[Development] Future of QtOpenCL

Sebastian Lehmann qt at leemes.de
Sun Jun 30 14:05:59 CEST 2013

Quoting Uwe Rathmann <Uwe.Rathmann at tigertal.de>:

> On Sun, 30 Jun 2013 12:36:18 +0200, Sebastian Lehmann wrote:
>> I see the strong developer community behind Qt as a huge advantage.
> Well, from the low traffic on the mailing lists on weekends it is quite
> obvious, that - even if being a real open source project - Qt is driven
> by companies with commercial interests. I would guess that there is only
> a very small minority of developers working on Qt projects beside there
> daily job.
> So IMHO in the end the right decision depends on if one of those
> companies will be interested in the code ( from the fact that it was dead
> so far the answer might be: no ). If not you might find yourself doing an
> one man show in an environment with tools and rules that are made for a
> big community project.

AFAIK QtOpenCL died because the office in which it was developed  
(Brisbane Nokia) fired a lot developers. In 2012, the office was  
closed. It simply wasn't continued in another office or by the  
community. And I want to change this. Maybe if someone does the  
initial work, a company is interested to continue. (I'm thinking of  
KDAB as an example, as they are maintaining Qt3D and I think they  
might be interested (in the future).)

> Over the years I have lost some of my illusions about open source
> development. I usually don't even get answers when asking for support for
> my project ( Qwt ) - stuff like testing code on specific platforms,
> setting up a nicer project page or helping with the docs. If you don't do
> it yourself, nobody does it.
> F.e. look at projects like pyqwt or qwtplot3d. Nobody ever took this code
> when the authors lost their interest - even if their outdated versions
> are still popular.

Thanks for being not overly optimistic. Maybe I'm too excited to see  
the problems. It'd be so nice having QtOpenCL as an official module.  
I'd love to put effort into this, even in my spare time. I don't say  
the module could be ready in a few months; it might take one or two  
years if driven by a few people from the community only.

I appreciate any further input. If it's a completely unrealistic plan,  
we shouldn't continue. But I'd love to.

More information about the Development mailing list