January 18th, 2013
Here we are again – it’s the break between barcamp and dinner on FUDCon day 1, and I’m blogging because I know I won’t get another chance for a while. Great barcamp sessions today as always.
I saw Aurelien’s talk on Hyperkitty, to which my response is simply “give me that yesterday” – I’ve wanted it since I saw Mo’s original mockup, and using mailman HTML archives for another two years since has only increased my desire. I have a feeling it’s one of those sleeper projects that will blow up huge once it arrives.
Peter’s talk on Secure Boot went off surprisingly harmoniously – probably because there were no journalists in sight and it was a pretty in-the-know crowd asking deep technical questions, rather than a general-interest crowd asking if Microsoft was trying to STEAL ALL THE INTERWEBS. Most of it went about 10,000km over my head (I have a reasonable working knowledge of SB at the practical, this-is-the-effects-of-what-it-does level, but the actual implementation of SB is super super crypto nerd stuff which I am way too dumb to follow), but it was nice to listen to the smart guys refine their wisdom.
The lightning talks were really good this year and highlighted at least three more YES I WANT THAT NOW things for me – pkgwat (yet another command line package query tool! but it actually does different stuff from yum and repoquery, and is totally worth having), fedocal (the new calendaring webapp, which looks like it’s coming out pretty well) and most awesomely fedmsg-notify, a desktop frontend for fedmsg, which lets you get desktop notifications for all kinds of events reported by fedmsg. Fedmsg is one of those projects which seems like it’s been going forever without anything visible happening, then suddenly mushrooms out into a load of great results – all the heavy backend lifting is mostly done now and the cool visible stuff like this is emerging. This is going to be just super-handy for keeping up to date on things without visiting fifteen different web frontends.
Spot presented a pretty radical plan for the release cycle which generated a lot of comment and discussion – I’m going to leave it to him to blog about it before going into any detail (he may have done so already, I’m not up to date on blogs yet), but despite raising a few concerns I actually quite liked it as a plan and it had the definite advantage of being a concrete, complete and achievable plan with easily enumerable and perceivable benefits. So I’ll be interested to see where we go. I’d hate it to get blocked up by the ‘but it changes this particular feature of the current process!’ crowd – someone raised the point that maybe it’s trading one set of problems for another set of problems, but that is going to be the case with any release process, so we can’t really let that objection alone derail any specific proposal for a revised release process, because it effectively precludes the possibility of any change at all. Anyhow, that’s all I have to say on the topic for now.
Tim did a great talk on the blocker/NTH process, which generated a whole lot of useful feedback and discussion – we had QA people and development people and ‘regular users’ all present and there was pretty unified agreement on which parts of the current process badly need improvement, so we should just go and improve them already. Tim and I have discussed using some of the hackfest time just to go ahead and implement some of the easiest and least controversial possible improvements in time for the F19 cycle – stuff that’s already been raised and discussed several times but which we have not had the time to do until now. It was great to see Dan (Mashal) and John (Dulaney) at all the QA-relevant talks providing input.
I gave a pretty ad hoc talk on the new installer in the final timeslot of the day, of which there should be video here at some point, I think (that should be a video of the whole day, and mine will be in the last hour or so – I’ll update the link if it’s broken or if a better one becomes available). It was really just a straightforward walkthrough of the new installer UI, explaining the design of the installer as a whole and of the storage process, and explaining specifically how each spoke works and some of the more commonly raised temporary issues in F18. I think I managed to cover most things, so I hope it went over okay and didn’t do anything to diminish the awesome work Mo and the anaconda team put in to the release.
Tonight brings the real highlight of any FUDCon: FUDPoker! So I’ll finish here. With any luck we’ll have some concrete results to report from hackfests tomorrow. It would help, of course, if FESCo could kindly nail down a concrete schedule for F19, so we know how much time we have to work with before the Alpha cycle starts up.