Re: Dictionary link types

to follow up on what Dave Raggett said:
> 
> Can anyone remind me of the motivation for being able
> to recognize the link is for a dictionary independent
> of what kind of dictionary it is?
> 
> Unless there is a strong motivation for this ability
> I would have thought that a single rel value for
> each kind of dictionary would suffice, e.g.
> 
>    rel=abbrev-dict
>    rel=phonetic-dict
>    ...
> 

Let my try.  I argued against trying to encode the dictionary
types in a flat list and using that list to qualify links
because:

Dictionaries are not single-function and hence dictionary types
are not a partition.  A hierarchical taxonomy would work to
characterize them, or a logical record type such as I discussed
earlier.

Dictionaries are relations and have multiple potential functional
applications.

Secondly, there are different views of the function of a
dictionary as applied to a document:

There is the a priori view of all functions that a dictionary
could support without regard for the document it is to be applied
to.

There is a reduced range of functions based on documented
characteristics of a document.  If the document is in English and
the dictionary is Slovenian-to-English but that is the best
available dictionary that connects Slovenian and English, an
English-to-Slovenian function could be generated using searching
in the Slovenian-to-English dictionary.

There is a reduced range of functions based on explicit selection
guidance from the document author.  This would be things like a
directive to expand acronyms using a particular dictionary.

Dictionary selection wants to fit our pattern of author
intelligence contributing to reader decision.  There are plenty
of examples from disability access to demonstrate that if the
overall decision flow isn't like that, the disabled will be hurt
because it is not reasonable to expect the author community to
understand or anticipate everything.

To support this style of decision flow, it is important for the
dictionary publisher and the author citing the dictionary to
declare capabilities in the broadest possible terms, and not
prescribe use uniquely when there are in fact options that can be
exploited to the advantage of browse-medium-diversity downstream.

-- 
Al

Received on Tuesday, 23 September 1997 11:53:27 UTC