[developers] pending/777: generation does not work (pending)

[Ann Copestake]

oe at csli.Stanford.EDU said:
> ann, can you comment on the intentions here?  without thinking hard about it,
> it seems that having to duplicate a letter set definition, so that it can be
> used for only one side of a sub-rule, is hardly desirable?

I think my earlier comments on this issue should be archived in the developers 
list.  I currently do not believe this will ever genuinely be necessary but am 
willing to rethink if there's demand.

oe at csli.Stanford.EDU said:
> i just checked in a patch to make it more robust (see attachment), 

Thanks

> but i
> guess we should decide on our attitude regarding rules like these.  i think
> either we should allow such rules (liberating people from having to put dummy
> %suffix() lines on irregular-only rules), or there should be a warning at
> grammar load time.  i sort of like the purity of using the rule-orthographemic
> p() test instead of the clunky NEEDS-AFFIX, but on the other hand i believe
> that some grammars use morphological rules that are (practically) only
> triggered from `irregs.tab'. 

I _think_ we've discussed this before but could be wrong.  The general point 
I'd make is that while there's lots of ways of changing things to be 
marginally nicer with the current set up, I am not particularly anxious to 
tinker with it, because the % mechanism is a messy hack anyway.  We need an 
approach which allows for inheritance of suffixation properties, which means 
we need to reify the spelling change rules - e.g., so they can be associated 
with types and be in an inheritance hierarchy.  I would prefer that, if things 
can be made to work with the current mechanisms, we don't make any further 
changes now until I have time to do the cleaner approach because otherwise 
I'll have to support more variants as legacy systems.  So, to return to 
specifics - the warning at grammar load time would be a good idea, I think.

I'd very much welcome further discussion of what the morphology should and 
shouldn't do, e.g., in principle, what do people want the relationship of 
irregular specifications to the lexicon and the regular rules to be?
Please look at the earlier stuff on developers if you're interested in this, 
cos that's the only documentation / discussion that exists right now.

Ann