John list of reasons to vote for Kerry was zero, so I figured I could cook up a few interesting links:
I guess am biased being the son of a scientist and all that.
A new piece from Chomsky.
I also found his blog. Wonder if Chomsky uses Emacs being from MIT. I have always wondered how approachable Chomsky would be to have dinner with my friends. Anyone have his email address?
A new publication: Gnome Journal to keep up-to-date on interesting bits from the Gnome world.
Gnome 2.8 was released: fancier, simpler, cleaner and more hip than ever.
I had dinner with the Fluendo guys in Barcelona last week at Thomas' place days after they had released their Theora Java applet. The quality is very good, and the best thing is: you do not need to install any software on any machine to watch live video streams from Theora servers. They will be happy to sell you a streaming server.
The Gnome Summit announced its dates; this year in Boston will be held in the Stata Center in MIT, the new gorgeous buliding across the street from the Cambridge Novell offices where we work from: Mark your claendars.
Jeff's company launched their new distribution: Ubuntu. Not only they picked a great name for their company, but they are adding timed releases to Debian, I would say the major problem for Debian's adoption (please do not send me email saying `just use sid/whatever' or `Debian testing is great for me').
When I started using C# on the early preview days of .NET, one of the first things I noticed is that the compiler did quite a lot of work to produce helpful error messages.
While implementing MCS, we noticed that it went beyond the obvious internal checks of the compiler: the MS C# team actively had to add extra work to produce a more meaningful error. This is in addition to C#'s definite assignment requirements which by itself catch plenty of errors that are commonly introduced in C.
Recently Marek Safar has been on a quest to refine our error and warnings to produce as many as the MS compiler does and to fine tune our reports. We are not there yet: today we catch 617 errors/warnings, we fail to report 18, and we report differently in 136 cases.
The parsing errors are the most difficult to cope with our compiler as we used Yacc to build the parser and the parser recovery is a lot harder than if you do your own top-down parser (something that we might do in the future).
I did a presentation on Mono last week at the Universidad Oberta de Catalunya, here is the: video.
I was exhausted when I did the presentation, having slept only four hours for the past fourty-eight due to the jetlag, the dinner following up was excellent, I had a chance to talk to a few folks afterwards.
Posted on 17 Sep 2004