From lou.burnard@computing-services.oxford.ac.uk Tue Sep 11 11:02:58 2001
Date: Mon, 10 Sep 2001 23:43:23 +0100 (GMT Daylight Time)
From: Lou Burnard <lou.burnard@computing-services.oxford.ac.uk>
To: Stewart C. Russell <stewart@ref.collins.co.uk>
Cc: editors@tei-c.org, Michael Beddow <mb@mbeddow.net>
Subject: Re: Summer reading... and a plea for help


Dear Stewart

Thanks for your comments on the dictionary chapter.
>
> As far as I can tell, I saw nothing that'd prevent the Print
> Dictionaries section being used with XML.

Well, there are a few problems with some content models in fact, but
we  already know about those.

>
> I didn't see any typos, but I hope the formatting of long lines of
> verbatim text (see 12.1, p254, definition of entity
> 'm.comp.dictionaries') is improved, as this example goes off the margin.

Yes, this is a persistent problem with the formatting of this version.

>
> I do have a couple of observations, which may require a bit of a rethink
> or some extra work:
>
> 12.3.1: <pron>: The notation used for phonetics in the examples is never
> defined. In order to reduce wheel-reinvention, some references to
> systems already in use (such as SAMPA) might be helpful.

Good point. Will do so.

>
> (The character entity names given in 'teiipa.ent' are hardly mnemonic,
> either -- &IPA322; for a schwa? I wouldn't -- or have any of the Collins
> editors -- use this system.)

Why not? They don't have to see the actual entity name. However, your
question reminds me that we need to check whether IPA symbols are defined
in Unicode or not.

>
> 12.4: <oRef>: acceptable values for the 'type' attribute are never
> specified, yet various (seemingly standard) values are used in the
> examples that follow.

This happens passim in P3 : it would indeed be a good idea to be more
specific about the need for consistency here.

>
> Maybe I'm too concerned with the strict definition of attribute values
> for computer-readable data interchange. I'm just starting a project to
> produce a grand-unified DTD for all our dictionaries and thesauruses
> (TEI.dictionaries is too unweildy, I have to say) so I could be being
> over sensitive.
>
> And a very minor comment on a statement made in 12.5:
>
> 	"In practice, publishers begin with the lexical view..."
>
> I'd disagree. Most dictionary publishers started with a lexical view
> derived from the typographic view, as typesetters' tapes were often the
> first electronic form of the data they had available. As the expense of
> rekeying a whole text in a structured form is high, tagged versions were
> often created from the typographic data.
>
> Tagging typesetter's tapes still makes up an uncomfortably large part of
> my job, alas. And yes, sometimes they are still 9-track open-reel
> tapes...
>

Maybe Collins is unusual! But rather than get into that fight again, I'll
tone down the prose a bit "Some publishers begin with the lexical view,
others with the typographic view..."

Thanks again!

Lou

