Hi Doru!
Thanks for having a look at Juraj's work.
A rule of thumb should be that shapes should be interchangeable. Consider the following Mondrian
One of the main hypothesis of Juraj's work, I have the impression, is that there is only one way to compose graphical widget, which is given by nesting ROElements.
The GridBuilder is about efficiently building grid, whereas MondrianBuilder is about efficiently build possibility nested graphs.
So far, I see that the GridBuilder should be a compelling mini language to build DSM, UML diagrams, Class Blueprint.
We have chosen the notation "view nodes: ... using: abuilder". Maybe the using: lexeme is not the best, since we mean grid builder here, and not shapes (as it was in Mondrian).
example: | builder | builder := MOFormsBuilder new. builder column; pref; center; row; fill; row; fill. builder x:1 y:1 add: (MORectangleShape new text: #asString; withoutBorder). builder x:1 y:2 add: MOChildrenShape new. view shape: builder asShape. view node: 42 forIt: [ view nodes: (1 to: 42). view gridLayout ]
In this case, I can keep the graph structure the same and simply use another shape, and it still works (In this case, the rectangle has a default way of handling children elements). view node: 42 forIt: [ view nodes: (1 to: 42). view gridLayout ]
Something that has always bothered me, is that it is not clear when the MOFormsBuilder is the appropriate construction to use. Consider the Jannik's DSM. How to visually represent it? By having a super-complex shape to which we do a " view node: myDsm" ? Or do you need an appropriate way to build the matrix?
You gave a nice use of the FormsBuilder. It makes me realize that it is useful to keep shapes and the structure separate. But why only shapes? And not layout or interactions?
I still think it's good to keep a shape be a way of specifying how elements look, not how elements are composed.
Yes. I agree!
I am realizing my ideas here are not well ordered. Defining and understand what is the scope of shapes, interaction and elements is tough.
Alexandre
The way we achieved this was through ChildrenShape, which simply offers a placeholder for where the children elements are rendered. The limitation of the previous design was that we only allowed one ChildrenShape. This is why we either showed methods or attributes in the uml class diagram.
But I think we could remedy the situation by having some sort of partitioning of the children. For example, we could tag children and then associate the tags to the different ChildrenShapes.
Ok, that is it for the moment.
Doru
On Apr 10, 2013, at 4:30 PM, Tudor Girba tudor@tudorgirba.com wrote:
Thanks!
Now, it will be easier to review :)
Doru
On Wed, Apr 10, 2013 at 4:27 PM, Juraj Kubelka juraj.kubelka@gmail.com wrote: Hi Doru,
thank you for the navigation. There is the script:
Gofer new smalltalkhubUser: 'JurajKubelka' project: 'RoassalProposals'; package: 'Roassal'; load.
ROGridBuilderTest browse.
Cheers, Jura
2013/4/10 Tudor Girba tudor@tudorgirba.com Hi Juraj,
Great to see this starting.
To make collaboration easier please do the following:
- create a separate repository on SmalltalkHub
- put your Roassal extensions there
- send a Gofer script to load this code easily in a Moose image.
(ask specific questions if you do not find out how to do it)
In the meantime, I will try like it is now.
Cheers, Doru
On Wed, Apr 10, 2013 at 4:03 AM, Juraj Kubelka juraj.kubelka@gmail.com wrote: Hi all, Hi Doru!
There is the kick-off of ComplexShape implementation in Roassal. If you file-in the attached code into https://ci.inria.fr/moose/job/moose-latest-dev-4.8/, you will be able to play with it a bit. There is ROGridBuilderTest test class with two examples.
Doru, I suppose you are almost only one who manage Mondrian's complex shapes. Would you please examine the code and give me your opinion? Thank you a lot. … Of course any one can give me some feedback :-) I would appreciate it.
For now I focused to do UML Class diagram. I know about blue-prints which can be done similar way. I have not seen other complex shapes. So there is an open space for other requirements which can impact interface and design.
Implementation breaks several tests because of change in ROContainer>>encompassingNestedRectangle. I removed a default extent (5@5) which is not wanted. I will fix the tests later when the implementation will be accepted.
Thank you for any response or question, Jura
<roassal-complexshape-export-of-roshape.png>
Moose-dev mailing list Moose-dev@iam.unibe.ch https://www.iam.unibe.ch/mailman/listinfo/moose-dev
-- www.tudorgirba.com
"Every thing has its own flow"
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
"Every thing has its own flow"
-- www.tudorgirba.com
"Every now and then stop and ask yourself if the war you're fighting is the right one."
Moose-dev mailing list Moose-dev@iam.unibe.ch https://www.iam.unibe.ch/mailman/listinfo/moose-dev