[openbeos] Re: BeAPI++

Adi Oanca wrote:
Axel Dörfler wrote:
I don't think it's worth to bury resources into something now that we don't need right now anyway.

Don't you think in the future?
Some things take time. A lot, I might say adding the research. If we wait until we finish R1, we need to do these things by ourself. Or we can recruit some new people to do this for us at that time.

If you know people who are chomping at the bit to do these sorts of things and extend the API, then maybe they should look into writing something like the Santa's Gift Bag package - that seems to have been fairly widely used and well-received.

I haven't had a lot of time to poke at Haiku development recently; I have a few ideas about things I'd like to add to the networking stack, and some other utilities that could be useful. But, I suspect that there are a lot of people out there like me who dream of being useful, but who may not actually end up contributing a lot just by being part of a "Team". I'm not sure that it's Haiku's job to give these people something to do; Haiku is about making the best (client) OS we can.

Instead, I'd argue for an approach similar to the Dr. Zoidberg model, where a group that's interested in improving the current status of something outside the Haiku R1 target work on their own project and get it working. If it is good, and they want to donate it to Haiku, I'm sure Haiku would be glad to accept it. On the other hand, if it's not the right API, then Haiku doesn't need to maintain a bad API, and can feel free to choose or develop something else. Be did this a few times with the APIs that they'd chosen - IIRC, the input_server API changed between R3 and R4.

So, if someone wants to work on an "XML Kit" or an "HTML Kit", they can certainly do so. I just don't think it should go into the Haiku tree until it's done.


Random thoughts from a wandering developer,
Evan Anderson <argent@xxxxxxxxxxxxx>

Other related posts: