HPSG Diagrams

[Mike Maxwell]

Luis Casillas wrote:
> ...The best thing to have would be some sort of system that allows
> you to write documents in SGML/XML, to be interpreted by some sort of
> compiler that allows you to write plugins to extend the syntax in your
> programming language of choice to do the things that LaTeX packages do
> (and build up a community that writes such packages, like LaTeX has
> around it).  This would allow one to have packages that directly
> formatted all sorts of XML data.

Years ago, when I was working in SIL on the LinguaLinks project, we had
a notion much like this, save that we were using a proprietary
programming language built on top of Smalltalk.  The notion was that
every object (of which feature structures would be one class), in
addition to its "programming" properties, knew how to display itself in
various forms.  Most of these views of an object were used in the
editors for the object or for editors of objects containing it, but
there was also the intention to make views suitable for use inside a
text document.

There were lots of good things about this (if you changed your mind
about a feature in your lexical entry, the view inside your glossed text
example inside your grammar paper also changed).  In the end it
foundered for a variety of reasons (memory hog, slow, poor user
interface, bugs, not enough developers...).  But in many ways it was
ahead of its time.

The current SIL work, under the FieldWorks project, envisions a sort of
XML database (built on an SQL db backend).  It's not explicitly built
around HPSG, though, and it's going slowly.  But the idea of embeddable
objects with views is still there.

On another subject, Ken Shan wrote (quoting me):
>> LaTeX is a programming language for words, equations etc.,
>> while MathType is a wysiwyg solution.
>
> That's a little bit like saying that Prolog is a programming language
> while HPSG is not. (:

Hmm--I agree with this.  So if it was intended to be ironic, it went
over my head :-).  As for the preview-latex app, I think it's great to
make the code-compile-run loop shorter, but there's nothing for a naive
user like eliminating the loop altogether.

     Mike Maxwell
     Linguistic Data Consortium
     maxwell at ldc.upenn.edu