W3C home > Mailing lists > Public > w3c-wai-gl@w3.org > April to June 2006

Some words about the terms "relative" and "absolute".

From: Marco Bertoni <mbertoni@webaccessibile.org>
Date: Thu, 1 Jun 2006 14:45:33 +0200
To: <w3c-wai-gl@w3.org>
Message-ID: <200661144533.666033@ocram>

Hi,
last week I was filling the surveys when I encountered the terms "relative" and "relative positioning" (in the Team A survey if I remember well).
I think we need some clarifications about those terms, expecially when they are used regarding both font sizing units and CSS positioning rules.

First off all in the "Comparison of WCAG 1.0 checkpoints to WCAG 2.0" [1] appendix we can read:

In General (Priority 2):
3.4: Use relative rather than absolute units in markup language attribute values and style sheet property values.

...that is related to:

WCAG 2.0 Success Criteria:
This maps to an advisory technique (Using readable fonts) for Guideline 1.4.

But Guideline 1.4 refer to visual and audio contrast [2] and I've not yet found, even in the "Techniques for WCAG 2.0", the advisory technique mentioned (Using readable fonts). Perhaps someone can help me to find what I'm searching for, I'm new to the group ;).

However even in the "CSS Techniques for WCAG 2.0" draft [2], in wich is correctly introduced the idea of "Using em or percent for properties that need to change" instead of "Use relative rather than absolute units...", when I look at the corresponding guideline I see mentioned "Guideline 1.3" that refer to the separation of structure and presentation.

Having said that, I must point up that using terms like "relative units" or "relative positioning" maybe somehow confusing:

1) I assume that the general accessibility principle in question is that users (expecially partially sighted people) must be able to enlarge fonts using their visual UA tools.

2) Microsoft Internet Explorer 6 for Windows - and all the previous versions - is not able to resize text if it is sized in pixels. But pixel are relative lenght units (they are relative to the viewing device resolution) [4].

3) On the other hand there are CSS absolute-size values that IE/WIN can perfectly resize: the absolute-size keywords [5].

So, IMHO it is really wrong and confusing to advise CSS designers that they should use relative length units to size fonts. Correctly, as I told before, CSS Techniques says: "Use "em" or % for properties that need to change". But we can make another stride to avoid to mention a unit identifier: e.g. here in Italy, in our recent accessibility law [6], we have introduced a requirement (Requirement 12) that says: "The presentation and textual content of a page must be able to be adapted to the dimensions of the browser window used by the user, without superimposing objects present or loss of information such as to render the content incomprehensible, including in the case of resizing, enlargement or reduction of the display area or characters in relation to the predefined values of these parameters.".  In other words with this approach we say that the designer must size charachters with a unit (not mentioned) that guarantee the enlargement of text in every UA. This is expecially useful to ensure forward and backward compatibility!. We cannot forget that, at the time of this writing, IE 6, with all off his bugs, cover almost all the market.

Directly following that we must clarify the use of the phrase "relative positioning": when a CSS designer hear that words he think about a CSS rule like this: #box { position: relative; } [7]. But in itself a positioning scheme isn't related to accessibility. Indeed (and correct me if I'm wrong) when we talk about relative positioning we are thinking about a concept like "use liquid design" rather than a specific CSS positioning scheme. So, IMHO again, it is wrong to advise designers to use a specific positioning scheme. Maybe useful instead to tell them that, for example, a liquid design is more accessible than a fixed design, no matter which technique they use to obtain that. Obviously we must first justify this suggestion.

These terminological problems are really serious: I've hardly discussed, here in Italy, with tricky designers that keep on using their beloved pixels for text sizing because these are relative units as checkpoint 3.4 of WCAG 1.0 suggest (no matter if partially sighted people using IE6/WIN cannot resize them). 

What do you think about?

Marco Bertoni

[1] http://www.w3.org/TR/WCAG20/appendixD.html
[2] http://www.w3.org/TR/WCAG20/guidelines.html#visual-audio-contrast
[3] http://www.w3.org/TR/WCAG20-CSS-TECHS/#syntax-data-types
[4] http://www.w3.org/TR/CSS21/syndata.html#length-units
[5] http://www.w3.org/TR/CSS21/fonts.html#font-size-props
[6] http://www.pubbliaccesso.it/normative/DM080705-A-en.htm
[7] http://www.w3.org/TR/CSS21/visuren.html#relative-positioning
Received on Thursday, 1 June 2006 12:43:31 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:47:46 GMT