On Wed, 6 Aug 2003, Tuukka Toivonen wrote: > However, Gerd disagrees: > >gstreamer IMHO isn't the solution through. I doubt making gstreamer a > >requirement to talk to v4l devices would be accepted by the users and > >developers. gstreamer does to much (it is more that libasound for > >video) and has to much dependencies (glib, xml, ...). > Hmm, it is possible. This is been discussed on Gstreamer mailing list. But > it doesn't have that many dependencies really, except for the plugins > (which are all optional anyway). glib dependency is inconvenient, but I > think that an object-oriented approach is necessary, and unless you want to > use C++ (which many people don't want to use), there's little choice. 1. stability. the main idea is that if there are places with unstable base, there can be less stability in the whole system. I like when applications tell me "SIG11", not the kernel. 2. c++ is not a good idea. note, that glib/gtk are not c++ed. and these is nice. are You are trying to get more comfort for You personally? :) > However, I'd rather avoid writing a conversion library from scratch, since > after all, Gstreamer already exists. I'll reconsider this again after I > have actually written some Gstreamer code (so far I have just read the > manual). so, why not to split GS, if the conversion block is nice there?
