From: Mitch S. <mit...@be...> - 2007-06-08 10:39:15
|
Aanensen, David wrote: > have you by any chance seen MIT's Similie project - Timeline widget? Yeah, I've looked at it and read some of their code. Actually, that was one of the things I used to try and justify the div approach. The code is MIT licensed, so we could incorporate it if we wanted to. I like the popups that point to the thing they describe. The reason I thought it would be good to go in a somewhat different direction was because they have a slightly different set of goals and constraints. Scale, like you mentioned, is one thing that I think we're much more interested in. And scale concerns have driven some of the differences in how I'm doing things, like using the browser's built-in scrolling functionality rather than scrolling by changing the position CSS attributes. The zooming thing is another difference. In a genome browser, I think you're really interested in the relationships between things that are vastly different in size. For example, you want to be able to see how a SNP relates to a gene that's tens of thousands of bases, but on a timeline it's less important IMO to see where a single day falls in a span of a few hundred years. So zooming is something that I think we want to spend more effort on. And getting that to work drove some more differences with Timeline. Generally speaking, it's good to set a high bar for rolling your own implementation of something, but on the whole I think that our particular set of needs is fairly unique and deserves some amount of bespoke code. I do want to make use of as much existing work as possible, though, from simile, prototype, scriptaculous, dojo, openlaszlo, gbrowse, catalyst, wherever. Mitch |