RobMensching.com /Blog
when setup isn't just xcopy

Sun buys Open Source, Microsoft and Google hire it.

This is just a flippant thought so don't read to much into it but I found it amusing enough that I thought I'd share. Last week there was much discussion about Sun buying MySQL. While everyone was discussing what this meant for Sun and MySQL and the "Open Source Business Model", I was trying to figure out why this deal felt unusual yet familiar to me.

Then it struck me. What I found interesting was that most of these discussions were about about Sun acquiring the company MySQL. Sun was "buying the M in LAMP". Then I remembered that Sun had done this before. Sun bought StarOffice years ago that ultimately became the foundation for OpenOffice.org.

Microsoft, on the other hand, tends to hire people with Open Source backgrounds rather than buy "OSS" companies.  I was trying to come up with a list of recent "Open Source hires" when this interview with Sam Ramji showed up that listed some of the most recent people to join the "Evil Empire" (*I* can say that, I actually dressed up like a Stormtrooper to represent the company in public <smile/>). I'd also might add Phil Haack, Scott Hanselman and Rob Conery to the list.

Google, like Microsoft, seems to hire Open Source people more than buy entire companies. I'm thinking of people like Chris DiBona, Sean Egan and Guido van Rossum. But it isn't like I really track Google's or Sun's or Microsoft's, for that matter, hiring and acquisition practices closely. So maybe I'm just making this all up.

Note, I'm not suggesting that "buying in bulk" is better or worse than "hand picking individuals" to build up a company's Open Source collection. However, I do wonder how those different approaches might affect a company's culture. I know that inside Microsoft it feels like Open Source acceptance is growing like slow winter thaw.

 

Investigating Blogging Server Software.

A few days ago I tried to post my second blog entry of the month. Windows Live Writer returned a very helpful error message from my server that told me I was out of disk space on the SQL Server. That was troubling because I don't have that many blog entries. It was late though and I didn't feel like debugging the server software so I called it a night.

I finally got around to digging into the SQL database this morning and discovered the problem. Subtext (my blogging server software here) stores referrals in a database table. Well, that table had finally gotten big. Too big. So I jumped on the SQL administration console and truncated the subtext_Log, subtext_Referrals, subtext_EntryViewCount, and subtext_URLs tables. The fact that you're reading this blog entry means that I finally fixed it.

Naturally the next thing I need to do in my "copious free time" is to figure out how to avoid this problem in the future. I know the best answer is to work with the Subtext community and upgrade to the latest build then add a feature to clean up these logging type entries from the database on demand (at least) or "automagically" (if possible). The problem is that I'm not sure I want to stay on Subtext.

The main concern I have with Subtext is that it is just big. There are about 45,000 lines of C# code (20,000 lines of comments). I expect all that code has its uses since Subtext has support for hosting multiple blogs and image galleries and such. However, I'm not terribly interested in all that stuff.

BlogEngine.NET is a new kid on the block and interesting because it's about 16,000 lines of code (5,000 lines of comments). It does seem to be targeted at an individual blogger and is up with the latest blogging technology. However, on my first high level pass through the BlogEngine.NET code I found the formatting of the code style to be pretty inconsistent which is an indication (in my mind) of how much the code is cared for.

Finally, Fredrik Grohn created his own blog engine with something like 300 lines of C# code. However it doesn't do comments, trackbacks or any other blogging functionality above and beyond RSS syndication. That makes it really bare bones but maybe a reasonable place to start from scratch... if I wanted to do that.

Ultimately, I'm probably going to stay with Subtext (and write a little utility to regularly truncate the logs) because I don't really have the cycles to spend getting into a new codebase and community. While I'm sure I could learn some cool new tricks playing with blogging server software I expect it would just be a distraction from the much larger projects that I'm working on in my free time. However, I have to admit that writing unit tests for BlogEngine.NET as a way to learn the codebase and help improve/maintain quality of the codebase has quite a bit of appeal to me.

Anyway, I'm back. More blogging to follow.

 

The much exaggerated death of the WiX toolset.

If you watch the front page of the WiX toolset's web site and get notifications for changes to the Windows Installer XML News on SourceForge and subscribe to this blog then this is the third time you've seen this content. I apologize for the message in triplicate but there seem to be a surprising number of people that seem to believe that the WiX toolset has been abandoned because there has not been an update to the WiX SourceForge page in a while.

To combat that belief I am posting this notice widely:

There has been some suggestion that the WiX toolset is a dead project because we have not released a drop to the SourceForge release point in a long while. Those fears are unfounded for three reasons. First, WiX v2 is stable so we are not expecting another build of that version unless a showstopper bug shows up. Second, weekly drops show up at the Weekly Releases release point regularly. Finally, pushing to the SourceForge release point is a manual process for us and thus a lower priority. In any case, development is still very much active and the wix-users mailing list is still the best place to get your questions answered.

At the same time, I'm going to re-evaluate whether ignoring the content on the (IMHO) "fugly" SourceForge pages is a good idea. The freshness of the data presented there makes a larger first impression than I anticipated.