So my C plus plus implementation of my mathematical model has thrown up a bug in the model. Not a good thing ™. I’m spending today fixing it and thus won’t be releasing the 0.1.0 version of my code. In fact the bug in the mathematical model may be another reason I’ve been having problems with the implementation.
Archive for June, 2003
Bug in my mathematical model
Thursday, June 26th, 2003Templated classes and bison parsers
Wednesday, June 25th, 2003I rolled back the design decision I made regarding the heavy use of the STL and wrote my own template classes for a 3tuple and 4tuple of objects. So in a night I got more work done than in the previous two weeks and I shrunk the codebase by about 1000 lines. I also think the new API’s are better and the core functionality is much more encapsulated.
I still have some small bugs in my Bison input parser, which _will_ get fixed today. Expect the 0.1.0 release sometime tonight or early tomorrow. I’m calling it the “Thus came a mountain from a molehill” release.
I saw the picoGUI guys using sentances to reference releases, much more fun than the 0.1.0 release or the insert_codename_here release. At least the sentances can describe somthing about the development process. I’ve not seen anyone other than the picoGUI guys do this, but I’d love to hear of others.
C plus plus, SCO, patents and Free Software
Tuesday, June 24th, 2003I decided last night to rewrite the section of code that has been offending me for the past 3 weeks. Originally the design decision was a trade-off between making a _lot_ of use of the STL or writing my own, more straightforward, templated classes. I opted for the former, the latter seems to have actually been the correct decision. I’ll be rewriting the code today.
The media flurry about SCO is really begining to piss me off. It’s all FUD people, however I wonder if all this high profile mention of GNU/Linux will do the marketing campaign favours.
I’ll be writing to all my MEP’s regarding the patent debate in the EU parlament next Monday. The only people that will gain from software patents will be large corporations like Microsoft and IBM (who seem to be doing fine without patents currently) and the big losers will be SME’s and Free Software. Alan Cox recently stated that if the EU software patents directive passes, software will be the only area covered by both patents and copyright. A bit foobar’ed. I view software as an alternative denotation mechanism for partially recursive languages, thus software is mathamatics and not patentable.
Theres a first meeting of FSF-ie in Messers Maguire (Dublin, Ireland) next Wednesday at 7:30. I’ll be there, so will others. Turn up if you want.
CppUnit strangeness
Monday, June 23rd, 2003For the past few weeks (too long) I’ve been implementing my mathamatical model (http://minds.may.ie/~balor) in C plus plus. It’s one of those things that needs a lot of testing as if the implementation does not correspond to the theory it may screw up some of my future work. Well basically I’ve found some very strange things going on in CppUnit, which are correct but not fully documented.
Because my mental model of how CppUnit works is different to how it actually works I’ve lost two weeks of code. It’s hard to maintain concentration on a single part of a pretty simple project for this long. I’m pretty jaded by it. Well we’ll see how it goes this week. Thankfully I’ve got Lori to keep me going.