W3C home > Mailing lists > Public > w3c-wai-ua@w3.org > January to March 2001

Conditional versus Optional: Preliminary Observations

From: <oedipus@hicom.net>
Date: Thu, 22 Feb 2001 21:58 +0000
Message-Id: <200102230258.VAA18289@ns1.hicom.net>
To: w3c-wai-ua@w3.org
Conditional versus Optional: Preliminary Observations

in the minutes from the 22 February 2001 telecon [reference 1], the 
following exchange was recorded:

  GR: "Required optional content" is a little weird.

  IJ: Good point! 

  Action IJ: Find clearer wording.

  GR: I propose changing "optional content" to "conditional content".
      I think that conditional doesn't presume that one form of
      content is preferred over another.

  IJ: I don't think "optional" suggests that optional content
      is lower class.

1. required bits are not "optional"--"required optional" is an 
oxymoron; what is "optional" is the discretionary portion of 
the requirement--for example, in the HTML4/XHTML world, deciding 
on appropriate ALT text for the IMG element...  the A, the L, 
the T, the equals sign, and a pair of quotes are required--what 
goes between the quotes is the optional bit...

2. "conditional" because what is delivered to the requesting UA 
is the derivative of the conditions surrounding slash containing 
slash initiating the transaction; moreover, the conditions under 
which content is delivered (or in which content is capable of 
being delivered) are not always/necessarily "optional", as they 
may (or are quite likely to) include both those over which the 
user has either no or limited control, or of which the user is 
ignorant (in a non-pejorative sense)--conditions can also be 
predicated upon explicit user choice; server side filters and 
transformations, including processing by proxy servers; 
configurations slash settings; hardware limitations; language 
preference (accept) settings; functional limitations, 
environmental limitations; markup support slash standards 
compliance ; etc.; the point is that the "content" (the message) 
is capable of being delivered by a number of potential messengers 
(content/file types), depending upon which is most appropriate--
e.g. when certain conditions (no matter their source) apply, send 
slash receive slash expose slash render X, not Y or Z, but if X 
does not exist slash has not been provided, Q will be acceptable...

3. "conditional" is completely neutral--no need to speak of 
equivalencies; doesn't champion slash pit one form of content 
slash modality over another, as it doesn't matter why the 
conditions exist, only that the UA respond to them 
appropriately, by providing content in the most appropriate 
content-type slash form slash modality, whether due to an 
explicit request for a particular content type, the explicit 
exclusion of unsupported slash unusable slash unwanted content 
types, or by preference slash cascade order

4. the term "conditional" captures the nuances of the term far 
more concretely, and far less ambiguously, than "optional", as 
it incorporates user configuration; negotiation transactions 
(such as those based on CC/PP, accept headers, etc.); SWITCH- 
and SWITCH-like mechanisms; the rendering order of nested 
OBJECT elements; SMIL test attributes; and the CSS cascade, to 
name but a few

5. "optional" is a dangerous term because the plain English 
language definition of the word "optional" is, according to 
the online edition of Webster's (http://www.m-w.com)

  involving an option : not compulsory

which (at least to my ears) eliminates the term from contention, 
as use of the ALT attribute for the IMG element is compulsory in 


PS: i know that the example is technology-specific, but that's 
simply because the case of IMG is the most familiar and clearest 
illustration of the point...


[1] (long URI warning!) 
Email sent using AnyEmail from http://www.hicom.net
Received on Thursday, 22 February 2001 21:58:43 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 14:49:29 UTC