Memes to spread
Some more crazy ideas for CIA‘s future…
- “Sparklines.” A pretty novel (but not new) way to present data history inline with text. I’m already building up some graphing infrastructure for CIA around my ‘fidtool’ library. Just imagine how cool it would be to, at a glance, see the activity history for all the projects on the screen.
- Wiki integration. The whole “documentation” section of the site is really just a read-only wiki at this point. It generates those pages from reStructuredText documents stored with CIA’s source code. If the documentation browser were promoted to a full wiki, users could maintain their own client scripts, installation instructions, and such without having to bother me 😉
But actually, those are the most tame ideas that have been circulating. Some even more outlandish ones:
- Why not build the entire CIA site on top of an existing wiki engine? It would be great to have a wiki-like way for anyone to edit project/author metadata, but with some form of version control to allow rollback in the case of abuse. Really, I could take nearly any wiki engine out there and give it a special namespace for stats targets.
- I’ve been searching for a new way to organize the real meat of CIA’s site: the stats browser. Each stats target (one project/author/host/vcs/etc.) can have several types of content attached to it. It can have user-assigned metadata, automatically generated ‘related’ links, the recent commits, and a list of children. The big problem with this is that the larger pages (project, author, gnome) just have too much content to display all at once. For any target, large or small, I really want to see the most recent information first. This suggests somehow merging everything into a single chronological list. I’m not entirely sure how this will work yet, but I’d like to focus on the commit list but attach information about related/child items as appropriate. An important part of this would also be allowing the user to choose where to display additional information. This will probably mean “More…” links at the bottom of the page, plus some way of interactively expanding the inline information attached to each commit.
Amanda found a really spiffy web service: pandora.com. It’s a streaming music server with an intriguing non-genre-based method of categorizing music. You put in a handful of artists or songs you like, and it tries to stream similar music for you. It actually seems to be working pretty well so far.
Thanksgiving this year was great- I spent most of my time back in Colorado with my family, whom I’ve spent far too little time with recently. I baked an apple pie with my brother’s help, ate far too many of my grandmother’s homemade rolls, and really just got some much needed time to catch up. Much of the Boulder crowd was busy with other things, but I was thankful for the time I was able to spend with David and Jen.
I’m not sure when I’ll be coming back next. I’m trying to conserve both my money and my vacation time at this point. I’d like to do something for New Years’ Eve, but any plans for that are still in their early stages.