Being CS

Meeting Miguel

Last week I met Miguel de Icaza at VM'04. Miguel gave Friday's keynote, about how the idea for Mono was conceived to make Gnome programs language and library independent. One of his best examples for this was Tcl/Tk, a scripting language that was extensively used for GUI design, only because Tk, its GUI toolkit was so easy to use. Prior solutions like Bonobo (also started by Miguel) failed (that one due to the cruft of Corba bindings, though I know little about Corba), and since the .Net Framework was relatively language independent, Miguel saw this as the proper solution.
Since then, Mono has expanded in scope, with Ximian handling many of the details of the actual compiler, run time and JIT, while relying exclusively on others to implement class libraries. He talked about the organizational structure of the project, and some of the key foundations (people were especially interested in the Boehm GC which Mono uses, and some of the limitations of their GC scheme).
Miguel also mentioned that Mono now works well on several OSes, including Mac OS X (with some tiny issues, in this case). After the talk I asked him whether it would be possible to theoretically run Dashboard on OS X, and we discussed some of the problems with this. I also asked him what Dietmar Maurer (an Austrian, who used to be in the Mono core team), is working on, and he told me that Dietmar had left Ximian, to found his own IT consulting company back in Vienna.
At lunch I also met Kirk McKusick, head of USENIX, and author of the famous Design and Implementation of the FreeBSD Operating System book, and it turns out he is currently working on a new edition. The second to last paper of the day was fascinating, though a lot of the papers that day were not that interesting to me. At around five I said goodbye to people, including Miguel, and started heading back home. Instead of going to the Mountainview CalTrain stration, I took the free bus to the Santa Clara station itself. After waiting there for 10 minutes, Miguel suddenly showed up.
We started talking about some of the papers at the conference, and how it seemed that some people were only really interested in publishing their own papers, rather than really hearing what else was going on in the field. Miguel is of course more interested in the practical applications of OS/VM technology, and few of the papers catered towards this. I was interested on how dangerous patents might be to the development of Mono, and Miguel gave a very convincing argument, of why Microsoft would/could not really use patents against Mono, and why it wouldn't be much of a problem if it happened. We also talked about other work that is being done at Ximian, and about how the company has changed since the early days, and what he will be working on after Mono (not sure yet). After that we talked about blogging, and how Miguel and Nat, and several others at the time, were using "blogs" a long time before the term was really invented, as a sort of activity log. He also explained some of the reasons for having an offline publishing system, and how comments were not always a good feature, especially with the rise of comment spam.
We then turned towards topics of the future of the software industry, and whether there would be any future innovation in Operating Systems. I also told him about some of my plans for the immediate future, and what I am going to do after college (innovative/research-based software, with practical applications, e.g. like VMware), and then, unfortunately, had to get out of the train at the Palo Alto station (Miguel was going to SFO). Overall, that Friday was probably one of the coolest days ever.