- From: Tab Atkins Jr. <jackalmage@gmail.com>
- Date: Tue, 6 Dec 2011 10:43:51 -0800
- To: Simon Sapin <simon.sapin@kozea.fr>
- Cc: "www-style@w3.org" <www-style@w3.org>
On Tue, Dec 6, 2011 at 5:02 AM, Simon Sapin <simon.sapin@kozea.fr> wrote: > Hi, > > Section 8 of Lists 3 has an algorithm to "generate a counter > representation". Step 2 says to use the absolute value for the rest of the > algorithm, but this is not what we want. For example, with the 'repeating' > type, we want the values -1 and -2 to be represented by the last and > second-to-last symbols, respectively (assuming two or more symbols). Using > the absolute value would result in using the first and second symbols. > > The "absolute value" part of step 2 should be removed and the algorithm for > some type adapted: > > * repeating: the modulo on negative dividends must be consistent with the > above > * numeric: take the absolute value before the rest of the algorithm > * other types: unchanged. I agree, and was thinking about this very issue yesterday. Good timing! Fixed. (I also made additive styles use the absolute value+negative sign thing, as they're used to generate some complex numeric styles.) > While we’re at it, it would be good to replace every occurrence of > "positive" and "negative" (for integer values) by either "strictly > positive/negative" or "positive/negative or zero". The CSS specs use the proper definitions of "positive" and "negative" (excluding zero) throughout, so I've stuck with that generally. However, I have adjusted a few uses of "positive" to "strictly positive" where I think it helps clarity. > PS: I write about various issues as I find them. Would it be better (less > spam-ish for the list or easier for editors) to aggregate them and send a > single email every now and then? Whatever is easiest for you. Multiple small emails are fine for me, as it's faster to address them, but a single large email is also fine for me. It makes no difference to me. ~TJ
Received on Tuesday, 6 December 2011 18:44:40 UTC