[Andrew Dunbar ([Wiktionary-l] Re: English orthographies) writes:]
> On 9/22/05, Jim Breen
<Jim.Breen(a)infotech.monash.edu.au> wrote:
> > Yes, in fact it is the frowning on redirects that led me to looking at
> > the UW proposals. I was looking at the Wiktionary structure to see if it
> > would be a suitable environment for my Japanese-Multilingual dictionary
> > database. I ran into a number of problems, one of which was the "no
> > redirects" policy, and someone suggested I look at UW.
>
> The "frowning on redirects" policy is largely due to the fact that we have
many
> languages in one "namespace". When a particular English spelling variant
or
> even a plural happens to coincide with the spelling of another word in another
> language then we have to have two pages anyway. This is not uncommon.
> We then decided it was better to try for some consistency rather than having
> some shared pages and some redirects. The other major issue was what to
> do when a dictionary is created for both the British (colour, centre) market
> and the American (color, center) market without us trying to force upon anyone
> which is the "standard" and which is the "variant", which
redirects lead to.
This second problem goes away if a search for an entry can be made on more
than one "headword". In fact single headwords is a limitation of paper
dictionaries that never needed to be propagated into electronic
dictionaries.
> > Provided:
> > (a) the essential information (senses, POS, etymology, etc.) only has to
> > be entered once, and remains the same for all the spelling and
> > orthographical variants;
>
> Sometimes some of these will be different. In British and the Commonwealth
> except Canada "tire" only means "become tired". In US and
Canadian English
> it also means "tyre", the rubber ring on the outside of a wheel. But these
are
> homonyms rather than senses though many non-lexography savvy people
> don't realise the difference.
Of course they are homonyms. With a relatively small set of phonemes,
Japanese is riddled with homonyms; there are cases of more than 20 different
words with the same pronunciation. You'd go (and be) crazy if you tried
to treat them as the one "word".
> > (b) the user, on entering either form, gets
the one collection of
> > information which shows all the alternative forms of the word, then
> > I really have no objection. I can't understand why they are in different
> > database records, and in the case of my own JMdict (XML) they aren't,
> > but then I don't use SynTrans, etc.
>
> Basically it's an arbitrary database design issue. UW is going for more
> granularity. In this way it's probably more object-oriented since it breaks
> things down into more, smaller objects. There is nothing intrinsic right or
> wrong about either approach.
Provided the design doesn't intrude into the operation (creation,
maintenance, lookup, etc.)
> > Not really. I don't know about the
languages I don't speak (i.e.
> > everything apart from English, Japanese, French and a little Latin), but
> > in general the spelling has little or nothing to do with the etymology.
>
> Sometimes one spelling is definitely known to be derived from another
> and both remain in use in various places. For instance the Spanish word
> for "peanut" was borrowed from Nahuatl in Mexico as "cacahuate"
but
> when it was later borrwed into Spain itself it became "cacahuete". It
would
> be a shame to not have a way to record such things in the cases we do
> know them.
I was really referring to the centre/center, colour/color situations. I
should have said "minor spelling differences".
Cheers
Jim
--
Jim Breen
http://www.csse.monash.edu.au/~jwb/
Clayton School of Information Technology, Tel: +61 3 9905 9554
Monash University, VIC 3800, Australia Fax: +61 3 9905 5146
(Monash Provider No. 00008C) ジム・ブリーン@モナシュ大学