Re: a vote for rotation

really, I was thinking more of text rotation of at least 22.5 degrees (1/2 of
45...), which would make 16 positions for the text. That would be sufficient for
simple animation (and, of course, regular text), but it would be much nicer if
it were arbitrary. Images, on the other hand, should be arbitrary. I don't know
how systems deal with rotation of any sort of font, but I know that most do it
(having seen it). I'm wondering at what CSS version they will include bending
text along vectors. This is going to get very complex.

Then, what about the rest of the point transformations? Reflection, skewing,
etc. These are all photoshop tricks, as far as I know, and you can't use skewing
very easily (it would seem).

Clive Bruton wrote:

> Chris Lilley wrote at 27/11/97 8:19 pm
>
> >Hey, great! "90 degree rotations of downloadable fonts acceptable, says
> >Clive Bruton" ;-)
>
> Yes sure, as long as they're encrypted, expire 3ns after they're
> downloaded, the font vendor gave permission, the font "broadcaster" paid
> for the priviledge and (some other stuff I forgot)... CSSvX supports
> rotation of type. :-)
>
> >
> >> Chris made a few points about some platforms not dealing with rotated
> >> text too well, this is probab<www-style@w3.org>ly true. However any that
> handles TrueType
> >
> >It was TrueType I was thinking of, because the hinting mechanisms basically
> >assume no rotation. Alas, my knowledge of TT hinting falls off when the
> >actual assembler-like codes are involved so I don't know if it is actually
> >possible to treat vertical  hints as horizontal hints ( a quick look at the
> >spec [1] seems to indicate that this is so)
>
> The argument that's always been made to me about vectors vs bitmap fonts
> is that the vectors can be rotated (in any increment) and maintain their
> shape due to their structure and their hinting. Maybe I'll check on that
> elsewhere...
>
> >And of course for bitmapped font formats, if rotation is restricted to
> >90degree increments, creating rotated 'virtual' fonts is computationally
> >trivial.
>
> Absolutely, because of the nature of pixels this is probably the only
> desirable way to do it anyway (but I guess that's just me trying to
> enforce "good taste" on someone else who wants free rotation), until we
> all have 200-300dpi antialiased screens.
>
> >
> >> But some provision for an "alt" tag should be made, just as for browsers
> >> that can't (or won't) handle images.
> >
> >In the context of CSS, if a property is unknown the entire
> >rule is ignored so authors can provide two rules, one for the rotated case
> >and one which will be used if rotation is not supported.
>
> Sounds good to me, as long as the authors allow for the change.
>
> -- Clive