RE: [css3-page] examples in 3.3.2 (page size) are 'US-centric'(?)

> [Original Message]
> From: Michael Day <>
> My preference for "a4" over "iso-a4" is exactly as I stated: the A4 paper
> size is known and widely used. If you specify ISO A4, many people will be
> confused and think you mean some variant non-standard size, as people
> don't know (and don't *need* to know) that A4 is an ISO standard.
> I know that the ISO B sizes and the JIS B sizes conflict. However, the A
> sizes don't, and they are in more common use. It makes sense to support
> shorthand keywords for the most commonly used ISO sizes.
> As to jumping the gun... you will find that we implemented page size
> keywords before they were added to the Paged Media draft, and then
> proposed them for inclusion back in April 2003 [1]. Discounting the view
> of every CSS implementor may give you satisfaction, but I doubt that it
> will lead to an improved CSS standard.
> [1]

I said discount, not ignore.  You have a vested interest in seeing that
the CR follows what you have already implemented which raises the
possibility of a conflict of interest.  That is one reason why it is
suggested that private extensions to CSS such as those that Yes
Logic uses have a vendor specific prefix.  That suggestion has
been pointed out to you before IIRC,

If the module goes beyond just A4 and letter, I am of the firm opinion
that the B sizes should be included as well, and that means that
a prefix is needed to indicate whether iso-b4 or jis-b4 is intended.

Quite frankly, I fail to see what is so difficult about iso-a4.  It is four
extra characters to type.  Anyone who will be using this property
will need to refer to some form of documentation. As for the
possibility of doing something like this for the metric sizes:


which would allow one to use a shorter form in those contexts
where the classname prefix is not needed; I think that would be
more confusing than to always use the classname prefix.

In short, I stand by what I said in my previous post.  If there
is a reasonable possibility that keywords for either or both
of the B series of paper sizes would be present in either
this version or a future version of the module, then the classname
should be used.  Only if they are not used and it is made explicit
that they will not be added should using non prefixed keywords
even be considered.  The problem is that I don't believe that
if we introduce keywords for paper size that the list will not be
extended to cover at least B4 and B5.

In an oblique way, you have made my point for me, Michael,
Suppose that, as in the current draft, only the keywords "a4"
and "letter" are supported.  What is there to stop a German
company from developing a user agent that adds "b4" as a keyword
as a shorthand for ISO/DIN B4 and a Japanese company from
developing a user agent that adds "b4" as a keyword as a shorthand
for JIS B4?  As Yes Logic has shown, the request of W3C that
proprietary CSS extensions use a vendor specific prefix will be
ignored by some.  In that light it would be far better for the standard
to say that all current and future keywords relating to paper size
will follow the <classname> "-" <sizename> form from the IEEE
PWG standard, even if the chosen keywords could be expressed
uniquely by just the <sizename>.  That way when those extensions
are made, and they are likely to be made, at least they will be likelier
to be interoperable and consistent,  The alternative would be to not
use keywords at all.  That  would cause much more problems for
the ISO sizes than for the North American sizes, but it would be
better than having a standard that would encourage conflicting
extensions to the standard.

I still think that keywords are a good idea, but given the likelihood
of extension to cover the B-series (either officially or unofficially)
I would much prefer the use of "iso-A4" to just "A4".

Note: I have noticed one more IEEE PWG sizename that requires
the use of the class name to distinguish the sizes.  Unlike the two c5's
or the two f's I noted in an earlier post, this might involve a common
size.  It seems that the PRC and the ROC have two different versions
of a size named 16K.  Furthermore, it looks like that size sees some
reasonably common use as I have come across a printer spec that
mentions the 8K and 16K paper sizes by name (along with ISO A3-A5
and JIS B4-B5, but not ISO B4 or B5 strangely enough) in its list of
supported sizes.

Received on Friday, 23 January 2004 19:38:50 UTC