it is a problem I want to have to solve, because I’d rather have to convince distros to hurdle a low barrier than a high barrier. But still… saying it doesn’t exist doesn’t make it go away.
Besides rant about VCS, not much to today- slept in a bit, watched some baseball, made some delicious eggplant onion pasta salad (also from Foster’s cookbook), cut down a bit on my email backlog (though still bad :/, encouraged the xchat-gnome guys to track down a very troubling xchat-gnome bug that is probably really some deep gtk 2.8/X interaction bug (it doesn’t happen with gtk 2.6 :/), and worked a little on my trademark essay/rant.
Also spent a bit of time wondering how hard it would be to ‘scrape’ the interesting flickr photos for CC-SA interesting photos. Seems like many of them would be great backgrounds :) Sadly, there is no rss feed for the interesting photos (AFAICT) which makes it slightly more scrapy and slightly less web-2.0-y. Oh well…