- From: Charles Pritchard <chuck@jumis.com>
- Date: Tue, 01 Nov 2011 10:24:26 -0700
- To: Lea Verou <leaverou@gmail.com>
- CC: Lars Gunther <gunther@keryx.se>, "www-style@w3.org" <www-style@w3.org>
> On 31/10/11 21:57, Charles Pritchard wrote: >> On 10/31/11 12:30 PM, Lars Gunther wrote: >>> Bringig this to your attention >>> >>> http://css-tricks.com/14771-a-call-for-nth-everything/ >> >> I've used "nth-char" as a step in an HTML rendering engine >> (canvas+js), not to select a single character but to update my >> current style for rendering inline text. It was handy and easier than >> specifying and calculating ranges. >> >> The nth-everything proposal does not include ranges. >> I'd like ::nth-letter(3, 10) -- an offset and a length. >> >> That'd save me a lot of unwanted <span> manipulation. >> >> -Charles >> > That could be done with a combination of ::nth-letter rules if we > become able to chain certain pseudo-elements (which is highly > requested in other threads, not sure how it's going): > Letter >= 3 is ::nth-letter(n+3) > Letter <= 10 is ::nth-letter(-n+10) > Combining them: ::nth-letter(n+3)::nth-letter(-n+10) Wouldn't a comma be a little more compact? > However, I have a hunch that these suggestions, even though undeniably > useful, will probably prove out to be very hard to implement in a > reasonably performant way :( They're going to be more performant than cluttering the DOM with span nodes. Given that nth-letter lacks semantic meaning, implementors won't have to clutter the accessibility tree either. nth-letter is specified in the same manner as ecmascript substr. Yes, it will break-apart ligatures when needed, just as <span> would. -Charles
Received on Tuesday, 1 November 2011 17:25:06 UTC