October 2010 Archives

Fedora Board seats

| No Comments | No TrackBacks
I just figured that I'd mention that nominations are now open for the 2 seats on the Fedora Board currently occupied by Matt Domsch and Chris Tyler (two folks who I've really enjoyed working with over the past 5 months or so on the Board). Sadly, I've noticed no nominations on the wiki page for doing that (and only two nominations on the page for FESCo). So go forth and nominate yourself or someone else, but just make sure that "someone else" is actually willing to serve first! :)

FAmSCo is looking slightly better, but they could use more candidates as well (they only have 8 for 7 seats).

If you want to become involved in the leadership of Fedora, these are really good ways to do it. Note that for FAmSCo and the Board you don't really need to be a rockstar C coder, and these are great ways for people who might not have that skillset to contribute to the leadership of Fedora. The Board actively eschews technical issues, preferring to leave those to FESCo.

Anyhow, get to nominating!
I generally think that optional insurance that you can get with your phone, etc. is generally a waste of money - you know, the pay $5.99 a month and get coverage in case you lose your device, etc?

Well, I get it on all of my devices, even though I think it's a waste since I rarely use it. Well, this weekend, I seem to have lost my MiFi - the personal wireless 3G hotspot thingy - the most convenient thing in the world, and how I participate in Fedora Board IRC meetings, etc. I was using it with my shiny new iPad (another wonderful thing - turns your wifi only iPad into a 3G iPad using something that I'd be paying for anyway, and on a respectable network!), and must have forgotten it somewhere or something.

The deductible was a very reasonable $39, and the online claim process (I searched Google for Assurion, expecting to find a phone number where I could talk to someone - instead, it's a "file your claim online here!" thing) couldn't have been easier. What's more, overnight shipping was free! I'll have my replacement on Tuesday, but unfortunately this means no Gobby for the Monday board meeting...oh well, I'm not assigned to take notes anyhow! :)

Now for the bad part. In doing my online research, I found that Verizon is going to start selling the iPad next week - no biggie, I won mine for free, right? The problem is that they are selling the wifi model only, with a MiFi (the exact same setup I have), AND GIVING A $10 DISCOUNT ON MONTHLY SERVICE (and no contract, presumably to compete directly with AT&T's similar offering). Oh, and if you already have either an iPad or a MiFi (or both), you need not apply for the discount - they won't give it to you. Boo. (though in looking through my account, I found that they're not applying my corporate discount to the MiFi, dunno if they don't offer it on data plans or if it was an oversight - either way, time to get on the phone with Verizon and find out!)
From Mike McGrath's excellent post the other day about the suckage of yahoo mail with Fedora Infrastructure.  This has come to my attention off and on ever since we migrated our mailing lists that Yahoo will randomly block us, I still see blockage in our mailserver logs as recently as today.

This in and of itself wouldn't be so bad, if Yahoo had a clueful mail admin team that could respond to our complaints in a sane manner, but alas, they're nowhere to be found. They keep sending us repetitive questionnaires that don't make any sense.

In short, if you care about your mail from Fedora (or likely a large number of other legitimate sources), don't use Yahoo Mail.

This message brought to you by Gmail(TM). :D
Enhanced by Zemanta

Etherpad FAD - Day 2

| No Comments | No TrackBacks
I'm now on the Acela back form the Etherpad FAD at Olin College, writing this on the wonderful Amtrak wifi service. This was an excellent and very productive weekend, even though we fell short of th emilestone of having an acceptable Etherpad package for Fedora. This is because of a few thingss:

There were more bundled libraries (some of which required packaging for Fedora, some which were already in Fedora) than you could shake a stick at, and extracting them isn't always as simple as we'd like - sometimes we'd have to BuildRequire the library, and copy the system version into the proper location in the source in order to have the build of the things that aren't bundled not puke.

Another issue that we found with one of the dependencies took a few hours last night, and is still not resolved.  A library called c3p0 (not the Droid, I'm not really sure what in the world it does, but maybe some Wookie cuddling would help) won't build in Fedora because it can't find ANOTHER library that it's bundled with, even though the system version of the library is installed, and the build.properties properly points to the system location of the library, but the build still fails saying that it can't find classes that are in the library. We've verified that the proper classpath is being passed to the Java compiler, and that's about the extent of my knowledge in troubleshooting Java builds.

This minor setback was, in my opinion, offset by the tremendous amount of success that we *did* have. A lot of good work was accomplished, though we're holding off submiting pakcages for review because we want to submit the entire stack at once, for fear of having "abandonware" in Fedora.

Speaking of abandonware, Etherpad itself could be classified as sort of that. The lineage is that a company called Appjet produced Etherpad, and they were acquired by Google, mainly in order to acquire some of their developers for Google Wave. They didn't really want the Etherpad product itse;f, so they open sourced it, and failed to cultivate a community around it. The current community around Etherpad is entirely non-Google, and there appears to be no canonical upstream - there's a version on Google Code, there's a version (or two or three) in Github, and I'm sure there are others hanging around.

This brings me to Jeff Mitchell's (of KDE, one of the lead developers of Amarok, member of the KDE Community Working Group, etc) talk around how to build (and destroy) community, and reinforced some things that I already knew, but hearing them over and over again, and ingraining them in everything that you do in the FOSS world, The part that really resonated with me was that "are we here for ourselves, developing stuff that we think is cool that scratches a personal itch, or are we writing software for users?"

You can't drift too far to one extreme or the other and still be successful. Too much of developing for yourself, and you wind up with no users, and no new developers. With no developers, you have no community. Too towards the user side, and you get developer burnout, and more importantly, users will slander your good name if they are displeased.

Another important concept, which should be incredibly obvious, but again, we all need reminders of it from time to time, is that there are three types of communities, each with dfferent needs and motivators. There's the user community, the developer community, and the commercial community

There are basically two types of people in free software communities - those whose primary motivation is loyalty to the concept and ideal of free software, and those whose primary motivation is towards monetary gain. Note that the two aren't mutually exclusive of each other, it matters what the PRIMARY motivation is, not that there's some of the other in there. Note that for people whose primary motivation is loyalty to free software, money can and does help - for instance, sponsoring attendance at FUDCon or other conferences, etc.

Typically, the user community will be motivated by monetary gain (which seems counter-intuitive, but they do indirectly gain financially, since they don't have to buy propietary software).  This is why, as mentioned earlier, they will have no hesitation to go somewhere else and slander your good name.

The developer community is typically, but not always, motivated by the loyalty to free software, which should be also failry obvious, while the commercial community is motivated by financial gain (though the people WITHIN that commercial community may be motivated by other things).

I've rambled on enough for now :).

Etherpad FAD - Day 1

| No Comments | No TrackBacks
Today marks the beginning of the Etherpad FAD. Andy came and picked me up at the Route 128 train station, but he got stuck in traffic on the way (getting here at rush hour on a holiday weekend was probably not the best idea). We then went to Olin, and joined the rest of the crew in the Campus Center.

Today was really productive, and I got alot done personally.  I proved that Etherpad can run against OpenJDK on Fedora, and helped to explain some of the infrastrucutre requirements for getting Etherpad running in Fedora Infrastructure.

As far as the packaging front, Etherpad, as with most Java applications, requires quite a number of external dependencies, most of which are included with Etherpad itself (unfortunately).  Some of them were already available in Fedora, and worked out of the box. Others weren't available in Fedora, and couldn't be - there was one with the stupid "do no evil" gimmick license that can't be included in Fedora, and others that could be, just need packaging.

There were other things that are available in Fedora, but the versions that were in Fedora weren't functional for Etherpad (again, not an uncommon situation for Java applications). So we have to make compat packages for a good number of things. Java libraries are generally insane, and don't maintain API from one version to another (even minor versions!)

One major problem that we've come across for Fedora Infrastructure, there is no version of Scala packaged for RHEL5, so we'll have to package the correct version of Scala for EPEL, and a compat Scala for Fedora, since Etherpad unfortunately doesn't work with Scala 2.8, the current version in Fedora.

Some of these are on tap for tomorrow, which should be another very productive day!

There were some
Enhanced by Zemanta