Will’s Miro dev blog

Short status this week….

I took Wednesday through Saturday off.

On Monday, I fiddled with my Windows build environment and finally got it working (again) on Tuesday. I was having problems getting fasttypes to compile. I was getting all kinds of errors when going through the boost stuff. After some skulking through Google results regarding boost compilation problems, I decided to try installing Visual Studio 7.1 without installing the service pack. That worked super–though I’m not entirely sure why. I updated the WindowsBuildDocs page with new urls and tried to break up the instructions into something that’s more digestible.

I also continued to work on the Firefox patch.

I spent Wednesday, Thursday, Friday and part of Saturday watching new bugs, helping users out with 1.0 issues, continuing to build a Windows VM (I’ve almost got it working again) and working on Mediabar.

I checked in a minor overhaul of Mediabar. There are two big issues with Mediabar that need to be fixed that involve architecture changes (I’ve been talking about this for a month now). I figured since I’m overhauling the code for that, I might as well overhaul the code and fix namespace issues and tighten up the existing architecture to make it easier to fix the big issues. In the process of making the changes, I noticed the flv extraction code doesn’t work. I’m not sure how it’s supposed to do what it does, though, so I’m not sure if it’s something I did or something that was pre-existing or something I’m misunderstanding. When I work on Mediabar again, I’ll talk to NPR and Dean about how it should behave and what kinds of things it should be picking up and write it down into an ad hoc specification. On a side note, anyone have any idea how to do agile-like development with Firefox extensions? Where does the testing code go and how do you kick it off?

On Sunday, I got worried that I’m going to miss the deadline for the Firefox patch I’m working on. The work is under bug #400059 in the Mozilla Bugzilla db. I spent Monday and Tuesday working on adding enclosure detection to the FeedProcessor and then adding enclosure support to FeedWriter so that you can see enclosure links on the feed subscribe preview page. When I get this working, I’ll submit it as a patch against bug #303645. Making those changes paves the way towards adding support for distinguishing between video, audio and text feeds and supporting applications for handling those different feed types.

I will be off of email and IRC for the rest of the week but I’ll be studying.

I hope you all have pleasant holidays or work days (depending on where you live)!

One of the things I keep reading in comments of various Miro-related reviews is something along the lines of “there’s no good content”. I think that’s utter bunk. There’s a lot of good content listed in the Miro Guide. The channel starter packs that we added to the first page when you start up Miro 1.0 make this painfully clear. This doesn’t even include all the content that’s not even listed in the Miro Guide.

I don’t have cable tv anymore because it doesn’t make sense to waste my money on it.

I also don’t watch a ton of shows with Miro. However, here’s the list of shows I do watch (some of them while testing):

• Ask a Ninja (Add to Miro) - Occasionally there’s an annoying show, but mostly I think it’s pretty funny. I saw the Recipe For Disaster episode just before a family reunion–couldn’t have had better timing. I first discovered Ask a Ninja when one of the NPR programs I listen to periodically (I forget which one it was) had their movie critic off for the week and they played the audio from the Ask a Ninja episode reviewing Pirates of the Caribbean: Curse of the Black Pearl. (Disclaimer: I have ninjas on my business card.)
• Make Zing :: Blog MAKE Podcast (Add to Miro) - I love this channel. The projects are really interesting and it covers a very wide variety of topics. Dean, my brother and I met Bre at PodCamp Boston 2 and my brother secretly thinks that the t-shirt cannon project was influenced by his story of kids launching potatoes at a river from a cannon in their chimney (long story–very odd).
• Galacticast (Add to Miro) - I met Casey at PodCamp Boston 2 and after hearing about Galacticast decided to look it up. It’s a great show!
• What you ought to know (Add to Miro) - The shows are almost all under 3 minutes long and they cover a variety of topics. It’s thoroughly educational in tiny bite-sized chunks.
• Wired Science | PBS (Add to Miro) - This is a great general science channel.
• WebbAlert (Add to Miro) - Morgan Webb and her crew do a really good job of distilling “tech news” down to a 5 to 6 minute program Monday through Thursday. I find watching this saves me the trouble of flipping through the series of blogs I used to flip through.
• Google Tech Talks (Add to Miro) - Some of the Google Tech Talks aren’t wildly interesting to me, but I’ve learned a lot from the ones I have watched. This channel is based on a Google Video search and so you’re going to want to set the Auto-download to OFF. Otherwise it’s likely you’ll pick up old videos you’ve already seen.
• Onion News Network (Add to Miro) - I usually test with ONN because it’s fantastic. The “Ninja parade slips through town unnoticed again” episode got a lot of play time on my systems–it makes me smile every time. (Disclaimer: I have ninjas on my business card.)

I think there’s a lot of other great content out there both to watch and to participate in.

So to people who shrug Miro and Internet video off because “there’s no good content” I say, “Buddy–this is 2007 and you’re missing the boat”.

What shows do you like and why? Toss your thoughts in the comments.

Update 11/16/2007: Fixed a grammar issue and somehow I managed to misspell Galacticast.

I did a bunch of release-management stuff, some minor bug triage work and some minor Gutsy work.

I passed a bunch of email back and forth with BDK and James regarding problems between Miro and the sun-java*-plugin packages on Gutsy and Feisty (bug 8444 and now bug 9064). BDK looked into it further but in the end either the test we’re using to determine whether the problem exists or not is bogus or we didn’t fix the issue. Regardless, after much discussion it was decided that the package conflicts were worse than the problem so we removed the conflicts for 1.0 final.

I got in touch with Dean’s friend Ben (not to be confused with BDK or my brother Ben), and he and I are going to go through our Gutsy and Feisty packaging and fix any outstanding issues (like bug 8716). I think this is pretty cool and hope that this is the first of many Boston-area Miro hack-fests.

I also worked on Mediabar. I’ve been doing a pass at cleaning up namespace issues and code cleanup. After I’m done with that, I’ll work on the tab rearchitecture and the rss discovery problems. Neil and I traded some email and he’s eager to work on things again. I’m currently the bottleneck on further Mediabar progress–I’ll be spending the rest of the week fixing that. I want to get back to working on the Firefox patch, too and get that done ASAP.

On a side note, I was selected for the Nokia n810 device program. I want to look into porting Miro over to the device and do some other development, too. It’ll be a good system for figuring out how Miro could work on “smaller devices” and what a slimmed down version of Miro can do. I also want to look into what it would take to get Miro working with Conduit so that Linux users can move video content to their n810 and other portable video playing devices.

As a side note, I live in Somerville, MA. If anyone (users, testers, developers, …) is interested in getting together to triage bugs, working out issues, fix problems, add features, … let me know. I’m totally game for hack-fests and getting together.

Miro 1.0 has been released! Yay!

I’ve only been with PCF since July (or maybe it was June–I forget), but since I came on board we’ve been working hard on stability and honing the feature set. Working on stability is hard because there are a near infinite number of combinations of library versions, video card drivers, operating systems, … out there and all of them are slightly different. Writing software that works on multiple platforms is non-trivial. It’s a huge testament to the community of users and testers and developers that Miro is at the point it’s at now.

One thing about 1.0 that I want to mention is that this is a snapshot in time of a continually evolving piece of software. If you look at Bugzilla, there are dozens of interesting features that we’re all interested in that range from starting Miro as a daemon process to viewing video as it’s downloading.

Chris, Nick and Ben are working on post-1.0 development already. There’s been discussions on the develop mailing list regarding reworking the user interface to use native widgets and make it much faster and more responsive. Paul is continuing work on the Miro Guide. Janet is working on making community testing easier for everyone involved and produce better testing data. I’m switching off to work on Mediabar. Dean and the Team Miro folks are working on honing the documentation and they’re doing a fantastic job of testing and identifying issues for release candidates and versions.

Miro development is moving along and its momentum is a direct result of us all working towards a common goal: building an Internet video player using Open Source and open standards that will enable the current generation of media content to flourish.

One other thing I want to mention is that we ditched the conflicts between the miro package and the sun-java*-plugin packages for Gutsy and Feisty. The problem between the packages still exists and it’s intermittent, but several conversations with people caused me to rethink adding the conflicts. So this doesn’t fix anything–it’s just trading one set of problems for another, however I’ve come around to agree that the conflict is more of a pain in the ass than occasional intermittent crashes.

I mothered the 0.9.9.9 release on Wednesday, which went really smoothly, and then the 0.9.9.9.1 release on Thursday, which sucked and took two days to sort out. I had problems doing svn checkouts on the Windows build box (took in excess of 1.5 hours for a single checkout) and then problems with the windows-xul setup.py script which had bugs when the version number contains more than four pieces. I had some help from Nassar working that out, but it sucked up most of my time and I didn’t really get much else accomplished.

Over the weekend, I helped out a few users who were having problems with 0.9.9.9 on Gutsy.

On Monday, I worked on Mediabar and continued building a new Windows XP vm.

Today I worked on legal stuff, bug triage (making sure I’ve got all the things I need to get done for 1.0 done) and Windows XP vm stuff.

Overall it was a week short on getting things accomplished and long on random problems and wasted time. Frustrating.

I plan on working on Mediabar and Firefox 3.0 patch work in the foreseeable future. I’ll probably be avoiding IRC for a while.

We released 0.9.9.9 yesterday and it took us 5 or 6 hours from tagging the branch to releasing builds. That’s pretty cool and makes for a smooth release. Also, I didn’t screw anything up this time.

I’m an employee of PCF so it’s my job to work on Miro (and the other things I work on). However, I’m just one guy and time is a limited resource. There’s no way I can triage, identify and fix all bugs. I appreciate all the help that I can get. I had a bunch of help this release.

I wanted to thank the following people for the time they spent submitting quality bugs, sending in patches, and helping me fix issues:

• Simon from dbus-python who walked me through fixing the dbus-python deprecation issue and reviewed the code changes I made.
• Markus who helped me with a variety of Gutsy and Gutsy packaging issues.
• Paul who noticed our README was out of date and suggested changes.
• Matthias who helped with Gutsy and Feisty packaging issues.
• Gotz who helped with packaging issues.
• Stéphane who helped with packaging issues.
• Marco who helped with packaging issues.

If there’s anyone else that helped out that I’ve forgotten, I apologize. Please kick me and/or comment below.

We’re well on the path to a solid 1.0. When 1.0 is finally out, we start working through the architecture changes required for many of the features that have been suggested.