From the ever-fabulous Lambda the Ultimate programming languages blog, I found an article on the recently published Revised⁶ Report on Scheme Status Report. Yes, it's a report on a report. Each new version of the programming language Scheme has an associated Revisedⁿ Report on Scheme, where n = a monotonically increasing integer. The current version is usually called "R⁵RS" Scheme, but a new version is in the works. The R⁶RS Status report is the status report on this new version.
And R⁶RS Scheme looks like it's got some very fun parts to it. First and foremost, a plan for a library (= module) system, which has probably been the single most difficult issue for making portable Scheme code. Next, a standard system for error-handling and exceptions. Also, standard records (AKA structures), and finally Unicode implementations.
I think this sounds like the finest kind of standardization process: working from the dozens of existing Scheme implementations and ad hoc community standardization efforts like the fabuloso SRFI (Scheme Request For Implementation) project. It's identifying a problem (differences between implementations) and supplying a solution.
There are some incompatibilities introduced too; top-level definitions are gone, which is a little sad but understandable. (eval) is going to have to (evolve) to deal with the new library format. These are unfortunate incompatibilities that serve a greater good.
It's strange to see some other ideas bandied about, though, like making pairs immutable (no set-car! and set-cdr! any more), and requiring the cdr of a pair always to be a list. That way lies top-down revisionist language design, which is madness for a language with the maturity of Scheme.
Codifying standard interfaces and implementations for features that are different in all Scheme implementations is a Good Thing; they're what everybody wants. Messing with the basics of the language is a Bad Thing; I don't think there's a Scheme programmer in the world who wants that to happen.
tags: scheme r6rs r⁶rs status report programming lambda the ultimate
NonCommercial
So, one thing that I find kind of sad is that some contributors to Open Source or Open Content projects make particularly emphatic anti-commercial statements. It comes up on Creative Commons mailing lists and even Wikimedia or Wikitravel talk pages. A typical comment is, ''I don't want anyone to profit from my works.''
It's strange that the worst possible thing people can think of happening with their work is that someone could make money from it. You rarely see, "I don't want anyone to use my work to twist minds, justify racism or break the human spirit," nor "I don't want anyone to use my work to hurt others or themselves." Why is the idea of generating value with a work the worst we can imagine, when so many more destructive purposes are served with pieces of information, art, and software every day?
It would be an interesting intellectual exercise to design a piece of software or a work of art from which no one could profit in any way, financially or otherwise. It would have to be uninformative, unpleasurable, inutile as a doorstop or sample bitstream. Of course, intellectual exercise is itself a benefit, so that'd have to be gotten rid of, too.
Here is my wish: I wish that my work on the Internet and elsewhere is of use to many, many people for many years in ways I cannot imagine today. I hope it makes people think, it guides them through difficult waters, and that it amuses and provokes. Good luck with that, whoever you are.
tags: sharing noncommercial cc profit money
FastCGI
So, I got Gauche-fastcgi installed on my server, as well as mod_fcgid. The builds were easy, and after a little bit of fiddling in the Apache configurations I had a sample FastCGI script running on this server.
With just a couple of lines of changes, I had my WiLiKi instance running FastCGI. Some httperf runs later, and I could see a 4-5x difference in speed between FastCGI and CGI. So, I couldn't resist: I've moved this whole site over to FastCGI.
I've noticed a considerably peppier response time. I'm going to try to add some more WebSoftwareFriendly optimizations to the WiLiKi software soon, which should speed up the site even further. Hurrah for that!
tags: fastcgi gauche-fastcgi fcgid wiliki httperf web performance
Long day
Maj and I went to see the local Movies 4 Mommies matinee this morning at the Guzzo Cineplex in the Marché Central. Movies 4 Mommies are special showings for parents with babies less than 1 year old; the babies can cry or yell or whatever and nobody gets upset because everybody's got one.
It was fun to go -- despite the name, I'd guess that 20-30% of the babies had daddies in tow, too -- and we met a few nice parents and cute bubbies. Unfortunately, the experience was ruined by the movie that was playing this morning: Click.
I'm not an anti-Adam Sandler snob; I'm fine with the fart jokes and the warbling voice. No sweat. But this movie was badly paced; the jokes were completely unfunny; the story was boring and the interactions were creepy and mean-spirited. It was our first day at Movies 4 Mommies, so that's probably the only reason we didn't walk out on this movie. It stunk really bad; the Rotten Tomatoes reviews of Click were pretty bad, too.
Afterwards we felt dirtied by the experience, so we stopped by Santropol for some lunch with our friends Tara and Chris. We got hugantic sandwiches and salads and veggies and Amita June made a fantastic mess.
For dinner tonight Niko stopped by; it was good to see him. He was in Paris while we were in California, so it was good to get back in touch. Maj made an excellent shrimp pasta and a big salad, which was good for this hot, humid weather we've been having.
tags: click movies4mommies adam sandler babies movies rotten tomatoes santropol amitajune tara chris niko
ClaimID
I just signed up for a new ClaimID account: http://claimid.com/evanprodromou . ClaimID is a service for accumulating links to your various profiles, clippings, and other media mentions. It means you're claiming an identity assembled on the Web.
I think this is a pretty good idea, but I'm not 100% sure why it's significantly better than my Me Other Places? page. I do have a lot of on-line accounts -- too many -- but I'm not sure ClaimID is a good way to organize them. I dunno.
It's great, though, that they support OpenID, which I think is the coolest shared identity system out there. Hopefully identity aggregators like ClaimID will start bringing some logic to our online identity systems. However... I'd sure like to see some other identity players, like TypeKey, starting to move towards OpenID.
tags: openid claimid identity typekey
People I like
Flipping through the Internet Identity Gang site, I tripped over a post by Bob Blakley on his blog, Ceci n'est pas un Bob. Bob was one of the members of the SAML committee when I worked on it for Outlook Technologies and then Securant. I found him to be an excellent explainer; I learned more on that technical committee about being clear and concise in technical communications than I'd learned before or since.
So it's great to see that he has a blog; I've bookmarked it for further reading.
tags: blogs bob blakley