As I said, I've attended Oleksandr Pavlyk's presentation of Mathematica 6 features, which was held on May 30 at Kiev-Mohyla university.
Lecture went off very well. Also there were many jokes both from the lecturer and from the auditory. There were about 15 people, roughly equally students and adults.
Oleksandr answered many of my questions. Here's what I found out:
- 2/3 of Mathematica kernel is written in C and 1/3 in Mathematica itself; also they use some libraries written in other languages.
- I wrote
about my frustration with the new help system. Oleksandr responded that this
was done to make help an ordinary Mathematica notebook, and the old system
was a "dirty hack". Another reason is huge growth of the documentation
WR is aware of slowness of help system on Linux — they think it's related to Java performance on Linux, and Java is somehow involved in displaying documentation.
- Mathematica uses number of opensource libraries (LAPACK was mentioned). They contribute back in form of bug reports.
- Their researchers publish papers, but they don't maintain anything like papers page by Google, as they don't see need for it. Oleksandr suggested searching the papers through Google scholar.
- There are free on-line seminars on wolfram.com, and they include not only training on some Mathematica topic, but also "talks with devs".
From the presentation I also learned about few new features in version 6, which looks quite useful.
Manipulatemakes experimenting with parameters easy. Although it looks like one of those bells&whistles in Mathematica 6, I find it useful.
RegionPlotis definitely a great thing, which I missed many times.
At the end there was training session, where everybody had an ability to try new features under the supervision of Oleksandr. There I got acquainted with Vladimir Bondarenko from Simferopol, who does very effective testing of computer algebra systems.