On 13-09-26 05:40 PM, Marshall Schor wrote:
On 9/26/2013 5:00 PM, Richard Eckart de Castilho wrote:
Personally, I'd not make any modifications to the descriptor at all, but rather
would just skip the delegate when programmatically creating the descriptor. We
do that all the time in our experiments. But if that is for some reason not an
option and the extension is a strong requirement, the change should at least be
made at the location that conceptually makes most sense (imho).
@Marshall: do you want to provide some more background why you do not simply
create the descriptors programmatically and externalize this skipping,
including, etc. into your experimental setup?
I'm probably not the right person to respond here; maybe other users want to
weigh in. Just to clarify, I'm not the person having the experimental setup
:-) .
Here's my 2 cents: To promote interoperability and wide-spread adoption of
annotator technology, the thought was there could easily be multiple kinds of
"roles" of people, having different skills, involved in setting up and running
applications. One role would be annotator writers, writing in C++ or Java or
.... Another role would be an "assembler" / configurer of pipelines from
parts. UIMA was intended to support the notion that some people would be
comfortable with programming; but others would not. Those others could be
comfortable with configuring pipelines using XML or using a more friendly tool
like a GUI editor (such as the Component Descriptor Editor, CDE) to do this.
I think that's a main source of this desire to have certain aspects of pipeline
assembly "externalized" outside of programming.
I do not know about the typical users of uima, but I can say that this
separation of roles does not apply to us at all. We build complex
pipelines programmatically using uimaFIT and are quite happy with the
results.
--
Alexandre Patry, Ph.D
Chercheur / Researcher
http://KeaText.com
