I've always got a list of projects simmering in the back of my mind. Regrettably few of them ever make it to the foreground. This is a list of the most interesting among them.
SVG Animations
While I was an apprentice to Kim, I created a number of animations of line drawings using AutoCAD 9 and a good bit of AutoLISP. As animations go, they were nothing special. I was experimenting with the use of animation to communicate complex ideas, and in that respect they were very good. Although that was a lot of fun, and LISP eventually taught me a lot about recursion, I don't recommend AutoCAD as an animation tool unless you happen to like programming and weird geometry problems. I've been out of touch with Kim for a long time, but I would be very surprised if any of those animations have survived the onslaught of technological evolution. For many years I've wanted to try to recreate some of those animations with more modern tools. It's probably some of my most inspired work and I have no way to show it off. Most recently, I've started looking at SVG. It's certainly promising. But there are two immediate challenges to overcome. I'll have to detail those challenges in another post.
Update: one small step
Avalon Groupware
Roll a Turbine or Avalon version of various apps described in Practical Internet Groupware, hopefully connected to James as an IMAP server instead of NNTP. If I ever make that happen, it would be fun to see that become an example application instead of the TDK. I originally thought of this as a good way to grok Practical Internet Groupware, and a collection of toy problems to drive other learning, but as I've thought more about it, seems like a larger scale project. In fact, as I'm becoming more familiar with blogging, it seems like syndicating blog feeds into an IMAP folder could be a pretty neat trick to add to the tricks suggested by Jon.
Werkflow, not Valves
Experiment with using Werkflow in place of the valves in Turbine 3. This is two things: a toy problem to learn Werkflow, and confirmation about my claim that my pipelines and state machines are analogous. Update: That second bit isn't going to work out since Bob has moved to PetriNets instead of a state machine.
SEDA
Experiment with combining Werkflow and Avalon and maybe Messenger to create a SEDA architecture. This is an elaboration on the previous idea. Another toy problem to learn Werkflow, Avalon, SEDA and apply Messenger to something. It's probably re-inventing a wheel, but I don't mind doing that if it helps me understand something better. There's an outside possibility that these things could introduce some very cool functionality too.
Online community development
I'd like to publish my mom's book on education reform electronically to create magnet content for an education reform online community. I was largely inspired by Phil and Alex's Guide to Internet Publishing. Hopefully this community would become a source of ideas for open source tools to support education reform. I'd like to find a way to connect some non-profit organizations with open source hackers to create tools that are affordable and stable, and to give us geeks something with more enduring social impact than yet-another-catalog-with-shopping-cart. I've been thinking about this one for a while. I described it in a slightly different way in the context of a personal vision I wrote a little less than two years ago.
Glut-o-meter
Creation of a glut-o-meter to help individuals appreciate how the consequences of small choices can have substantial environmental impacts. Partly inspired by scorecard.org. I'd like a site where an online app could incrementally request info from a user as they explore the resources on the site and generate a consumption profile based on the information provided. It would then offer small measures to adjust their lifestyle based on the information gathered. It starts to spin into the idea of centering an online community around these apps and content to gather and evolve information about reducing the gross over consumption in the US. There are other people working on the problem of overconsumption, for example Arcosanti, and Segway.
I've been bloghopping all day. I keep stepping away from the keyboard, shaking my head about the time I've poured away. Each time I asked myself, "Self, what were you supposed to be doing again?" The original goal was to try to identify some blogs to aggrigate. Early this morning I managed to get the Movable Type plugin working with Blagg. That turned out to be a simple adjustment to mt.cfg.
Cool beans, I can now aggrigate blogs. But I immediately noticed that the aggrigated entries all had today's date, undermining some of the cool navigational instrumentation offered by Movable Type. So I dug into the code for Blagg and Movable Type to see if I could pull the dates from the RSS feeds and use those in the MT entries. I ran directly into the RSS morass. Now I understand what all the fuss is about.
"Self, what do we do now?"
"RUN, COWARD!"
Rather than joing the RSS battle, I decided that a little patience and judicious use of cron would let me enjoy Movable Type's instrumentation in my aggrigator. If I aggrigate the feeds daily, they'll be sufficiently in sync for my purposes. Only this first pull will be hosed.
That's what I did this morning. Then I decided to figure out which blogs I want to aggrigate. That's where the Energizer Rabbit started bloghopping. And kept going and going and...
The Movable Type blaggplug wasn't working for me until I discovered that I used a relative path for the DataSource in my mt.cfg file. It worked like a charm once I changed that to an absolute path.
I attended the Colorado Mac Users Group OS X meeting last night for the first time. The first presentation was observations from a couple guys who attended the Seybold conference. A presentation about why Rendezvous rocks. The short story is it's easier than AppleTalk, isn't chatty, and leans on completely open technology. A review of why changes in Darwin are important to even the non-geeks. It was a long list, but the thing I liked best: the simple addition of Ruby, Python, and the aquafied Tcl/Tk opens MacOS to 30,000 programs from the Linux world.
Anyway, the first presentation is the purpose of this entry. It seems the Graphic Arts industry is not switching to OS X. The consensus in the industry seems to be that, contrary to what Apple would have you believe, OS 9 is still the world's best graphic arts platform. For small outfits, or individual designers who mostly need to push Photoshop and Illustrator around, OS X is fine. But for the businesses that are running production printing operations, OS X isn't there yet.
In particular they cited "font madness", the lack of printer and scanner drivers, the lack of plugins for Quark and Photoshop as the largest contributing factors. The cost of upgrading was also cited, but seen as trivial in comparison.
Their impression is that the graphic arts industry looks a bit like a "deer in the headlights" at the moment. It's clear that the change to OS X will be painful and expensive. They seem to be supressing a question in the back of their minds: "is it time to change to Windows?"
In contrast to this there's Tim O'Reilly's article about Mac OS X Switchers. A lot of alpha geeks are coming to appreciate the Apple community's long focus on the real power of ease of use (as opposed to the power of MHz and good shells). When most things "just work" it leaves a lot more time for hacking on the fun stuff.
It'll be interesting to see how Apple adapts to a changing user base.
Need to come back to this link too. Update (12/17/2003): Wired has an article about Pierre Teilhard de Chardin and his noosphere: A Globe, Clothing Itself with a Brain. It's likely to be a more reliable link.
Noosphere: The Emerging Web of Consciousness
Many years ago (1991 or 1992, maybe), early in my exposure to the 'Net, I was surfing Gopher sites and found myself pulling documents from a computer in Europe. That's when I first groked the power of the 'Net. Some time after that I described the 'Net to friends and family as a nervous system for the planet, where the people on the other side of the computers were neurons and the computers and networks were the synaptic links between brains.
In the early days of the Web everyone was using the "Information Superhighway" as the metaphor of choice for the 'Net. That describes what the network packets are doing, but it doesn't describe what the people receiving those packets are doing. I tried hard to get friends and family to think in terms of the people who were being connected by these computers, and not about what the packets were doing. I shuddered at the ugly imagery in early ISP marketing: "your onramp to the information superhighway." In retrospect, I'm not sure why greymatter should be more attractive than highways. Maybe because it's organic? Anyway, seems I independently found myself talking about a noosphere.
Tim Berners-Lee has persuaded me that Cool URIs don't change. He advises that URIs be designed without classification or categoriztion information. Those groupings are likely to change over time which will work against the goal of stable URIs. In contrast, in Practial Internet Groupware, Jon Udell suggests that the URI is prime real estate for metadata. Although Jon makes an excellent case for the value of meta data in the URI, I favor Tim's advice in this case. The meta data will change over time. I decided to sacrifice the use of that prime real estate in favor of the stability of the URI over time.
Given Movable Type as my chosen publishing system, I had a few other constraints within which to work. Both Tim and Jon recommend including the creation date in the URI. It's valuable meta data and it won't change over time. However, Movable Type generates the permanent entries with a serial number and not a creation date. I lost more meta data, but protected against changes in the calendar system as some have proposed. :-)
The default paths for blogs in Movable Type steered me in the wrong direction at first. By default the indexes go in www.example.com/ and the permanent archives go in www.example.com/archives/. That makes for longer permanent urls. I wanted something that was more like Jacob Nielsen's Alertbox: http://www.useit.com/alertbox/20010204.html. I also liked the suggestion from Tim to put the recent stuff behind a foo/recent/ URI. Fortunately, Movable Type allows for this sort of organization, it just isn't suggested by default. The organization of Jacob's Alertbox suggested I should come up with a column name for my blog.
It took surprisingly many variations to finally settle on putting the indexes in dobbse.net/, and the permanent archive in dobbse.net/thinair/.
Edit: I was wrong about limitations imposed by Movable Type. I found this comment while bloghopping: (like Movable Type, where your archive URLs can just be named by the title of the entry, or the date, etc) Stupid Fool: Complexity of Implementation
Edit: After learning how to control the naming of permanent links in Movable Type, I broke the rule about not changing urls. This site has only been live for a couple weeks, and from the stats I get back from the web host, I'm almost the only one hitting the site right now. I won't have another chance to make this change without breaking links. So I put monthly archives in /thinair/yyyy/mm/index.html and permanent links in /thinair/yyyy/mm/0000id.html. I don't think that even on a busy day I'll post so many items that I'll need to separate posts by day (like this: /thinair/yyyy/mm/dd/0000id.html), so I chose to keep the urls a few characters shorter.
Update: I've done a minor redesign of my URLs.
At 1,655 meters (5,430 feet) above sea level, and roughly 1850 meters (6,070 feet) below timberline, the air here in Boulder is pretty thin. "thinair" is a persistent disclaimer that there's a remarkable lack of oxygen around here, just in case I post something that needs disclaiming. :-)