Re: [css-shapes] basic shape syntax in prose

On 9/27/13 3:07 PM, "Tab Atkins Jr." <jackalmage@gmail.com> wrote:

>On Fri, Sep 27, 2013 at 1:54 PM, Alan Stearns <stearns@adobe.com> wrote:
>> On 9/27/13 1:19 PM, "Tab Atkins Jr." <jackalmage@gmail.com> wrote:
>>>On Fri, Sep 27, 2013 at 1:06 PM, Alan Stearns <stearns@adobe.com> wrote:
>>>> The normative grammar-based definition is not readable when precise,
>>>>and
>>>> not precise when readable. I think there's a false sense of accuracy
>>>>in
>>>> the format for functional notation, as we seem to be willing to fudge
>>>> whitespace rules everywhere.
>>>
>>>What do you mean?
>>
>> This definition doesn't precisely define what's actually allowed in
>>hsl():
>>
>> hsl() = hsl( <hue>, <percentage>, <percentage> )
>>
>> It's actually something more like:
>>
>> hsl() = hsl([ ]*<hue>[ ]*,[ ]*<percentage>[ ]*,[ ]*<percentage>[ ]*)
>
>That's not being loose - the definition of the grammar explicitly says
>that whitespace is allowed between any tokens.
><http://dev.w3.org/csswg/css-values/#component-whitespace>
>
>> -inspired is the key. CSSWG additions and conventions are what trip
>>people
>> up.
>
>They're still documented, and unambiguous, unlike prose.

But I see that prose is sufficient for defining these whitespace
conventions and the functional notation itself :)

I genuinely fail to see the ambiguities introduced by the prose versions
in the shapes module, but I'm OK with adding grammar back in as long as I
can avoid the misleading {3,5} and unreadable repetitions of
<length>|<percentage>. If <arg> should not be used as a local shorthand,
what do you recommend? Shall I make a local definition of #{A,B} that can
eventually be added to Values and Units level 4?


Thanks,

Alan

Received on Friday, 27 September 2013 22:26:52 UTC