gzz-dev
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Gzz] Re: [Gzz-commits] gzz/doc/pegboard 1008/PEG_1008.rst 1009/PEG_


From: Tuomas Lukka
Subject: Re: [Gzz] Re: [Gzz-commits] gzz/doc/pegboard 1008/PEG_1008.rst 1009/PEG_100...
Date: Mon, 7 Oct 2002 17:06:24 +0300
User-agent: Mutt/1.4i

On Mon, Oct 07, 2002 at 11:37:37AM +0200, Benja Fallenstein wrote:
> Tuomas Lukka wrote:
> 
> >>+                     (Benja:) Ah. Now I understand somewhat... However,
> >>+                     even then, you cannot satisfy scaling in two 
> >>dimensions,
> >>+                     so it would have to be ``scale(int into, float 
> >>scale)``.
> >>   
> >>
> >
> >No, the point is that we define rendering text into a coordinate system 
> >that has
> >been scaled anisotropically to be undefined in AWT.
> >
> >And like I said, coordsys already allows you to do such a scale. scale() 
> >is just
> >shorthand for coordsys.
> >
> 
> Good point. It should be removed (and the functionality archieved 
> through some other means). 

I disagree completely. ;)

> Specifying width and height through scaling 
> isn't very sane anyway, since if you take scale *seriously*, stuff like 
> the lines of a box would have to be scaled, like in your images in the 
> vob techreport.

No, no. It all depends on how you interpret the box. It's not about taking
it *seriously*, it's about taking it as one transformation.

See below.

> One option is for coordsys to have width and height atop of everything 
> else. I don't like it-- 

Yes, me neither.

> it's much cleaner to have coordsys = 
> interpolatable transformation of pixels... Another possibility is to 
> simply use a translateXY coordsys, and make cs.toScreenCoords(0, 0) 
> return the (w, h) pair of the coordsys. That's hacking by using coordsys 
> as transformations, but it could be fine. 

No way.

> A third possibilty is to allow 
> vobs to have any interpolatable parameters, not just coordinate systems. 
> This is maybe the best general way to do this.

Actually, they already may; coordsys can be freely defined.
This would be one possibility, but read on.

These problems are pretty much solved by PEG 1010: with that PEG, we can
finally combine the coordsys in ways we want, and use this to get the effects
we want. For example, basing the text coordinate system at the (0,0) point of
the box coordinate system, offset by a small amount *in global coords*.

It's all finally coming clear.

        Tuomas





reply via email to

[Prev in Thread] Current Thread [Next in Thread]