Re: Font bug in IE 5.5

'It is recommended that authors use this conformance profile to take advantage of forward compatibility. Authors should be able to use style properties with an understanding that the cascading rules are processed correctly and that unknown properties and values are ignored. For example:

'  body {
    background-position: center center;
    background-position: 45% 55%;
  }
  
A MP-UA that can process percentage values for the background-position property will process the first background-position specification and then "write over" that value with the second background-position specification. A MP-UA that cannot process percentage values will process the first background-position specification and ignore the second background-position specification.'

This needs more clarification. I don't like 'process' at all - it's 
very vague. Secondly, this [whether unsupported values predicated of 
properties should result in the declaration being ignored or not] is an 
issue that isn't dealt with in CSS2 proper [and whereas I concurred 
with the aparent meaning of this paragraph, Ian Hickson took a contrary 
view], where I think it deserves addressing. I believe 'process' 
actually means 'support', as in 'bothers to support', rather than 'is 
capable of processing' as is perhaps implied, in that *any* UA that 
can 'process' center center is necessarily capable of 'processing' 45% 
45%.

In addition, consider this: under these rules, background: fixed url(back.png) blue;, a perfectly good declaration in the past would be *required* to be ignored by compliant mobile CSS UAs, thereby 
potentially making pages illegible.

Secondly 'specification' should be 'declaration' in CSS parlance. Also, I think ignore should be 'ignore', with that defined (or at least 
linked to the appropriate part of CSS2) since the WD doesn't precisely 
state its relationship to that specification (e.g., at the end it states that it subsumes certain parts of CSS (such as 'inheritence' [sic], but not others). Therefore, a section 'Relationship to CSS2' 
should be added:

'This document extends CSS2, subsuming all of its sections and 
requirements, with the following exceptions:
....
Where it differs from CSS2, this document is authoritative.'

[This would for instance, ensure that the requirement that UAs may
treat all border types as solid be restated.]
--

In the table of selectors, 
div.warning   The same as div[class!="warning"]   Class selectors   Yes 

obviously should be class~=. In addition, I think a note as to whether 
multiple classes are a compliance requirement should be included.

--

inline | block | list-item | none as the required property values for display support is wrong - inline is missing

--

<length>, auto, and inherit are the required values for width and height
support; this is entirely inconsistent with margin, defined thus:

'margin' Yes <margin-width>{1,4} | inherit 

where <margin-width> includes %; therefore the UA is required to support
% for one but not the other.

--

'visibility' shouldn't be required, because it's almost completely 
useless in the context that it is defined - visibility is only really 
useful when positioning properties are supported. In addition, 
'collapse' is definitely completely useless in the context of the set
of properties defined by this WD.

--

I would say that float (and, implicitly, clear) shouldn't be a requirement - it's complicated to support, and pretty useless without 
width % support. In addition, I would support many other properties 
before this one. Actually I don't think the document makes too much 
sense; presumably a simple CSS-enabled phone would just (physically) 
support a few properties anyway, whereas the sort of device (e.g., a 
$600 palmtop) that would support things like float would be able to 
support the full CSS2 specification anyway. It seems to me, from the 
contents of this document, to be massive overkill to devote a whole 
specification to one (admittedly rather fashionable) medium; after all, 
things like non-support of 'azimuth' are already covered (and better - 
a mobile device might be a mobile speech reader)by CSS2, but this does 
not [although the document does not define its application, which is 
part of the problem - does this apply to all devices that match media: 
handheld?; or is it only for explicitly MP-UA-compliant devices?]) 
define itself.

--

The font property is broken (or at least the definition doesn't make sense) by the final paragraph of section 2 (which states that 
declarations that can't be processed should be ignored) - line-height is
not required to be supported, but font is, and line-height is one of the
parts of the font property.

Received on Sunday, 15 October 2000 12:24:43 UTC