> So it looks like there is a need to have something like Eigen in Qt, at least 
> for Qt 3D purposes (faster, better optimized classes, more complete API).

Not sure how that conclusion was drawn. More like the opposite. People who need 
something like Eigen can just use Eigen. There is no need to reinvent a 
half-baked alternative in QtCore. Whereas from Graphics (Quick, Quick3D, Qt3D) 
perspective this part of the discussion is irrelevant as the needs there are 
different, and those needs are better served by the existing gui/math3d, or, 
when it comes to 3rd party packages, glm (but definitely not Eigen).

Best regards,
Laszlo

 
-----Original Message-----
From: Jaroslaw Kobus <[email protected]> 
Sent: Friday, January 24, 2020 11:02 AM
To: Laszlo Agocs <[email protected]>; KDAB Mike Krus <[email protected]>; 
Konstantin Tokarev <[email protected]>
Cc: Konstantin Shegunov <[email protected]>; [email protected]
Subject: Re: [Development] Moving math3d classes from GUI to CORE



________________________________________
From: Laszlo Agocs <[email protected]>
Sent: Thursday, January 23, 2020 5:26 PM
To: KDAB Mike Krus; Konstantin Tokarev
Cc: Konstantin Shegunov; Jaroslaw Kobus; [email protected]
Subject: RE: [Development] Moving math3d classes from GUI to CORE

> Indeed, getting some of the SSE work from Qt 3D into Gui could be useful as 
> well.

> When it comes to 3rd party solutions, the graphics stack would most likely be 
> fine and happy with using glm (and so math3d could just go away), but that 
> would mean pulling in more 3rd party dependencies, which is not necessarily 
> ideal either. (in any case, that's a topic to be discussed separately)

> And yes, Eigen is probably a good example of something Qt should not be 
> pretending to be competing with.

So it looks like there is a need to have something like Eigen in Qt, at least 
for Qt 3D purposes (faster, better optimized classes, more complete API).
So either we use it directly (this option is rather not possible, looking at 
the previous feedback: licensing issues, big codebase, etc...) or we start 
developing it in Qt.
Yeah, definitely a topic for a separate discussion :) And there is a good time 
now for this discussion I guess.

Just my 2 cents.

Jarek
_______________________________________________
Development mailing list
[email protected]
https://lists.qt-project.org/listinfo/development

Reply via email to