Dictionaries (was: "... comments on ACRONYM")

Rob (wlkngowl@unix.asb.com)
Mon, 21 Jul 1997 19:33:12 -0500


Message-Id: <199707212342.TAA24538@unix.asb.com>
From: "Rob" <wlkngowl@unix.asb.com>
To: Dave Raggett <dsr@w3.org>
Date: Mon, 21 Jul 1997 19:33:12 -0500
Subject: Dictionaries (was: "... comments on ACRONYM")
CC: www-html@w3.org, Aymeric.PoulainMaubant@enst-bretagne.fr

Summary: an idea for dictionaries in HTML documents

On Mon, 21 Jul 1997 Dave Raggett <dsr@w3.org> wrote:

> On Thu, 10 Jul 1997, Aymeric Poulain Maubant wrote: 
> 
> > Numerous comments have already been done here on the ACRONYM
> > element. 
> 
> Compression will help when a given acronym occurs repeatedly
> in the same document.  Another approach would be to use
> a LINK to a dictionary of acronyms, as has been suggested
> for abbreviations.

Perhaps ACRONYM should be deprecated for a more general element [I
just deleted a few paragraphs I wrote about why ACRONYM is
problematic-- mainly with differentiating one similar abbreviation or 
acronym from another when the context is vague].

Something like DICT for "DICTionary defintion" (DFN is
already used for defining instance.)

The DICT element could be used in the header or be included using
<LINK  Rel=Dictionary ...>

Some examples,

 <HEAD>
  <DICT TYPE=Abbrev ID=Street NAME="St."
     CONTENT="Street"></DICT>
  <DICT TYPE=Abbrev ID=Saint NAME="St."
    CONTENT="Saint"></DICT>
  ...
 <BODY>
   <DICT REF=Street>Elmo St.</DICT>

The CONTENT (in this case "Street") would be substituted by a speech 
reader, or show up as a pop-up window.

If CLASS (ie, <DICT CLASS=Street... >) is used markup can be
customized using style sheets.

Note the use of REF is latter instances or a definition since an ID
can only be used once. If no ID or REF is specified, it would be
default. TYPE would also be optional but recommended.

  <DICT TYPE=Acronym NAME="HTML"
    CONTENT="Hypertext Markup Language"></DICT>
  ...
 <DICT>HTML</DICT>

A possible expansion is to allow a DICT defintion in the body of the
message at the first occurance (latter definitions without IDs would
be ignored).

Another possibiliy: If NAME is left out, it is assumed to be the 
element's content, so

  <DICT TYPE=Acronym NAME="HTML"
    CONTENT="Hypertext Markup Language"></DICT>

is the same as

  <DICT TYPE=Acronym CONTENT="Hypertext Markup
    Language">HTML</DICT>

Although the latter is inappropriate for HEAD elements in older 
browsers.

An HREF could be used to link to an optional resource.  How this 
would be handled: only the first occurrance in a document would have 
a link. UA's might also provide a menu of dictionary definitions.

The use of a dictionary scheme like this would not be limited to 
abbreviations or acronyms, but could also include people (with 
information about them, their relation to the document-- optional REL 
attribute?, their titles in an organization, etc.), institutions, 
geographical names, other proper names, keywords, foreign words or 
phrases (with definitions), definitions of technical terms, etc.

Robots/indexers could also recognize certain type, and meta-data 
schemes could use the dictionary when building them.

Examples,

  <DICT HREF="~/jdoe" TYPE=Person REL=Author>Joe Doe</DICT>

  <DICT TYPE=Definition LANG="la" NAME="ad infinitum"
    CONTENT="without limit">

> Dictionaries in principle can also be used to associating
> hypertext links with terms appearing in documents. This was
> the subject of a talk at the recent Web conference (WWW6)
> in San Jose.

In SGML, something like

 <!ENTITY % dict.types
  "abbrev|acronym|definition|
   none|other|
   person|organization|location"
 >
 <!--
  abbrev        abbreviation
  acronym       acronym
  definition    definition
  person        a person
  organization  an organization/institution
  location      a place (city, town, country, etc.)
  other         none of the above
  none          this is not a definition/do not substitute
 -->
 <!ATTLIST DICT
  %attrs
  href          %URL            #IMPLIED -- resource with more info --
  name          CDATA           #IMPLIED -- the word/phrase defined --
  content       CDATA           #IMPLIED -- the definition/expansion --
  type          (%dict.types)   other    -- the type of definition --
  rel           CDATA           #IMPLIED -- relation to document --
 >



Rob
 
---
Robert Rothenburg Walking-Owl (wlkngowl@unix.asb.com)
Se habla PGP.
http://www.asb.com/usr/wlkngowl