Re: [css-text] New transform

> On Apr 2, 2016, at 00:14, Jonathan Kew <jfkthame@gmail.com> wrote:
> 
> On 1/4/16 15:09, Florian Rivoal wrote:
>> 
>>> On Apr 1, 2016, at 22:38, Jonathan Kew <jfkthame@gmail.com> wrote:
>>> 
>>> On 1/4/16 09:03, Henrik Andersson wrote:
>>>> The text-transform property is nice for the various effects that it can
>>>> do. However, I think that it could do more.
>>>> 
>>>> As such I propose a new value for the property.
>>>> 
>>>> The new value is rot13. When this new transform is applied the
>>>> characters in the text will be replaced according to the rot13
>>>> substitution cipher.
>>>> 
>>>> Here is an example of it being used:
>>>> 
>>>> <style>
>>>> .spoiler {
>>>>     text-transform: rot13;
>>>> }
>>>> .spoiler:hover {
>>>>     text-transform: none;
>>>> }
>>>> </style>
>>>> <div class="spoiler">Today is an important day.</div>
>>>> 
>>>> 
>>> 
>>> I don't think I support this -- it's much too limited in scope, being essentially an English-only feature. (What will rot13 do with Latin-script text that includes lots of diacritics? Let alone non-Latin alphabets?)
>>> 
>>> If -- and it's a fairly big "if", in my mind -- we want new text-transform capabilities, I think we should instead consider a more generalized feature that allows authors to specify the desired transform or substitution.
>> 
>> I agree. There are lots of potentially useful text transforms people can think of, but many of them are either gimmicky, or serious but targeting niche use cases. A generic mechanism for author defined transforms would be a better use of WG and UA vendor time.
>> 
>>> At a simple level, maybe a function modeled on the "tr" utility, which allows rot13 to be pretty trivially specified, but also allows alternative transforms, and could handle extended alphabets, etc:
>>> 
>>>  .spoiler {
>>>    text-transform: tr("a-zA-Z", "n-za-mN-ZA-M");
>>>  }
>>> 
>>> Or for a less cryptic (and more extensible) approach,
>>> 
>>>  @text-transform rot13 {
>>>    match: "a"-"z" "A"-"Z"; /* could also support U+XXXX notation? */
>>>    replace: "n"-"z" "a"-"m" "N"-"Z" "A"-"M";
>>>    /* could also think about other kinds of transform... reverse?
>>>       arbitrary text-munging functions? etc. */
>>>  }
>>> 
>>>  .spoiler {
>>>    text-transform: rot13;
>>>  }
>> 
>> I've looked into an @rule based approach a few years ago with some depth:
>> https://wiki.csswg.org/ideas/at-text-transform
>> 
>> I've added the rot13 example at the bottom.
>> 
>> I'm happy to reopen this topic if people are interested, as I think it's a neat idea, but it wasn't obvious then that this would gather implementor interest. If people are going to be interested in this for more than a day, I'll happily resurrect this.
> 
> Personally, I'm suspicious the original suggestion wasn't intended to last more than a day in the first place..... ;)

I suggest you go have a look at the rot13 example I put in the wiki, and decode it :)

 - Florian

Received on Friday, 1 April 2016 16:01:30 UTC