W3C home > Mailing lists > Public > www-style@w3.org > December 2010

Re: [CSS21] 4.3.2 Lengths (reference pixel?)

From: Dr. Olaf Hoffmann <Dr.O.Hoffmann@gmx.de>
Date: Fri, 10 Dec 2010 10:58:28 +0100
To: www-style@w3.org
Message-Id: <201012101058.28728.Dr.O.Hoffmann@gmx.de>
David Singer:
> I don't understand your question, and I don't think you understand what is
> defined here.

Well, due to my job (experimental physics at university) and education I have 
experience with units and due to previous work with drafts and 
recommendations of the W3C groups I have some experience with this 
as well.

And what is defined is explained in the related chapter:
'Lengths refer to distance measurements.'

Measurements and distances is something, I have a lot
of experience with, therefore no problem to understand
(this might be different for a few other people, that 
contributed to discussions here or in the public-fx mailing list ;o)

>
> In the cases where there is no 'display surface' to measure on, 

I think, lengths are only relevant for some kind of graphical presentation,
not for example for aural style sheets, removed anyway in CSS2.1.

> or when it 
> is viewed at distances that mean that some other size is appropriate (e.g.
> when a normal web page is shown on an electronic billboard), this set of
> relationships is used.  

It is another question, whether it is useful or not to define a fixed relation 
between px and meter. If you think, this causes problems, you should 
sent a comment about this issue to the list (I agree, that for specific
devices the current relation between px and meter will cause accessibility
problems, resulting in a need of the audience to switch off the styling
at all).

> Did you read the section where it talks about 
> choosing an anchor, that measurements are either anchored in physical
> lengths on the page, or by the reference pixel?
>

The whole point about this is, that the relation between px and meter
is already defined before with the cited sequence. If something is 
(127/480)mm, it is precisely defined and there is no need anymore
to choose something, it is fixed, because mm is always a precisely 
known and reproducible size - whether it is a useful unit for some
devices or in some situations with limited information on how to
present a mm for the user agent are a completely other questions.
This should not be mixed up.

> The only thing I'd maybe complain about here is that we appear to define
> that an inch is exactly 2.54 cm, 

This is no practical problem, because the official definition how to derive
an inch from a meter is the same, therefore the inch is derived from meter,
what is derived from a second, therefore it is only a simple scaling factor
like the velocity of light.

> whereas I think that's not a relationship 
> for us to define, merely inform about.  

Yes, this is true, only information, no definition, this is the same
as for mm and cm - it is well defined as well in other resources.



Boris Zbarsky:
>I think you just misunderstood what the spec says.

I don't think so. There is a simple definition, that finally
results in 1px = (127/480)mm
Because we know how much 1mm is with good accuracy, 
with this definition we know, how much 1px is.

Therefore the following section is not useful anymore.
It was in previous versions as CSS2.0, because there it
was required to know, how large 1px is. Because with
the new definition, we already know how much it is. The 
old section about reference pixel either contradicts now
with the new definiton or it fits only accidentally.
That someone can choose something is misleading,
because the definition already choosed what applies.

Therefore I propose to remove this misleading sections
to avoid confusion.
Alternatively, if as mentioned above, there is a fear, that
the new definition causes problems, this should be removed
to avoid the contradiction.

>The CSS spec doesn't say that its "mm" or "cm" units are the same as 
>actual physical mm and cm, if you will note.

It defines:
"
in: inches  1in is equal to 2.54cm
cm: centimeters
mm: millimeters
"
This is quite simple to understand and to test with good accuracy in our
time. 

If the CSS-WG thinks, that they do not need centimeters and millimeters
or that they need unit identifiers that express something different, they 
should use other names for it to avoid confusion.
But it is maybe a little bit late to skip absolute units like cm and mm
in CSS2.1. They still exist in CSS2.0 and as far as I understand this,
CSS unfortunately has no version indication. Any inconsistent change
of a feature means for authors simply not to use it anymore, because
it is defined in different ways in different versions of CSS.

>Yes, this is somewhat odd.  But that's the way CSS is used in the real 
>world, it turns out. 

If the current draft is odd, I think it is time to change, to aviod, that
finally the complete CSS gets odd ;o)

>I suggest reading the extensive discussion on this  
>issue that's already happened.

My observation about this at least with the discussion within the
public-fx list is, that several people seem to have problems to understand
what standards are for and what a unit like cm or mm really means ;o)
Well, it is not a trivial thing indeed, it is physics, logic and international 
standards, not just arbitrary styling, but if there is a need for it in CSS,
one has to care about it somehow in a qualified way.
The current draft and the comments indicate, that there seems
to be some need for help and I try to help to write something 
consistent and meaningful, avoiding confusing/misleading oddness.

Olaf
Received on Friday, 10 December 2010 09:59:03 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 17:20:35 GMT