[CSS21] some feedback on PR

> From: www-style-request@w3.org [mailto:www-style-request@w3.org] On
> Behalf Of Bert Bos
>     http://www.w3.org/TR/2011/PR-CSS2-20110412/


I used the PDF version:
        http://www.w3.org/TR/2011/PR-CSS2-20110412/css2.pdf



For much of the CSS2.1 spec, I’m a first time reader so bear with me...
1.      p32, font shorthand - It would be good to see some discussion of how “h1 { font: bold 3em/5em Helvetica }” behaves.  My reaction on first read was that if you try to break the shorthand into individual properties, the interaction line-height and font-size with initial state will lead to some “might be surprising” results.  I may be misunderstanding though.
2.      p45, Intrinsic dimensions
a.      I don’t see “intrinsic ratio” defined anywhere, though referenced frequently.
b.      Can it ever be negative?
c.      Can it ever be zero or infinity?
3.      p63, below “ex” - Formatting issue: extra bullet.
4.      p75, “For example, ...” - Formatting issue: the Greek letters are showing up as boxes in the PDF form.
5.      p82, 5.8.1 - Can these attribute selectors be applied to the style attribute itself?
a.      h1[style~=blue]
6.      p123, double - “The sum of the two lines and the space between them equals the value of ‘border-width’.”
a.      I found this phrasing confusing.
b.      Also, a specific dashed rendering qualifies as a valid “double” rendering by this definition.
c.      Note that this definition is reused on p295.
7.      p146, image at the top - First sentence of second paragraph misspelling: “propery”.
8.      p151, 9.6, 3rd sentence - When I first read this, I found it confusing.  I was under the impression that the APE content shouldn’t flow around floater siblings but *should* flow around floater children of that APE.  And now, I’m not sure if that’s correct or not.  Examples would help make it crystal clear.
9.      p161, diagram, Inner Contents - The axes of the coordinates are backwards.  It should be (-100, 200) as it is in the diagram on p162.
10.     p162, hyphen example - Is there an expected relationship between the width of the pair of hyphens, 1em, and 10px?  Or does the example break if the font size or face is changed such that the dashes become relatively wider?
11.     p174, <percentage> - “...the resulting layout is undefined...”  A good example describing a scenario that can trigger this would be helpful for people trying to understand this behavior for the first time (like me).
12.     p193, <length> - Does “means the same as”, um, mean the same as “behaves as if” in CSS3?  Or does it mean something more/less?  I guess what I’m getting at is that I’m unclear how far this goes.  Do they have the same computed value, for example?  In this case, it seems like that’s not the intent.  Is there somewhere where “means the same as” is defined in CSS2.1 (or CSS1)?
13.     p196, last paragraph - “it should be inserted between the inner border edge and the outer padding edge.”  Prior to reading this, I thought that scrollbars were supposed to eat into padding space, and then content (but I’m not 100% on that).  This phrasing suggests it should allocate additional space of its own - such that you have the content box, then the padding box, then a scrollbar width, then the border box.  A diagram would be helpful here.
14.     p199, 2nd to last paragraph - “Content that has been clipped does not cause overflow.”  I think the intent was to say that the content *excluded* because of clipping should not cause overflow.  Inserting the word “out” after clipped would make that clearer.  On this page and subsequent pages, “clipped content” is used to refer to elements that have clip set at all, elements that have clip different from border box, the part included in the clip rectangle, and the part excluded from the clip rectangle.  I find it confusing to loosely refer to “clipped” without being explicit about “clipped out”, “clipped in”, “clip excluded”, or “clip excluded”.  The first paragraph of p200 has similar issues.
15.     p207, ‘quotes’ value - There’s a formatting issue in the PDF file.  In “[<string><string>]+”, the brackets are *very* subtle because they are somewhat obscured by the boxes around both “<string>” occurrences.  Recommend adding a space before the < and after the > to avoid the formatting issues.  (Or add margin around the PDF box style if that’s possible.)
16.     p213, both examples - I think it would be useful to reiterate that the outputted string is one less than the value in the “(=1)”, “(=2)”, etc. portion of the comment for each row.
17.     p213, 2nd example - “the style rules given above to number chapters and sections” refers to p212, where there are 5 separate such style rules.  I think it would be best to just repeat which style rules are being referred to, instead of having the unclear and prior-page referencing.
18.     p214, “h1::before” - Typo (two colons)?
19.     p224, “The page selector...” - Second sentence of this paragraph is missing a period.
20.     p228, below “right” - “A conforming user agent may interpret values ‘left’ and ‘right’ as ‘always’.”  Similar concern as with the p193 language “means the same as”.  If a UA decides to exercise this “may”, can it return “always” as the computed value?  Or would that be non-conformant?
a.      This language is used again on P259, just above 16.3 Decoration.
21.     p235, 14.2 - “either a color or an image”  When used elsewhere, “either A or B” is used to express alternatives.  Here you can specify both a color and an image.  Recommend: remove the word “either”; also consider changing to “and/or” or add a trailing “, or both” to the sentence.
22.     p238, ‘background-position’, Value - Can’t we simplify the grammar here?

-----Current
[ [ <percentage> | <length> | left | center | right ] [ <percentage> | <length> | top | center | bottom ]? ]
|
[ [ left | center | right ] || [ top | center | bottom ] ]
|
inherit

-----Simplified (Removes an || and makes it clear that the reason for the 2nd rule is to add the cases where a horizontal keyword can come second)
[ [ <percentage> | <length> | left | center | right ] [ <percentage> | <length> | top | center | bottom ]? ]
|
[ [ top | center | bottom ] [ left | right ]? ]
|
inherit

23.     p238, ‘background-position’ - The term “initial position” here is troubling.  The diagram on page 237 refers to the corresponding location as “center image”.  I think it would be best to come up with an official term for “the base image position” (which is also not a great name, but hopefully gets the point across for this discussion).  Using “initial position” is easily confused with “initial value of background-position” which is something totally different (“0% 0%”).
24.     p239, center - The block here is addressing both the meaning of the value and the interaction with other fields.  I think it’s attempting to say too much, and complicating things.  Recommend: “Equivalent to ‘50%’ for whichever axis it is being applied to.”
25.     p241, 15.2, first paragraph - Isn’t “ensure” a better fit here than “insure.”
26.     p260, Note - Does this same behavior apply to overline, line-through, and blink?
27.     p262-p263 - If an author wants text to have zero spacing between letters (or words) how is this achieved?  Put another way, is the “default space” between characters/words exposed in some other way?
28.     Unresolved references:
a.      p70, first paragraph - gamma correction
b.      p109, tty - pixel units
c.      p151, Note - applied to all elements in CSS1
d.      p164, 2nd full paragraph - opacity
e.      p166, 1st paragraph - bidi class B
f.      p298, below example - SVG cursor