W3C home > Mailing lists > Public > www-style@w3.org > April 2011

Re: [css3-lists] of lists and castles

From: Tab Atkins Jr. <jackalmage@gmail.com>
Date: Mon, 25 Apr 2011 17:40:37 -0700
Message-ID: <BANLkTik5kSo5G+_egFgh2s81E=smyeuOfA@mail.gmail.com>
To: David Singer <singer@apple.com>
Cc: www-style list <www-style@w3.org>
On Mon, Apr 25, 2011 at 5:26 PM, David Singer <singer@apple.com> wrote:
> On Apr 25, 2011, at 17:00 , fantasai wrote:
>> On 04/25/2011 04:11 PM, Tab Atkins Jr. wrote:
>>> On Mon, Apr 25, 2011 at 9:42 AM, Tab Atkins Jr.<jackalmage@gmail.com>  wrote:
>>>> Okay, then.  I think the styles are valuable to support, given that
>>>> they are used in real life.  Should we perhaps just limit the styles
>>>> to the range 0-9999?  That would cut out a decent chunk of complexity
>>>> (as it would limit them to a single "group") and still support the
>>>> *vast* majority of use-cases.
>>>>
>>>> I'd have to review, but I think this would also allow me to define
>>>> several of them using the 'additive' type.  A few would still have to
>>>> be explicitly defined (the Chinese ones, in particular, due to the
>>>> zero-collapsing rule they have), but it would be less than the current
>>>> set.
>>>
>>> Upon review, yes, I could do the Japanese and Korean styles as simple
>>> additive styles if I limited them to the range [0,9999] (or
>>> [-9999,9999]).  Chinese would still have to be specially defined, but
>>> it would be significantly simpler if also limited to that range.
>>>
>>> This seems like an adequate compromise - 10k should be enough for anyone, right?
>>
>> I suggest first publishing a draft with what you have now, and then cutting
>> it down later. Even if it doesn't wind up in CSS3 Lists, it's probably useful
>> information for other people, and we might want to use those algorithms in a
>> spec at some point in the future.
>
>
> You could set an explicit conformance limit, and note in the algorithm
> wrinkles and steps 'for those wanting extra credit'.  It's not harmful
> to document a robust and general algorithm, after all, just put "Note;
> this part is only needed for numbers >X, and X>9999 (the conformance
> limit)"

It is harmful, however, to have MAY requirements if they're not
strictly necessary.  Unless there's a very good reason, every
requirement should either be a MUST or shouldn't exist at all.  So I'd
rather have a solid strategy and just spec that.

~TJ
Received on Tuesday, 26 April 2011 00:41:25 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 17:20:39 GMT