So, someone pointed out in the comments that the Elisa packages for F10 are very old. I got in touch with the maintainer to see what's up. It turns out he would dearly love to be updating them every week and have them bang up to date, but he's been waiting to have the new sub-packages that came in with 0.5 reviewed for nearly a year: see this bug report. This is a shame. :( If anyone reading this has package review privileges, please review the packages - the Elisa ones, obviously, and also python-twill. For now, anyone wanting up to date Elisa packages for Fedora 9 and 10 can grab them here. They're updated every time a new release comes out.

Now THAT was a hell of a game. I'm gonna stop screaming at the TV now.

There will be a Test Day focusing on the 20 Second Startup feature for Fedora 11 on Thursday, 19th February, in #fedora-qa on Freenode. Test Days are a great way to get involved with QA as you know there'll be lots of people around who can help you out, and this topic is one everyone can help out with and most people have an interest in. The page for this test day is available already, and if you like, you can get a head start by adding your data to it now (as I have done). There's full instructions on how to generate a set of boot time results for your system and upload them for the developers to analyze. It should only take about an hour to run a full set of tests. So please provide your input and come by for the test day on Thursday!

Elisa just keeps on rocking more each week. They release an update to the 0.5 series every week with useful bugfixes and small new features, and rarely break anything. This week's update - 0.5.28 - is particularly dear to my heart, as they implemented a feature I've been bugging the poor Fluendo guys about for months now - in-container soft subtitle support. This is a feature widely used by anime, which is mostly what I use my HTPC for. Rather than being simply burned in as part of the video stream itself, or supplied as an entirely separate file, subtitles are provided in text-with-timing-information format, to be rendered by the player, within the video container (MKV is the container format usually used to do this). Gstreamer has basic support for rendering these subtitles, but it's utterly non-configurable, and shows up in rather ugly - and small - yellow text, which was pretty unwatchable. Finally, with this update, they get rendered in nice, sensibly sized, white letters by default in Elisa, and you can configure how you want them to look as well. Awesome stuff, guys. With all these quiet weekly updates, Elisa is turning into a real contender. With this fix I can pretty much use it full time for my HTPC. This is nice because it's a very pretty and clean interface, and also because Gstreamer actually seems to render video more smoothly than xine or mplayer - I get quite a bit of tearing, especially on HD videos, with xine and mplayer, but not so much on Gstreamer. Not sure why this is, it's a bit of an opaque area for non-experts to investigate. It's not maxing out the CPU, so that's not it, and I did test briefly with VDPAU, which didn't seem to make any noticeable difference. I'm still packaging Elisa for Mandriva, because my HTPC is still running Mandriva (don't really see any need to go through unnecessary work to convert my servers and HTPC to Fedora, Linux is Linux...). Fedora has good Elisa packages already, maintained expertly by Matthias Saou, so my services are not required there. He doesn't update quite as fast as me, though. ;)

I'm happy to announce that phase 1 of the previously discussed QA group Wiki space revamp project is complete. You can now see the new front page and Joining in page in their final locations. I hope everyone agrees that they're an improvement on the old model. As well as those two pages, we made significant changes to a few others, breaking out a big chunk from the messy Testing page (which will probably go away soon) into a new page, revising the Rawhide page, and a few more. A huge thanks to all the guys from the QA group who helped out with this project, including Leam Hall, Jóhann Guðmundsson, Christopher Beland and others.

Well, time to start posting here about what I'm actually doing for Fedora! There's a ton of stuff on my to-do list and lots of interesting discussions and ideas I'd like to write down, but I'll try and stick to one thing at a time. One project we're doing quite well with is revising the Fedora QA Wiki space. That link points to the current incarnation of the front page, which is...well, not as great as we'd like it to be. :) So I've started a project on the fedora-test-list mailing list to revise it. With the help of some great people from the existing QA community, we now have some good drafts for the front page and a new 'how to join in' page: my draft, Jóhann's draft. For me, a Wiki area like this is of huge importance for driving increased involvement in a project: it's the public face of the project, the place people go to find out what it actually does for them and whether they might want to help out. So it needs to be simple, readable and attractive.

Whew. So I'm back home after a very busy week in Raleigh. I had a fantastic time, actually. I met a bunch of great people there. We - that's both 'we' as in the internal RH 'Fedora QA' people and the wider Fedora QA community - have come up with some plans that we should be able to really get to work on in a short timeframe and that should have concrete positive results very soon, which is great. I'll write in more detail about some of the things we're planning soon, but for now I'm just incredibly tired. I didn't sleep much the whole week in Raleigh (can't sleep well when I'm on the road) and was working out every night and waking up at 7 am, so I'm pretty frazzled. :) I'm going to finish setting up my mail server to handle my Red Hat mail and then I'm going to go to sleep, for a long time, and spend the weekend definitely not working. So I'll be back at the coal face on Monday. Thanks a bunch to everyone in Raleigh who was kind enough to answer my dumb questions, and to everyone both within and outside RH who's welcomed me to my new role, you're all great. I'm really looking forward to getting started with some projects.

Forgot to mention in the previous test - I am proud to present the first of what will no doubt be many exciting and highly useful concepts I have been lucky enough to participate in developing! Will, James, myself and Vincent (Danen) were all chatting and got on to the topic of the discussion about what compiler optimization flags to use, that's come up recently on the Fedora development mailing list. We decided the basic problem was setting the bar in terms of just *how* old hardware you bother supporting. Someone pointed out that they've found a working i686 system in a dumpster, but it was in turn noted that it was a dumpster *in Raleigh*, which may not have been a fair test. So we perfected... The Kentucky Dumpster Test which is a simple method for determining minimum acceptable hardware. To whit, the lowest spec machine you should support is one for which you can find a working example in a dumpster in Kentucky. Anything older than that gets no love. A fine and practical concept, I'm sure everyone will agree! I'm off to file an expenses claim for a fact-finding trip to Kentucky...

So, I am here in Raleigh sitting in James Laska's cubicle, writing this on one of the four systems lined up on the desk next to me. :) I've been here a couple of days now for orientation, and meeting up with some of the rest of the Fedora QA team I'll be working with here. It's been a lot of fun so far, and from this little sample, RH definitely is living up to the expectations I had. This afternoon I sat in on the desktop training session for the non-Linux users in the orientation group - it was run by a nice guy from the helpdesk team, and it was really interesting to see how he ran that, and how the people in the class took it. It went more smoothly than I honestly expected, which really backs up the idea that if used in a properly controlled environment, Linux isn't any harder (or particularly easier) to use than anything else. As the guy who I was shoulder-surfing said, "in the end it's basically just 'point here and click this' no matter what system it's running". It kind of crystallized for me the idea that in a corporate environment the operating system is a very minor issue to the end user in itself - what's far more important is how the whole infrastructure works together and how it's presented to the user. Quick update: the day I got here my mail server VM fell over (it seems to do that, for no reason at all, on occasion, while the web server VM stays up unless I actually do something bad to it - odd), and the stuff I set up to leave a hole through my firewall to let me log in to my desktop via ssh just exactly so I could fix things in this eventuality seems not to be working. So I'm stuck without access to email on any of my existing addresses. I'll pick it all up when I get home, nothing's lost, but if it's something really important - you can mail me at my spanking new Red Hat address, which is awilliam@ . As I can't contact my mail server I'm just accessing that direct on my laptop at the moment, so it works. I'll get a different alias later - RH default alias has an eight character limit, which is why it's a bit odd. :) If there's any RH people in Raleigh reading this, do feel free to seek me out and say hi, I'll be here till Friday, probably somewhere in the Fedora QA group. It's nice to meet as many people as I can before I head back to the moosey wilds!

So I'm sitting in a cafe, typing this on my 1.5lb, 8" laptop with the aid of a Bluetooth mouse, via the 3Mb internet connection provided - again through Bluetooth - by the phone in my pocket. I remember downloading the shareware release of Quake to my spanking new desktop's 210MB hard disk over a 33.6k modem. It took three hours. The future seems to have happened while I wasn't quite paying attention...