Re: Advice on dictionary conversion

I wish to convert a bilingual dictionary from MS-Word format to
"properly"-tagged XML, and I hope I may ask for some comment on the
feasibility of this, using XSLT or otherwise.

I have some experience with converting dictionaries from a
"typographically" tagged format to a semantically tagged format, and
found that this, even with a well organized dictionary is a very
time-consuming job.

The right question to ask is thus how far do you need to go, which
depends on your target application. If this is just to make the
dictionary searchable for human users, your probably are already far
enough when every entry is encapsulated in <entry> and <headwords>
stand out, besides of course retaining the typography. If the
dictionary needs to be smarter, you will find a lot of cases where the
typographic rules have been used in a lax way, such that any program
fails and requiring human review to get right. Quite a daunting task
for a 5 megabyte blob of text.

Having said that, I've created a few XLST scripts to "up-tag" a
typographic dictionary (assisted and glued together with Perl).
Running this not only requires tweaking on the conversion scripts, but
also on the dictionary data, to convert necessary tagging errors and
omission. I found myself using a lot of nested <xsl:for-each-group>
statements to group tenses and meanings together based on typographic
hints.