Re: [css3-text] combining transforms

On Mon, 13 Feb 2012 06:23:13 +0100, Koji Ishii <kojiishi@gluesoft.co.jp>  
wrote:

> I understand that and I agree. I'm just trying to list up pros and cons  
> of Florian's proposal, I said "weak" in the first two reasons and that  
> includes this one too.

Right. In my mind, the fact that ePub forked an unstable spec does not
prevent us from changing the spec later, but it does mean it would
be somewhat irresponsible for us to change the spec *without reason*.

That said, since the combinations that can be expressed with the existing
syntax can trivially be converted into the one I proposed, so epub UAs
have a relatively easy way out.

> As far as I think, I came up with 3 weak reasons to keep the current  
> syntax. If you don't want to include this even as a weak reason, that'd  
> be 2 weak reasons.
>
> On the other hand, the only reason to change the syntax is "the same  
> thing is doable by @text-transform rule," in a lot lengthy way, and  
> which is still under discussion.

The @text-transform rule is somewhat longer, and more expressive. It
is not incompatible with the existing way of combining transforms, but
it makes it redundant.

Also, if you consider the situation from a java-script point of view,
it would be somewhat unfortunate that the same combined transform could
be exposed through CSSOM in two different ways, making manipulation of
transforms more complicated than strictly necessary.

selector { text-transform: uppercase, full-width; }

and

@text-transform full-upper { conversion: uppercase, full-width; }
selector { text-transform: full-upper; }

would look very different under the CSSOM, despite doing the exact same  
thing.


> As I said, I'm totally fine to change the syntax if it is reasonable to  
> do so, or the proposal is strong, so I wanted to find reasons to change  
> but without luck.
>
> If we don't come up with any other reasons behind this, I'd prefer  
> keeping the current syntax. Again, it's not strong and I'd love to hear  
> any other good reasons.

Likewise, I think the two ways of combining transforms are not  
fundamentally
incompatible, so they could coexist, but for simplicity's sake, I'd prefer
having only one way.

Received on Monday, 13 February 2012 08:33:33 UTC