[developers] For Grammar Engineers: requests for anecdotes

Francis Bond fcbond at gmail.com
Fri Feb 22 06:47:08 CET 2013

I am not sure which category this falls into, but I have often found
maintaining Jacy harder than it should be because of old analyses
remaining in the grammar even when replaced by new ones, it was often
time consuming (for me) to work out which bits were actually being used.

On 21 February 2013 19:24, Antske Fokkens <afokkens at coli.uni-sb.de> wrote:
> Dear fellow delph-in grammar writers,
> Emily and I are working on a paper on linguistic hypothesis testing through
> grammar engineering and CLIMB.
> We would like to include some anecdotes from grammar writer's experience on
> problems while changing analyses deeply embedded in the grammar.
> We'd therefore like to ask you whether you have some examples of cases where
> you either:
> 1) decided not to try out an alternative approach or change an analysis,
> because it was embedded too deeply in the grammar
> 2) changing an analysis was particularly labor intensive and complex,
> because of the same reason
> 3) you made a decision between alternative approaches only to notice it was
> the wrong one after it became much more complex to adapt
> Any feedback or input on these issues is very welcome, level of detail is up
> to you (though at least some indication of the phenomenon and main issue
> would be helpful).
> Thank you very much,
> Antske

Francis Bond <http://www3.ntu.edu.sg/home/fcbond/>
Division of Linguistics and Multilingual Studies
Nanyang Technological University

More information about the developers mailing list