Definition of REQUIRED

The misunderstanding of the Definition of REQUIRED comes up a lot on
this list so I thought I would put it out as a separate memo rather
than buried with another topic.

When we dropped to just two ratings - we looked at lots of options and
came up with REQUIRED  and RECOMMENDED.
Perhaps they are not well chosen but here is how we came on them

1) The guidelines are just that, Guidelines, not requirements.  There
is no certification of sites as accessible etc. Therefore we cannot
have REQUIRED guidelines.  W3C does not have the authority to require
anything as I understand it.  But we could say that doing this or that
was required for some users to be able to use the pages.  So that is
what we did.   So the definition of Required is

[Required]
Required, otherwise it will be impossible for one or more groups of
users to understand the page.


2) We then just used recommended for the second level - those things
that would make the page easier to use but are not required to make
the information accessible.

[Recommended]
Makes page easier to understand and use.


We have asked for comment on these definitions - but have not gotten
any that suggest other definitions or approaches.  So we have
continued with this.  (We dropped from 4 to 2 levels after
recommendations at the GL group meeting in Texas)

Sooooooo

Required does not mean that we require it.   It means that it is a
fact that it is required if all users are to be able to access the
information on the page and understand it.       All other
recommendations go into the recommended category, even if we think
they are very important.

As always, comments on other ways to approach this are welcome.   I
think we need to have a factual base though for putting things into
the required category since we don't really have a mandate for
arbitrarily requiring things of web authors and we don't have a
certification process in place.     RC group is looking at these
aspects.

Having said all this, I will say that this whole aspect is a bit
murky.  If you try to apply things absolutely strictly you quickly end
up with lots of things that could go either way based on
interpretation.

So let us know what you think.    Do these seem to work (mostly)?

Or is there a better approach.

Thanks Much

Gregg
For the Editors/Chairs

Received on Tuesday, 14 April 1998 23:36:14 UTC