Quote:
Originally Posted by tompe
And gcc is still here. If you had had another business model and contributed your superior stuff to gcc then the world would have been a better place...
|
We considered that. But we were unable to find a business model that would allow such a contribution while still paying us to do the work. Perhaps a smarter group of people has invented or will invent such a model; we didn't find one.
And the megacorp that bought us out simply trashcanned the product line, with no willingness to contribute to gcc (or other open source projects).
Supporting a team that works consistently at the cutting edge (but not quite the bleeding edge) of compiler technology is quite expensive; a number of large corporations and some universities (via gov't and corporate research support) do so anyway. Supporting a team that keeps up with the cutting edge,
while also producing tools that are reliable enough and user-friendly enough to be usable for real-world development projects is even more expensive. And it's pretty much not done anywhere any more. The folks at the cutting edge (or even the bleeding edge!) are building research artifacts. The important output for them is papers, not products -- which is exactly as it should be. The folks building usable tools track well behind the cutting edge. Again, this is normal. It's less risky and a whole lot cheaper than trying to ride the cutting edge. And gcc tracks well behind the best of the commercial folks (although
not behind the worst of them by any means). There's nothing wrong with that picture, by the way. It's the usual model for deep technology development.
Xenophon