• Macros in Python
• Ideas for Mutable Strings in Python
• Explanations of itertools
• Python 2.3b2 with new IDLE

Norman Walsh is touting John Cowan's TagSoup as a solution for HTML escaping. Andy Clark's NekoHTML is also a solution in this space.

• Erann Gat's The Complete Idiot's Guide to Common Lisp Packages
• The Common Lisp Cookbook

Patrick Logan made some comments on the whole RSS/Echo situation that really resonated with me.

What's the best thing about aggregators? The fact they mainly visit weblogs? No. The fact they use RSS? No.

The best thing about aggregators is that they are better user interfaces for browsing the web than Mosaic, which is really what we're still using when we use IE.

Well at least Mozilla and Safari have tabs, for crying out loud.

What do we need, another Echo? Nope. We really need better applications. What happened to that Web Services revolution anyway, that the heat generated in 2003 is about HTTP.

Aggregators are such a better user interface that the aggregator has replaced the web browser as my UI to the web. Except for the fact that the aggregator that I'm using, Aggie, outputs a huge HTML page, due to my style of processing the aggregator output. You can't ignore the fact that the improved UI of the aggregator is due to a common data format, in this case RSS. I think that it may not be possible to improve the applications without improving the data format. The parts of Echo that are most valuable to me are the archiving use case and an improved/enriched editing/posting API. Tightening up the meaning of parts of certain parts of the spec would be valuable as well.

During my vacation, the ASF board adopted a resolution replacing me as the chair of the Apache XML project. I've been the chair for over two years, and it was time for a change of faces. Berin Lautenbach, the new chair is full of energy and motivation. I'll still be involved as a member of the PMC, working to facilitate smooth operation and growth of the project.

If you accept the position that standardizing stuff is a good thing -- and I think it is, but only for stuff that has matured, then Mats Henricson points out that the JCP is unusual in producing not only specification documents but also reference implementations and test suites (TCK's). The IETF's practice is to require two independent/competing implementations for something to be standardized, but doesn't require the test suite.

I agree with everything that Dana Blankenhorn wrote in his post on how to fight the RIAA.

But the best thing you can do is keep your wallet in your pocket. No matter how much "law" the RIAA may claim to have on its side you have the ultimate power. If you don't like the terms and conditions under which a product is offered to you, don't buy it.

There's one other thing he left out that we could do. Find a way to support artists who are producing music outside of the RIAA.

Well, we're back from our brief vacation. The Packwood area is beautiful (it was our first time), and our friends' cabin has a creek going right through the back yard. The kids loved it, Julie loved it, and I loved it. The only thing I didn't love was how short it was. As I slogged through the aggregator backlog, I came upon Russell Beattie's post on Americans and vacation.

I visited Spain last October, and I saw a lot to like, particularly in the way that Spaniards (and Europeans, by extension) use their time. I suppose being the best in the world at running ourselves into the ground is an acheivement, albeit pyrrhic.

I'll be offline for a few days. We're going to take advantage of a friend's cabin in Packwood. No internet, no telephone, and no TV reception!

• Eric Raymond talks about the relationship between hacking and refactoring, and between the open source community and the agile programming community. It's nice to see someone else picking up on this -- I submitted a talk proposal on open source and agile to OSCON, but it was rejected.
• ESR's piece then pointed to an old Philip Greenspun piece on professionalism. It's not exactly the hacker ethic, but its related to it.

Yasser Shohoud has an interesting article on the relationship of .NET and the WSE:

WSE is logically a part of the .NET Framework. Keeping it physically outside the .NET Framework allows us to ship new versions at a much higher rate than the framework.

This is the kind of modular delivery that I was saying Microsoft wasn't doing, but at least in this particular case, I appear to be wrong. I wonder how many other pieces there are like this one.

Hats off to Chris Brumme for his detailed and informative post on CLR reliability. A double hats off because he knew he would get beat up about it. I don't see the Sun VM people blogging about this kind of thing on java.net.

Over at Sam's wiki, folks are talking about a roadmap for some work that would unify the various flavors of RSS and build an archiving and blog editing protocol on top of that. I think that seeing some convergence in these areas would be a good thing. I'm not as emphatic on the standardization part -- looks like IETF is the direction at the moment, based on all the uses of Internet Draft.

Much has been made about Eric Kidd's essay on the future of the software industry. Today Krzysztof Kowalczyk posted a response that takes the opposite point of view.

My thinking lies somewhere in the middle. For all the noise that is being made about open source, we are still at the beginning of understanding what it is all about, how to organize it, and how to make a living off of it. Jon Udell's post about fit and finish can't be ignored. The open source community is still experimenting with different ways to organize itself. The organization of Linux, the FSF, the ASF, and JBoss are all different. It's not clear yet that there's a really good model for this. Just look at the discussions that we at the ASF are having this week regarding membership (which is related to how the ASF is structured). There's a lot more room for experimentation around many aspects of open source.

I'm an open source guy, but I also feel an affinity towards small developers -- must be all those years on the Macintosh, where the best apps were done by the small shops. The entire fabric of the software industry is changing, but that doesn't mean that there's no future. I don't believe that open source means that people doing software don't make a living. I do believe that it means we have to do it differently than we've done it before.

One last thought on this from Cory Doctorow's notes on Tim O'Reilly's Reboot talk (via Sam)

What keeps me up at night?

...

* Users don't own their data -- who cares about source when your data is locked in?

If you are a hacker (in the Paul Graham sense), then having the source means your data is no longer locked in. But for most people open source doesn't change this. This is one of the reasons I got interested in XML and open source, but I don't think that this problem (which is a user side, fit and finish kind of problem) has been solved yet.

It's a Macintosh kind of day today.

The WWDC is running, and Panther and the G5 are no longer vapor. (Where's that 15" G5 Powerbook?!)

I have to say that the Mac developer community is where the buzz is. I was totally bored reading the various Java One blogs this year, but a couple of the blog posts from MacHack just stirred the envy in me.

Eric Albert recounted the increase in Rendezvous usage. This stuff is very cool. I don't get why we aren't seeing more activity on Windows and Linux. Looks like a good opportunity for a small developer. Oh, wait, Microsoft and open source have exterminated all of those.

James has the most amazing journey. Somehow he manages to score a ton of old NeXT hardware. I was excited by the Lisp Machine that was on eBay last week, except that it had no display or keyboard. That and the fact that I have nowhere to put something that big.

Sometime today the reality distortion field will wear off. Till then I'm waffling over whether my next machine is a Mac or not.

David Lamkins, a major Lisp hacker, has written the book Successful Lisp, providing an overview of Common Lisp for the working programmer.

If you want your head to spin, check out this picture of the state of XML.

Don't use toString().

Phil Wolff's contribution to Sam's wiki is interesting because it hints at the direction we could go in terms of syndicating more structured information in addtion to regular weblog posts. I'm not sure whether this is feasible without some kind of safe/downloadable code model for processing structured information when your aggregator sees a new kind of information. His example of computing local soccer statistics in an aggregator is beyond the abilities of any aggregator that I'm aware of.

Naval Ravikant gives us a peek into what's hot, at least what VC's are seeing nowadays. Here's a summary of what they are seeing.

Predictions: most killer apps will emerge first via web-based GUIs (client side) unless they involve 3D graphics or heavy filesharing, in which case they're Win32 apps. Server-side killer apps will more easily emerge on Linux than on Windows. Some of the more interesting consumer-facing server apps are emerging just as quickly on Linux as on Windows (PVRs, online photo albums, music jukeboxes). Fear the Penguin, indeed.

Jeremy Hylton reports on the new itertools module in Python 2.3. I read the example and thought "this reminds me of streams from SICP". Then I went to the itertools docs and saw this:

This module implements a number of iterator building blocks inspired by constructs from the Haskell and SML programming languages. Each has been recast in a form suitable for Python.

Very cool.