I always thought that the "currently subscribed to" web service is pretty cool.  It took some doing, but I've gotten it working.  I had a strange roadblock on an error related to the first declaration of the channelUrl variable in Jon's script.  I changed its name to channelLink, and all's well.

The Radio help isn't much help at all; it's a hacker's environment.  You just start poking around and look at the samples and things start to come together.  I really started to figure things out when I explored the Radio application itself a bit.  The application itself seems to be a sort of object browser.

We (that is, we technical types) focus intensely on form when we write software. ...who else but programmers could seriously debate the merits of Python's significant whitespace? So it has always seemed odd to me that these same people...pay scant attention to form when writing email or other kinds of purposeful communication. Why not? The reasons are the same.  Jon Udell

Why not indeed?  I just ran into Dave's site the other day and spent a good hour reading.  I didn't even realize what he was doing, but I was hooked.  The whole idea of writing in a constraining format seems like nothing more than an excercise, yet Dave makes it work so well I never realized he was doing it.  Many of Bach's works were written as excercises, but a good musician makes that irrelevant.

Bad couple days for computers & me.  Saturday, couldn't get BSD to install.  Yesterday, trying to fix a problem with IE that cropped up (tags with target="_blank" quit working), I managed to completely nuke the .Net runtime.  Can't uninstall it, can't get the install to run.  Looks like it's time to reinstall the OS. 

Joel says that VS.Net is "sluggish".  What I've noticed is that it's really pretty zippy on my dual 733 MHz at work (well, it was until I hosed everything yesterday).  I also have a 900 MHz machine that doesn't run VS.Net nearly as well.  My guess is that the IDE is multithreading, probably to support intellisense.  I wonder if this is the future, are we going to need MP boxes to run applications well?

Xml2CodeDom [authorities here are alert]. I actually think that this is pretty cool. Yes, useless as a direct format, but I wonder if something like this could be used as a substitute for the unimplemented ICodeParser interface. I experimented with CodeDom back in August when I started working on automated unit test generation, but abandoned it because existing tests couldn't be maintained without ICodeParser to "rehydrate" the CodeDom from existing code. This could be a useful tool for that sort of application as well. I need to look into this (do I say that too often?)

Sam Gentile's Rant. Sam's rant has already generated a lot of discussion, so I'm sort of showing up after the party's ended. It seems like the crux of his article is that the important thing about .Net is the CLR & BCL. I suppose that's like saying the important thing about Windows programming is the i386 instruction set and the Win 32 API. This sits well with me because that's how I learn; I usually don't understand something until I understand the minute details. So learning .Net has been a bit different because Visual Studio is trying its best to hide things from me. I never really had that problem before because the programming environments did little or no hiding when I was learning new environments. So reading Richter has been hugely educational, because he's writing at the level I like to learn at.

I think that what set Sam off is something different though. What he starts off with is a complaint about a general tendency for people to get onto a Q & A forum, USENET, or a mailing list, and fire away with basic FAQ type questions, or questions that have been previously answered. First of all, I'm not speaking directly towards Sam's behavior; he contributes greatly and generously to several mailing lists that I subscribe to. I'm just talking generalities. Anyway, last week I read a post somewhere in Dave Mc Cusker's weblog (can't find the specific ref now) along the lines that programmers he works with don't read documentation, they'd rather speak face to face. This is more a problem of how to manage online dialog than it has to do with the intelligence of the person asking the question; people generally like to communicate as directly as possible and like instant responses. I haven't run into many unmoderated online forums that handle this well. Perl 5 Porters used to be good, when I was on that list in 1996. There was actually a great post there once, addressing the question "Why hasn't my question been answered?", that enumerated a list of reasons: because we're all busy, because nobody knows the answer, because the answer is RTFM, etc. I wish I still had a copy of that post. Anyway, it is the questioner's fault for asking a question where the answer's readily available , but it's the responder's fault for answering the question improperly. When something's in the archives, the answer is "read the archive", and if you want to be generous, a link to the archived thread. When it's in the manual, "RTFM", or "read the faq". And if you asked the question, please don't be insulted. The others in the forum are trying to manage the signal/noise ratio and increase the overall quality of the content, meaning less redundancy, in part. There is the problem of specific individuals who dominate the discussion with questions they could have answered by non-interactive means. I work with a guy who has a pretty good solution: he only answers a question once. People learn to bring a pencil & paper with them before they talk to Eric, because they won't get a second chance. It seems unkind, but Eric's time is in demand, and this way he gets to mitigate the impact of interruptions.

There you have it: answering a rant with a ramble! :-)