[Kst] The problem with auto-sizing labels

[Brisset, Nicolas]

> > 2) when a user just clicks (without dragging to delimit a 
> rectangle) 
> > to create a label, it should use the default font size and 
> start from 
> > where he clicked
> 
> Like the current behavior, or something different?
Yes, that's the current behavior and I don't see any reason to change
it.

> > 4) at least in layout mode it should be possible to resize 
> a label by 
> > dragging the control points, adjusting font size automatically as 
> > required. Moving and resizing with the mouse is clearly the easiest 
> > way, especially compared to calling the dialog to try a 
> smaller/bigger 
> > font size until finding the appropriate one !
> 
> Agree - nice wishlist item.  I would suggest that only width 
> is adjustable.  
> Height is autoset from the resultant font size...
Sounds good to me, height is not the primary concern. Dragging the
corner control points should work, though.
 
> > 6) I don't see a problem with resizing labels due to 
> contents changing.
> > In many cases there should not be major length differences 
> and even if 
> > there are, I think the best hint at how to show the label is the 
> > bounding box delimited by the user at the beginning. If he does not 
> > like the new size, he can enlarge it easily (see 4) above)
> 
> Hmmm... this requires some thought.  What is more important 
> to the user?  The font size, or the box size?  You change the 
> content from within a dialog that has font sizes.  So you 
> would expect them to apply.  Sounds like we need a 'fit to 
> box' check box for the label font size which is auto set if 
> you drag-resize a text box.
I was thinking about the case where contents change due to some [item]
changing, but you are right in the general case and the "fit to box"
proposal you make sounds good to me.

> >The only thing that needs more consideration IMHO is defaults: 
> 
> And who is actually going to do the work... and what order to 
> do it in.  Some 
> of that code is in pretty bad shape - it should be cleaned up 
> before features 
> are added.  In my mind, the number 1 priority for kst right 
> now needs to be 
> stability!
You are right, that's a concern. But as I said at the beginning of the
mail, this was a user perspective and I was only trying to identify what
behavior we want. As long as that's not clear, we can't expect anybody
to do the work anyway. George indirectly asked for a clear spec wrt
labels recently, and I was trying to answer that request... And my
experience with kst is that most of the time developers are prompt at
implementing something once the design is clear. That said, I have not
looked at many places in the code so far and I can understand that parts
of it need to be cleaned up before we implement what I was suggesting.
Apart from that, I'd like to say that I find kst quite good
stability-wise right now, better as not so long ago (around 1.3.0). To
me, the only glitches are some parts still lacking a few features and a
bit of polish to be really useful. Probably my major expectation in the
short term is to be able to work efficiently with labels and metadata,
as almost all the functionality is here but there are still some
annoying gaps (like defining user strings, or the "flow-around" option
for view objects I asked for some time ago so that some metadata could
be pinned to the sides and not be covered by graphs when
redoing/cleaning the layout). Maybe what I should do is post a message
explaining the workflow I imagine with metadata ?
 
> >I can't believe we left this aside for so long just because...
> 
> And because no one has stepped up to do it.... (shocking!)
Right, that's a good reason :-) As you've noticed, I like to ask for new
things but I leave it to others to implement them (mainly because I'm
not a good enough coder and not familiar with most of the code anyway) !
What I meant here is that it did not sound so complicated (or dependent
on something else) that it should have taken so long. But I also forgot
about it, meaning that it probably wasn't high enough on the priority
list...

Thanks for your answers, at least I have the feeling that we have a
clear enough spec of what we want for labels now. Or are there different
feelings out there ?

Nicolas