On Fri, 2006-02-10 at 18:58 +0100, Martin Gercke wrote: > Is there any rule based on which you could suggest who might look into > switching to 2.0 yet and who could have a look but should probably wait > because it's too soon? > For instance concerning the different platforms: mica2, micaz, telosb ... > and maybe concerning something like module-related, specific things ... > can't think of anything but probably there are some things that are > known to be not really "stable" or fixed in design yet so that switching > is not really recommended for now or the near future.
If you want to build on top of a bunch of existing 1.x services (e.g., collection, dissemination, time synch, localization), then switching to 2.x is probably not a good idea. While most of those services are going to appear for 2.x eventually, they aren't there yet. The only major part of 2.x that's expected to change soon is the decomposition of the MSP430 USART, so that it can better support resource arbitration. I also suspect that I'm going to tweak TOSSIM a bit, but this will mostly involve adding some additional functionality than changing current abstractions. Basically, if you want to build systems starting at the HIL level (sending packets, timers, sampling sensors), then I'd recommend moving to 2.x. It will make your life a lot easier. If you want to build systems that touch hardware specific functionality (e.g., crypto on the CC2420, a new sensor board), then I'd recommend moving to 2.x. If you would rather use a bunch of existing services (understanding that you'll probably have to spend a good deal of time getting them to work together), then I'd stick with 1.x for now. Phil _______________________________________________ Tinyos-help mailing list [email protected] https://mail.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-help
