Re: performance of goocanvas

[Ed Griffiths <edgrif sanger ac uk>]

Damon,

Yes that's a good point, we ended up creating far too many and the performance 
was shocking so we had a rethink.

There's probably an argument for some kind of two tier system where there are 
"proper" canvas objects and relatively few of these but then lightweight objects 
that are drawn within a "proper" object....

.....but I haven't thought any of that through.....

Ed


On Thu, 21 Nov 2013, Damon Chaplin wrote:

> On Thu, 2013-11-21 at 09:48 +0000, Ed Griffiths wrote:
> > Damon,
> >
> > so that's where we got to with the foocanvas....we didn't create a custom item
> > but in effect we did have one as we used a group parent and did all our drawing
> > within it's space....
> >
> > Interesting really that every now then a need pops up for a really fast 2D
> > canvas that deals with large numbers of lines/boxes etc...not an easy problem to
> > crack and have a good interface which is I guess why one doesn't really exist.
>
> I think part of the problem is that the GObject stuff was designed with
> widgets in mind, where you typically only have tens or hundreds of
> objects. Hundreds of thousands is a bit too much for it! (I wonder if
> the Qt Canvas is any better. I tested it once but can't remember.)
>
> People seem reluctant to create their own custom items too. Maybe if
> that was easier it would have helped. (I made it a bit too complicated
> with all the model/view stuff and the use of interfaces.)
>
> Damon

--
 ------------------------------------------------------------------------
| Ed Griffiths, Acedb/ZMap development, Informatics Group,               |
|  The Morgan Building, Sanger Institute, Wellcome Trust Genome Campus   |
|  Hinxton, Cambridge CB10 1HH                                           |
|                                                                        |
| email: edgrif sanger ac uk  Tel: +44-1223-496844  Fax: +44-1223-494919 |
 ------------------------------------------------------------------------


--
The Wellcome Trust Sanger Institute is operated by Genome Research 
Limited, a charity registered in England with number 1021457 and a 
company registered in England with number 2742969, whose registered 
office is 215 Euston Road, London, NW1 2BE.