uima-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Steven Bethard <steven.beth...@gmail.com>
Subject Re: Removing descriptor files from ClearTK
Date Fri, 05 Jun 2009 18:17:19 GMT
On Thu, Jun 4, 2009 at 9:06 AM, Philip Ogren <philip@ogren.info> wrote:
> 1) maintaining descriptor files is a giant pain in the butt.  The developers
> of ClearTK are two graduate students and a postdoc and we do not have the
> resources (or patience) to maintain these files.  We have found that as we
> have evolved and refactored our code that our descriptor files are
> constantly breaking and are absurdly burdensome to maintain.
[snip]
> 2) We cannot create all possible descriptor files that might be needed to
> use ClearTK in the ways desired by the user.  Our library relies heavily on
> dynamic class loading driven by class names provided in configuration
> parameters.

We had a couple of nice discussions about these issues at SETQA-NLP.
I'm still thoroughly convinced that the canonical version of an
AnalysisEngine descriptor is better represented in Java (or C++) code
than in an XML descriptor - for the sake of refactoring, type
checking, a much simpler mechanism of setting configuration
parameters, etc. However, given that the AnalysisEngineDescription
class has a number of toXML() methods, it should be straightforward to
automatically generate the appropriate XML descriptor from an
AnalysisEngineDescription object.

A nice side effect of this approach is that our users can use the
simple factory methods for creating AnalysisEngineDescription objects
with whatever configuration parameters make the most sense as defaults
for them, and generate a new XML descriptor for themselves. And
they'll get their configuration parameters appropriately type checked
before they ever have to actually load the XML descriptor.

Steve
-- 
Where did you get the preposterous hypothesis?
Did Steve tell you that?
        --- The Hiphopopotamus

Mime
View raw message