FWIW:

I won't get into it much for now, but I have used a "band format" for lexical data. There are other names for this type of free form database. I wrote a crude elisp routine to recover entries into LaTeX formatted files.

A "band" is a record, so to speak. I am not very well qualified in this, but was able to use it to record lexical data. You may find some linguists' websites where this or similar formats are elucidated. A record starts with a double dotted key, and information categories may be made up on the fly, as marked by single-dotted keys, preceded by at least two spaces. I think it's convenient for a record to be delineated by a line feed, as well.

..HW <headword> .D <local dialectZ> .GE <English gloss>   .NS <scientific name> .NCE <Common Name>   .NCs <Spanish Common Name> .R <remark> .RC <Remark on Cultural Signficance>

This is just a made up case, but perhaps you can catch the drift.

Here are a couple of simple cases from my files:

..hw tutubi   ,lang vis .nce dragonfly    .source FSD
..HW sigai    .lang vis .ge (mollusc) shell, when empty
..hw soksok .ec gecko .cg .la ilo .src hanna .n
..hw locus .ec octopus .cg .la ilongo .src hannah .n see nucus [vis]; kuus [chuukese]
..hw tikling .ec heron .cg .la vis .from fsd
..hw nucus .ec octopus .cg .la vis .src fsd, hannah .n related to chuukese kuus

Fairly straightforward elisp would scan a record and wrapping each item in a particular typeface.

To get an idea of the output. Each line was output as an \item in a list. This got to be a LITTLE cumbersome, perhaps, and someone good at coding would do it differently. The idea is that a lisp routine scans the records and spits out list items. This could be any kind of output, and perhaps org mode would be a good way to rig a routine to scan list items and output different band types as slanted (\sl), roman, or italicized components.

\item [{\sl k\'{u}\'{u}s\/}$_{3}$]   \index{k\'{u}\'{u}s} \quad     Small, night-time octopus.   HADJ E\'{e}t.

\item [{\sl k\'{u}\'{u}s\/}$_{4}$]   \index{k\'{u}\'{u}s} \quad     Daytime octopus.   {\sc syn\/}:\ {\sl nippach}.    {\sc alt\/}:\ {\sl k\'{u}\'{u}h}.    \HADJ F\'{o}n\'{o}.

\item [{\sl k\'{u}\'{u}sen neepwin\/}]   \index{k\'{u}\'{u}sen neepwin} \quad    {\sc see\/}:\ {\sl nippachin neepwin}.    Even though this is not said, it would be the correct way to say it. \HADJ Wonip.

This may not be an appealing approach. I am still pleased with the ability to flexibly add band keys on the fly, during data entry, and the potential to use LaTeX as a frontend. HTML would also be useful, depending on how you wish to read your dictionary.

Not a perfect system. Linguists have done better. Robert Hsu of the University of Hawaii built a system around SPITBOL and maybe SNOBOL4. I was hopelessly lost trying to use those, but elisp did what little I needed. I think that it may be possible to organize a database using org-mode.

For now, I have a capture template for data entry, such as it is:

("=" "lex" entry (file+headline "lexicon.org" "Unsorted") "* ..hw %^{Headword} .gs %^{Scientific Name} .ge %^{English Gloss} .ec %^{English Common Name} .cg %^{Category} .la %^{Language} .src %^{Informant} .n %^{Note} %? .dt %u " :prepend t :immediate-finish t)

Again, FWIW. To me, a great deal. Maybe to others, not so great of a deal.

Alan

On Mon, Jun 6, 2011 at 7:38 PM, Christian Moe <mail@christianmoe.com> wrote:

Hi,

Is anybody using Org-mode to build an advanced dictionary with sub-entries, tags etc.? Would you be willing to share a setup?

For example, the obvious way to build a dictionary would be to use a dictionary list (I borrow a few English-French lines from the wonderful WordReference.com site):

- pine ::
(/paɪn/)
1. /m noun/ [bot.] pin; *stripped ~* pin décapé.
2. /intr verb/ languir (*for* après; *to do* de faire)

This looks nice, but unfortunately, you cannot set tags or properties on dictionary terms, so it's not particularly amenable to fancy searching, mapping etc.

On the other hand, you could do something like this:

* pine
:PROPERTIES:
:Pronunciation: /paɪn/
:END:
** pin :bot:
:PROPERTIES:
:Word_class: noun
:Gender: m
:END:
*stripped ~* pin décapé.
** languir
:PROPERTIES:
:Word_class: verb
:Transitivity: intr
:END:
(*for* après; *to do* de faire)

It's a pain to do, and because of outline folding, it could be a pain to look up meanings, and you might need to do some serious post-processing on the export to make it look anything like a dictionary. But when you're done, you could extract a list of all botanical terms (:bot:), or of words and pronunciations only... etc.

So for my growing pile of translation notes, I might like to keep that kind of thing. But there are so many ways it could be organized - what do you put in subheadings? what in entry text below subheadings? what in tags, what in properties? etc. So if someone has an example that works for them, I'd like to see one.

(Org may not be the best tool for this job, of course, but it's the right tool for me...)

Yours,
Christian