Of (proper) names, aliases, and languages

Hi...

I was just thinking about knobots and machine generated indexes for
search engines (as I often do for lack of something useful to
occupy me ;-)...

And the following struck me:

In texts, you can have:

wolfram		for	tungsten

constantinople	for	instanbul

consumption	for	cancer

king charles	for	charlemagne

mark twain	for	samuel langhorn clemens

etc.  Ie, there are words and names that in other languages, times,
contexts, dialects, etc. might have several "aliases".  We need to
be able to tag these associations (equivalences) so that "unworldly"
software (search engines, possibly AI, possibly not) can still
make intelligent selections.

Anyone else feel there is a need here?

Then there is the case of names being language dependent, like
budweiser (german) being natively budevice (czech)... but that's
a different can of worms.

-Philip

Received on Tuesday, 20 June 1995 18:05:47 UTC