Re: My proposal (Was: Shapes layout proposal)

[Lars Clausen <lrclause cs uiuc edu>]

On Thu, 14 Jun 2001, Andre Kloss wrote:

> On 14 Jun 2001, Lars Clausen wrote:
> 
>> > 1. Abstraction of subshapes
>> Sounds good.  Nothing problematic there.
> I'm glad you like it ;)
> 
>> > 2. Subshape-Arrays
>> With this comes the possibility of multiple levels of nesting and
>> resizing based on contents.  Fortunately, there's been a lot of research
>> on that for GUIs.  The question is what of it we don't need.  Do we want
>> for instance shapes that divide the containers space evenly between
>> them?
> Well, _I_ don't need them. ;) It depends on what is easier to
> implement. I think the best solution will be to give the user the most
> possibilities to push things to where he wants them.

The reason for GUIs to have the equal-spacing setup is that the actual
widget layout may be determined by somebody else.  I don't think we really
need that here.  Just using relative resizing is probably more useful and
easier.

>> > 3. Subshape resizing behaviour
>> Also sounds good.  It has always irked me that the UML class is not
>> resizable.  And there's no reason why text couldn't resize as well, by
>> changing font size (it might not be a fully proportional resize, but the
>> shape layout algorithm could take care of that).
> This goes in fact beyond the scope of my proposal...but why not? If
> someone implements this, fine!

I do want text to be resizable eventually.  If you don't have that, you
can't, say, have classes of different sizes in your diagram look nice.

> Another point is the behaviour of non-minimal-sized subshapes if a
> shape is made smaller. There are some possibilities:
> 1. Make every subshape smaller if possible
> 1.1 by the same value (ie. x1=5, x2=27 -> x1'=4, x2'=26
> 1.2 with the preservation of size relation 
>     (ie. x1=6, x2=8 -> x1'=3, x2´'=4, What happens when min(x2)=5?)
> 2. Take the first subshape and make it as small as needed or possible,
> then go to the next and so on until every shape has minimum size.
> 3. Don't let the user make shapes smaller if it's "filled" with
> subshapes. (Zero-implementation)
> 
> I think 1.2 is probably the best-looking solution, but It's somewhat
> hard to implement, I presume.

If text can be resized, there's no reason to have a minimum size for
anything, is there?  Then resizing would be scaling, and thus 1.2 is the
Right Thing.  Would you want resizing to do other than scaling?  That might
be useful, but makes resizing more complex.

-Lars

-- 
Lars Clausen (http://shasta.cs.uiuc.edu/~lrclause) | Hårdgrim of Numenor
"I do not agree with a word that you say, but I    | Retainer of Sir Kegg
will defend to the death your right to say it."    |   of Westfield
    --Evelyn Beatrice Hall paraphrasing Voltaire   | Chaos Berserker of Khorne