Ok, so if I try to understand a bit the structure:
=> MOCanvas is just the 'global' Morph (and so not really a canvas ?) that will contains all the elements's morphs
=> This global morph know the MORoot element  .
=> This MORoot element know each sub-MONodeElement.
=> Then the MOCanvas Morph can be set to any MorphWindow (to the MOBrowser or MOEasel for example).
=> It's only when we try to open such a window, that we concretely 'draw' each node (dispatching the work to the corresponding MOShapes).
=> So the morphic library says to the MOCanvas: drawOn:, wich dispatch the work to the MORoot, wich dispatch to each MONode, which dispatch tho the corresponding MOShapes.

Like alex said, the concrete canvas to use is provided by Morphic, and if no one should manually instantiate it, Morphic should also be modified to use by default the RomeCanvas instead of the current ones (?).

What seems to be sure, is that we can't specify easily a specific canvas to be used in mondrian   

2010/6/15 Alexandre Bergel <alexandre.bergel@inria.fr>
> I did not say that he did not say anything. I just said that it is difficult to find where problems are by looking at a very complex case. When the problems were raised, we did look into the issues, but it was just too difficult especially when we had multiple problems in the same time.
>
> Now we are in the situation in which multiple people noticed multiple problems, there are some fixes, but there are still problems left. So, we have to dig deeper and we should do it more systematically by creating simpler cases that reproduce the problem. It's not an issue of "I told you so", it's an issue of "We need help to identify all problems" :).

+1

Alexandre

>
>
> On 15 Jun 2010, at 14:01, Stéphane Ducasse wrote:
>
>> Doru
>>
>> Jannik said it several times - check the list. So do not tell him that he did not want to help.
>>
>> Stef
>>
>> On Jun 15, 2010, at 1:25 PM, Tudor Girba wrote:
>>
>>> One more thing.
>>>
>>> It can also be that when you have expensive traversals in your blocks, you can easily get to quadratic algorithms.
>>>
>>> For example, suppose you have something like this:
>>>
>>> view shape
>>>     height: [:each | (classes select: [:target | target invokes: each]) size].
>>> view nodes: classes
>>>
>>> In this case, for each class, you would traverse all classes again, so N^2.
>>>
>>> That is why we need canonical examples first that expose complex graphs. And those people that have a direct interest in getting Mondrian fast would be good to help in this direction, because it can get difficult to understand the particularities of each model :)
>>>
>>> Cheers,
>>> Doru
>>>
>>>
>>> On 15 Jun 2010, at 13:13, Tudor Girba wrote:
>>>
>>>> Hi,
>>>>
>>>>> now something important is that mondrian is barely usable for us: DSM, package blueprint, torch are all slow.
>>>>> I would really like to see some pragmatic solutions to be found. Like not computing all the blocks all over the time.
>>>>
>>>> This is what we are already discussing, but until now it is not clear what exactly generates the slowness. I believe the problem comes from the edges, but we have to take a more systematic look.
>>>>
>>>> The other thing that has an impact is that right now Mondrian computes everything lazily, so even if the visualization appears, scrolling might still be problematic the first time you go through the entire picture. So, testing should also take this into account.
>>>>
>>>> Doru
>>>>
>>>>> Stef
>>>>>
>>>>> On Jun 15, 2010, at 12:43 PM, Tudor Girba wrote:
>>>>>
>>>>>> Hi Cyrille,
>>>>>>
>>>>>> The canvas is defined in MOCanvas. This canvas is embedded in MOBrowser and MOEasel, which in their turn create a StandardWindow that holds the MOCanvas.
>>>>>>
>>>>>> Just to inform the others, Cyrille is experimenting with adding the Athens canvas (which provides an abstraction over Cairo, Balloon and possibly others) behind Mondrian.
>>>>>>
>>>>>> Cheers,
>>>>>> Doru
>>>>>>
>>>>>>
>>>>>> On 15 Jun 2010, at 11:42, Cyrille Delaunay wrote:
>>>>>>
>>>>>>> Hello,
>>>>>>>
>>>>>>> I have been looking a bit for the place where we set the canvas to use in Mondrian, but I didn't find.
>>>>>>> I know that all shapes are drawn on a FormCanvas, but there is no references to FormCanvas in mondrian classes.
>>>>>>> I just saw that there is a 'defaultCanvasClass' method in the class Form for example, so maybe such a method is use somewhere in the code.
>>>>>>> Does somoene know the way to set a new Canvas to Mondrian ?
>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> Moose-dev mailing list
>>>>>>> Moose-dev@iam.unibe.ch
>>>>>>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev
>>>>>>
>>>>>> --
>>>>>> www.tudorgirba.com
>>>>>>
>>>>>> "Presenting is storytelling."
>>>>>>
>>>>>> _______________________________________________
>>>>>> Moose-dev mailing list
>>>>>> Moose-dev@iam.unibe.ch
>>>>>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> Moose-dev mailing list
>>>>> Moose-dev@iam.unibe.ch
>>>>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev
>>>>
>>>> --
>>>> www.tudorgirba.com
>>>>
>>>> "There are no old things, there are only old ways of looking at them."
>>>>
>>>>
>>>>
>>>
>>> --
>>> www.tudorgirba.com
>>>
>>> "In a world where everything is moving ever faster,
>>> one might have better chances to win by moving slower."
>>>
>>>
>>>
>>> _______________________________________________
>>> Moose-dev mailing list
>>> Moose-dev@iam.unibe.ch
>>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev
>>
>>
>> _______________________________________________
>> Moose-dev mailing list
>> Moose-dev@iam.unibe.ch
>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev
>
> --
> www.tudorgirba.com
>
> "Sometimes the best solution is not the best solution."
>
>
> _______________________________________________
> Moose-dev mailing list
> Moose-dev@iam.unibe.ch
> https://www.iam.unibe.ch/mailman/listinfo/moose-dev
>


_______________________________________________
Moose-dev mailing list
Moose-dev@iam.unibe.ch
https://www.iam.unibe.ch/mailman/listinfo/moose-dev