On Thu, Jul 17, 2008 at 9:37 PM, C. Scott Ananian <[EMAIL PROTECTED]> wrote: > Lots of reasonable points made on this thread. > > The two cents I'd like to throw in are: > $0.01: we shouldn't feel like shipping unsugarized apps is a failure: > better an working app w/ crappy UI than no working app at all! > $0.02: my suggestion to "replace" Browse wasn't to eliminate the > sugar-specific UI work, simply to suggest that we could more > profitably base it on Firefox than Gecko. Similarly, minimizing the > differences between upstream Abiword and write is (IMO) a Good Thing. > We should keep our forks as small as possible, so that we can most > effectively use the work being done upstream. > > For Firefox, that means (for example) that we can use upstreams > Awesome Bar instead of reimplementing our own url completion. For > abiword, it means acknowledging that a lot of our initial Tubes port > was/is simply unnecessary now that we have a stream-based > collaboration mechanism, and we can/should be able to strip down Write > as a consequence. It's possible that we can most fully utilize > Abiword/GTK's theme mechanism to make Sugar UI "upstreamable" as well. > Again, the point is to reduce our diffs with upstream.
Yes, I agree that this is a goal that makes a lot of sense. Unfortunately, my experience says that the approach you are suggesting won't be less work than what we are doing right now, because the software components you mentioned aren't so easily malleable as you seem to think. Check out the sources for abiword and gnumeric and grep for MAEMO, do you think those projects will let everyone add their ifdefs to suit their UI choices? Checkout microb-engine from maemo, they include their own patched mozilla. This approach might work well for Nokia and their dozens of engineers working on Maemo, but for the Sugar guys? At this time we would be even more insane than we are and we would have provided a much worst experience to kids. Seriously, embedding a gtk widget like the ones we have in Read, Write and Browse gives a pretty sweet spot in customizability. Adding some buttons and calling methods on that widget is not hard, we actually reuse all the hard work in the upstream project while choosing carefully the way in which we expose that functionality to users. If we count the amount of man-hours that went into those activities and told the nokia executives in charge of maemo, I think that they would be quite surprised... And then, having children and activity authors in general being able to read the code and embed those widgets in their python activities... that's invaluable, in my opinion. A maemo-tinkerer would need to set up a build box in order to add a button to the toolbar of one of those apps. Regards, Tomeu (sorry if I have offended anyone regarding Maemo. I know little about it, just have seen how they integrate with upstream projects and wanted to make the point that this wouldn't work for us) _______________________________________________ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel