FACET - What it Means
Common senses from which the technical ones below (#1-#4) try to borrow
"Facet" is not widely used as a technical term. (DMOZ ignores it). But these may be interesting.
[0a] - http://dictionary.reference.com/search?q=facet
Various reference sources
[0b] - http://thesaurus.reference.com/search?q=facet
Roget's, in a very old edition
[0c] - http://www.wordsmyth.net/live/home.php?script=search&matchent=facet&matchtype=exact
Bob Parks' combined reference
Sense 1 - Scripts associated with a data type which can manage its constraints and defaults
This is the oldest technical sense - came in late '70's as a way to model Minsky's Frames. Facets were originally processes (in Lisp, Smalltalk, or Simula) which (on request) checked constraints or computed default values.
[1a] -http://www.cs.umbc.edu/771/current/presentations/frames.pdf
FRL, KRL, KL-ONE, LOOPS, KEE; example facets shown in KIF. Historical references cited
[1b] - http://www.ndsu.nodak.edu/instruct/juell/cs724s03/foils/frl.html
Overview containing links to original FRL def by Goldstein & Roberts
[2a] - http://www.lexikos.com/words/scripted.jsp
Similar version
now being implemented for topic maps. Uses WORDS scripts as
processes
[2b] - http://www.lexikos.com/psi/words/aspect/
PSIs for
occurrence types of the (slot-like) aspects holding such (facet-like)
processes.
Sense 2 - Features that characterize a numeric data type, including value constraints
This replaces the procedures of sense 1 with a list of primitive data features
that constrain values sufficiently to let application code check
them. Default values seem forgotten in XSD, which barely touches
on sense 3 (see 5d).
[3a] - http://www.w3.org/TR/xmlschema-2/#facets
XSD definitions
of related typing. Data only. The programs are kept separate
[3b] - http://edocs.bea.com/wli/docs70/classdocs/com/bea/web/validation/FloatWord.html
Example of such a program from the BEA application server world
[4a] - http://www.altheim.com/ceryle/api/org/ceryle/tm/Facet.html
Larger set for
Topic Maps, based on TM4J and PSIs. Includes code for fixed
defaults
[4b] - http://www.infoloom.com/pipermail/topicmapmail/2003q4/005440.html
A string for such a value OR a constraint on one. Next message gives examples
Sense 3 - Features that characterize enumerated data types, including classification trees
In any formal language, enumerated types
differ from scalar types and need different treatment. Options
with no associated order define sets. Imposing a hierarchy
on them creates an FC tree, which can degenerate into a comparative
series.
[5a] - http://www.kmconnection.com/DOC100100.htm
Top homepage
from Google for "faceted classification" - gives library science links
[5b] - http://www.poorbuthappy.com/fcd/
An active mailing list for users of the FC type of data
[5c] - http://xfml.org/
An exchange
format for FCs based on TMs, seeking compatibility with RDF, and RSS
[5d] - http://www.w3.org/TR/xmlschema-2/#rf-enumeration
XSD explicitly
excludes "entity" order - so it just does sets - no trees or series
[6a] - http://www.topicmapcentral.com/wiki/Wiki.jsp?page=FacetedClassificationPattern
Kal's published
design pattern for building FC hierarchy out of topics and associations
[6b] - http://www.lexikos.com/psi/psitree.jsp
Dan's pattern for building a static FC hierarchy from PSIs and anchor strings
[6c] - http://www.infoloom.com/pipermail/topicmapmail/2003q4/005432.html
Carlos' observation that FC might best be kept peripheral to Topic Maps
Sense 4 - The original meaning in HyTime, and its comparison to RDF
When I Googled "facet RDF", I got very little back of interest except
from Topic Map people, whom I suspect are the only ones seeing the
comparison. We have since officially dropped related specs, so
this sense is best labeled archaic, obsolete, and dead
[7a] - http://www.w3c.rl.ac.uk/SWAD/deliverables/8.1.html
Top spec-like
hit in Google from the RDF world. It seems similar to sense 3
[7b] - http://www.isotopicmaps.org/pipermail/sc34wg3/2003-April/001631.html
This was good for background; and it starts a thread examining this sense some
.