under a Creative Commons License.
[link] 2006-03-12T17:43:38 SHDH8 report: My semfile project is not especially active, but when Jim said he wanted to write some UI for storing local tags on local files, I decided to clean up semfile. My hope is that instead of starting over (which is tempting, since this type of project is very easy to start), he'll build on my semfile design. (That design so far consists mostly of "use rdf".) Anyway, I did some cleanup and wrote test cases for the cmdline tools. I also revised the file URIs to include hostnames, which is going to be an ongoing issue. Home users have various networked and non-networked files that they will want to tag. Someone told me in passing that beagle uses extended file attributes on certain file systems for some of its tracking. In particular, it may leave an id in the attrs to rediscover files that are moved. semfile has that concern, and I haven't even started to think about what solution I want. Suchi (no home page?) is interested in analyzing human behavior from passively-collected events, and I have years of logs from my various home automation projects. Kragen's cool PWM encoding demo on an atmel microcontroller reminds me that I should try my pixel project with atmel, since that project is largely stalled due to issues with programming PICs from linux (both the getting-my-code-written-in-non-asm and the install-code-into-the-chip types of programming). I have other projects that do boring queries on RDF stores, and they want to give really fast results. Kragen pointed me to some truth maintenance research. We also talked about how to optimize cases like mine, although I'm really interested in general-case optimizations that don't have to change when I adjust my queries. While SHDH7 was low on food I liked, this one was quite well-stocked.
| Unless otherwise noted, all content licensed by Drew Perttula under a Creative Commons License. |