The third day of OSCON was where all the little talks started happening and was very exciting. More people were here and the conversations I had in the halls were all wonderful and gave me lots of ideas. My underlying theme for the day was "virtual machines".
I started off with "Exploring the Microsoft Shared Source CLI (aka "Rotor")", where Peter explained the differences between Rotor and the commercial CLR. He's a good speaker and communicated well and he covered all aspects of Rotor. It seems strange that a Linux port hasn't been blessed by Microsoft but he mostly covered things I was aware of. More on the CLR later on.
Then Michael talked about "One Year of PHP at Yahoo!", which was in a very packed room. It was interesting in how he used PHP in an enterprise environment, but there were some little things which really threw me off where he looked like he was micro-optimising (stat calls, for example). Also, he complained that in PHP4 the object orientation was slow, so he recommended people to not use OO in their PHP code. This is unthinkable for anything other than really small projects. I so don't understand.
After a free lunch from Microsoft we had the Perl lightning talks, which of course were all good. I still don't understand why everyone seems to want slides for a five-minute talk, though. Autrijus' talk was by far the best. He rocks.
Then Dan gave his state of Parrot. It's all very positive and we basically have a fully functioning engine at the moment. These are exciting times.
It was time to present my "Little Languages in Parrot" talk, which went really well. I think I scared everyone a little with Befunge, but the whole point of my talk was to convince more people to help and I think I may have been successful. ThI finished a little early and gave Arthur some time to introduce Ponie and why it's so cool. That went down well too.
A quick spicy salmon and p5p-stats later and I go to the "Dynamic Languages on the CLR" BOF, hosted by Microsoft people. A lot of good people turned up: the parrot crew (sans Dan), Larry, matz, Activestate peeps. Unfortunately the Microsofties spent too much time trying to lead the conversation (no, a lot of people don't care about managed code) and it looked like they hadn't done their research into dynamic languages at all. Very disappointing. Oh well, I'll grab them later on and explain why we started parrot.
Then I went to a party, drank beer, talked about ponie and CPANTS and jobs and deprecating v-strings and eventually went home to sleep. Mmmmm, sleep...
seriously though, I am very cynical about what rotor can offer to people who aren't activestate - mono provides 95% of rotor but without the cancerous / parasitic licensing.
Shared source is typical microsoft 'heads we win, tails you lose' marketechture and best avoided.
If microsoft really wanted CLI to be an open standard they would have followed suns example, now java is flourishing as java is a proper open standard without restrictive licensing.
Re:rotor and open source contamination
acme on 2003-07-10T21:38:54
I hate licensing. mono and portable.net are reinventing the wheel for no good reason, apart from licensing.Actually, the license for rotor isn't too bad. It even contains the phrase "You may use any information in intangible form that you remember after accessing the Software".
God I hate licenses.
Re:rotor and open source contamination
TeeJay on 2003-07-11T08:56:44
I think there is more to mono than licensing, end even if it were just about licensing it brings some kind of diverity if not healthy competition.I am glad KDE and GNOME both exist because I prefer the approach the gnome people have taken and they have fed back into each other.
At a later point maybe mono and rotor can share ideas that makes the CLI better implemented or more powerful. Of course this is limited by microsoft rather than ximian - again this boils down to licensing - there would not be a need for mono if microsoft had done as Sun have and released a totally public domain reference model.
