Re: [edict-jmdict] Spelling regularization

[Glenn Maynard <glenn@********>]

On Mon, Aug 9, 2010 at 3:32 AM, Paul Blay <blay.paul@googlemail.com> wrote:
> Why not have an [altsp] tag for glosses?

That seems overly complex.  Data embedded in XML shouldn't require
significant parsing to access, beyond XML itself--that's the whole
point of using XML.

I originally recommended a gloss attribute, eg. <gloss lang="en_UK">,
but on further examination I don't like that; it'd lead to too much
duplication of text.  Everything like this:

<gloss>favorite book (favourite)</gloss>

would become:

<gloss lang="en_US">favorite book</gloss>
<gloss lang="en_UK">favourite book</gloss>

... and then there's also the question of tying the "identical"
glosses together, so you can figure out which items can be discarded
automatically, and maintaining the extra glosses through edits.  It's
way overcomplex and not worth it.

A practical solution is much simpler: just pick which spelling
variants to use, use them consistently, and drop all of the others.

Anyhow, Jim doesn't want to change this.  (Jim: I really am trying
hard to understand your position on this, but I just don't get it...)

-- 
Glenn Maynard