Re: Motivation for having a 'text color' that is invisible

[Jens Getreu <getreu web de>]

Le vendredi 26 octobre 2007 Ã 09:03 +0530, Sameer Sahasrabuddhe a
Ãcrit :
> On 10/25/07, Lars Clausen <lars raeder dk> wrote:
> > > I would vote for correctness than ease of coding. If it takes some
> > > effort to go look at every object, then so be it. What you really want
> > > is that displaying some built-in text on some objects should be
> > > optional, then that is what it should be. Making it invisible obscures
> > > the intention. Who knows, when someone gets around to coding this,
> > > they might notice an abstraction that helps in ways not currently
> > > foreseen.
> >
> > Well, if text could be made invisible, then why not other things?  It
> > could make sense to have an "invisible" color -- any renderer can handle
> > that simply be not rendering the thing in question.  That way we provide
> > much more than by having an extra option to turn any text field on/off.
>
> This is a case of the classic arguments about content v/s style. To
> make a perfect analogy, if appearance defined the role of an element,
> HTML wouldn't need the H1 tag. An invisible object is not the same as
> an undisplayed object. It is still around, and participates in other
> interactions. For example, an invisible object near the border could
> result in a strange blank area on the periphery if you export a bit
> map. Demanding that every renderer handles this correctly can get
> tricky. If you group an invisible object with other objects, the
> border of the group would include a similar strange blank area. 

In TeX this effect is widely used as "struts" and quite convenient. 
Sometimes you want a group of objects or a page to take more space than
the visible parts of it.



Jens

> The
> invisible colour can also get reset when you are using the properties
> dialog to set the colour of multiple objects.
>
> Full transparency support is of course, extremely desirable as Mike
> says. But it is not the solution for this problem. And about notes for
> objects, we could actually define a standard "notes" property that is
> available for all objects!
>
> Sameer.