Flutterby™! : Microsoft tools 2003-05-31 01:39:33.020281+00

Microsoft tools

2003-05-31 01:39:33.020281+00 by Dan Lyke 5 comments

There have been a couple of entries over at Idle Words recently that deserve front page coverage here. Two days ago Maciej reported being contacted by a recruiter for Microsoft. But today he reports that the recruiter called back, having trouble understanding why no one was using Microsoft products to develop web search engines:

I've been scratching my head over this baffling little datum. Take me, for example. Why did I write my web crawler in Perl, using the excellent Web crawling modules already available for free on the CPAN, instead of paying $2,000 for a MSDN subscription? Why did I download existing GPL'd code for my language identifier, instead of taking three weeks to write a C# text parsing library from scratch? And why did I store the whole thing in a MySQL database on Linux, at a cost of zero dollars, instead of paying a couple of grand for a Window/SQL Server installation that would do the same thing?

That's right, because I'm a religious zealot!

Now I got my MSDN license for free, but I still find the development environment and tools available under Linux far superior in terms of usability and capability.

But it goes deeper than this. At the .NET Server 2003 launch, one of the reasons even the most Microsoft friendly of my coworkers rolled his eyes and left in disgust is that Microsoft isn't solving the problems that innovators need solved. All this .NET stuff is about building a better COBOL, and big corporate customers will not be where the innovation comes from.

Which leads us to patents and IP issues. Idle Words also has a report of finding "prior art", but what most amazed me about the details of this is that I've done some principal component analysis code to play with protein expression datasets, and as soon as I understood what I was seeing I thought "hey, this would be really cool for grouping documents". That's what the patent (and the long previous paper) covers. How do we reform patents so that the obvious stuff, the things nobody even bothers to point out, don't get put in place because the patent examiners don't work in the field? Alas, when we have multiple patents with the same claims coming through, I suppose we have larger issues in the USPTO.

Of course even solving these problems won't make patents ethical.

I think you've touched on the true viral nature of the GPL. Developers who use GPLed (and freely-redistributable) software are more likely to use that license on their own software. As this phenomena continues, more Free Software arrives. Its a nice little feedback loop.

After the past 20 years of the GPL, I think we can say that its viral nature isn't quick, but it is powerful.

I don't release my own code under anything but the GPL; anything else is just too much work. (Latest project here.) "Not quick, but powerful" is probably a good way to describe any system powered by human laziness.

Their development tools have locked-out programmer-centric tools like vim and emacs. They ignore the millions of man-years developers have invested in training their fingers to hit ^X^F or jjjksw etc. To force MS editing tools on developers is literally PHYSICALLY STRESSFUL and I simply don't use them whenever I can avoid it.

VC6 at least let you remap things -- .NET denies the most power of power users simple tools to model their work environments. There can be no possible explanation other than direct and open contempt for their users, customers, and the creators of the products they themselves thrive on. :/

VC6 at least let you remap things -- .NET denies the most power of power users simple tools to model their work environments

Which is really funny because the NT development team uses (or at least used to) vi-based editors and various *nix-like command-line tools.