Dictionaries

Rob Rothenburg <wlkngowl@unix.asb.com>

HTML markup for "special words" (acronyms and abreviations, proper names-- people and geographical locations, definitions, 'foreign' words) has been ambiguously addressed at best. In September '97 I posted a suggestion to the list [1] recommending a possible solution using the LINK element to reference a "dictionary file" that in HTML, that consisted of a definition list for defined terms.

I came across RFC 2229, a "Dictionary Server Protocol", which may provide an even better solution. The INS, DEL, Q and BLOCKQUOTE elements in HTML 4.0 have the CITE attribute for referring to a URL as a citation.

Something similar could be done with the ABBR, ACRONYM and DFN [3] elements where the citation refers to a dictionary server or a document that contains the definition. This would supercede the previous suggestions, and would be work well with older browsers and agents which do not recognize citation attributes.

For example:

  <ACRONYM TITLE="Hypertext Markup Language"
    CITE="file:///E%7C/Docs/html4/html40.txt">HTML</ACRONYM>

  <DFN CITE="dict://dict.org/m:shortcake">shortcake</DFN>

An optional reference to a dictionary or glossary (using the LINK element) would also be useful.

How a browser (user agent) handles these citations is not specified. An agent may display citations as hyperlinks, or provide a separate menu of definitions. (An agent might even be configured to look in a default dictionary server for definitions of uncited terms, or to override citations in a document should the links be invalid or the server unavailable.)

Multiple citations for a term in a document should be discouraged, but not forbidden since there are times when an acronym or term may have multiple meanings.

Pronunciation (aural style sheets) and other language-specific issues for abbreviations and acronyms are not addressed here.

The document type returned by dictionary servers is also not addressed here (nor is it addressed in [2] RFC 2229) although presumably in the future a "definition markup language" will emerge. Note that that RFC 2229 is not an official standard.

Notes and References:

[1] "A less painful way of adding dictionaries to HTML" http://lists.w3.org/Archives/Public/www-html/1997Sep/0394.html

[2] "RFC 2229 A Dictionary Server Protocol" ftp://ftp.isi.edu/in-notes/rfc2229.txt

[3] The DFN element represents the "defining instance of the enclosed term" (according to Sec. 9.2.1 of the HTML 4.0 specs) implying the first use of a term in the context of a definition.

[4] What is missing is an element for marking up "proper names" (names of people, geographic locations, institutions, or even scientific names such as genus/species) and other special terms ("keywords" that may not be appropriate for including in the META elements but useful for users of a collection of documents).

DFN could cover some of these, and certainly others could be used as standard links (the A element). Whether one would like to mark these attributes for user agents (such as search engines, or authoring tools which could be used to build an index) is a separate issue not covered in the message.

Author Contact Information:

Robert Rothenburg Walking-Owl
PO Box 1327
Stony Brook, New York, 11790
USA

E-mail: wlkngowl@unix.asb.com