How the LACITO markup scheme fares with Native American dictionaries
Mike Hammond

We have a project running to develop on-line dictionaries of Native American
languages of the southwest. The idea is to make these dictionaries available
to learners, speakers, and researchers. Our system is based on applets and
servlets, and we are developing an XML-based markup scheme for our
dictionaries. I tried out the LACITO markup scheme using some of our Navajo
materials. I think it has several nice properties.

GOOD THINGS ABOUT LACITO:

1. It's based on XML.

2. It's real simple.

3. I could imagine embedding some version of it WITHIN more specific XML
applications, like our dictionary scheme.

NOT SO GOOD THINGS ABOUT IT:

1. It does not provide the additional structure we need for dictionaries.
(Presumably, analogous concerns would arise for any document with more
structure than a "text".)

2. It would seem to commit one to a rather simple model of sentence- and word-
structure.

3. I may have misunderstood, but they seem to have invented a new attribute
"lang", where XML already provides one: "xml:lang". Moreover, the XML proposal
includes specific proposals for values for this attribute for many of the
well-known languages.

4. I notice that while the linguistic elements "s", "w", and "m" allow for a
"transl" element, nonlinguistic elements like "title" do not.

5. We have found it quite important to have links and this system doesn't seem
to have these.

6. At the moment, our own system does not include sound files. Most of that
part of the dtd is not useful for us (at the moment) and a bit problematic if
we want to use this for simply structuring a document.