On Sun, Sep 28, 2014 at 07:28:20PM -0400, Karl vom Dorff wrote: > Again, I think the GSOC projects are already well defined. I suppose it's > just a matter of defining how much should be spent on each project. The GSoC projects are (on purpose) items that are not part of our R1 target. We tried putting important features for R1 there in the past, and it has shown that the work done by the students was usually not up to our expectations, which in the end did delay the feature rather than push them forward. However, it does get us new contributors (including Pawel and myself), and that part works well. Haiku, Inc is trying to push Haiku towards R1. As such, it funds ideas that developers think are important/needed for R1. We all know where the money will go if we put the decision in the hand of users instead: USB3 support, UEFI support, and OpenGL acceleration. There are all nice features and would be great to have, but putting our developers to work on those will delay R1 further. Once again. Sometimes you have to let the developers finish the important groundwork tasks (such as the network kit and app_server enhancements I've worked on during the WebKit port to get the OS in shape for a release. As bounties are feature driven, they increase the feature creep. > > However, wait, wait, wait. If Haiku Inc. has money in the bank, then it > should/could be the backer - not waiting again for someone else to finance > it. > > I really do appreciate the things Adrien works on, but there are other > things Haiku needs as well - this is compounded by the fact that one can > already use a fairly decent QT browser. I think the whole process needs to > be handled more diplomatically. The users are those that use the system and > want feature x, y, z. At least a poll could be held again that asks users > what they want to see in Haiku (I'm sure times have changed and goals have > been met since that last one). If they all say a native web browser, then > so be it. I want a list of webcams that work under Haiku, maybe someone > else wants some functioning scanners. Have you tried Qupzilla before calling it a decent browser? It crashes every 2 minutes and has a very low html5test.com score for example (about 260, my current WIP of webKit are at 410 or more points). Things are being worked on on the Qt size, and competition is a good motivator for me. Finally, remember that it's nice to setup bounties, but you have to find a developer who wants to work on the features. I would start by asking the developers how they want to be paid and what they would like to work on (which is what Haiku, Inc. is doing). Bounties that don't get claimed are useless. And once again, bountysource is up and running and waiting for people to inject money into it. Karl, you can advertise this at Haikuware. We can also add information about it on the Haiku website. But I still think this doesn't need to involve Haiku, Inc. They CAN donate to bounties there if they decide that's a good way to spend their money. -- Adrien.