Disk Clock 0.8. Another fairly restrained release due to a martial arts seminar. A few small changes with big visual impact, and some big changes with no visual impact ;^)
- Startup delay is hacker configurable.
- Reinvented default clock setup.
- Reversed default color gradient.
- Experimental feature: auto show.
- Internal debug refactoring.
More parameters exposed in the constructor options. The startup delay introduced last time can now be changed easier, and the default hover mode and set of disks is exposed.
Reinventing Time, Again
While in principal the week disk allowed you to see the progress of a day in a smaller visual field, in practice the day/night coloration of the day disk provides pretty good context and I wasn’t really using it. Things are a bit less colorful without it, but I’d rather have the greater visual resolution.
I also found that I was having to scan the entire 15-minute disk too much. I replaced it with the hour disk, but then found it’s marks too small, so I added 15-minute back again. That leaves it the same size as before, but it wasn’t really suffering for resolution anyway.
I would sometimes notice the changes in the moon, and then find it curious that this surprised, given that I’ve got a moon disk in the default hover configuration. I don’t hover very often, however. (Perhaps I will more often without the week on the main display). In any case, there is an experimental feature to switch to the hover mode for a few seconds when activating Dashboard.
I’m fairly uncertain about this one so far, because it often reduces my resolution on time right when I’m trying to look at it. It might be better to pop the calendar up as inside disks, but architecture isn’t designed to handle this right now – everything is sorted by scale, which make arbitrary combinations match up better. The timeout is hacker-configurable, and zero will effectively disable it.
I originally did the basic colors on the metaphor of the time period filling up from white to color. It never felt quite right however, so I reversed it for a radar-sweep metaphor. So far, I like it much better.
I rewired my internal debug reporting system to support a more plugin like architecture. I don’t know yet if it would easily support external plugins, but I did accomplish my immediate purpose: there is a new store-and-dump backend that will allow me to collect message before the DOM gets set up, without resorting to alert() or debuggers. Of course, since the debug system was itself the most significant code change, the new system hasn’t really been battle tested yet.