On 19 March 2014 18:19, Alexandru Florescu <[email protected]> wrote:
> Hey,
>
> Basically, as we talked before, we agree on having a FST and do some
> searches (probably breadth-first) over the analyses. Right?
>

Who is 'we'? You're missing context here.

Breadth-first, yes, because it would need to visit all analyses, and
it's a simple extension to the existing analysis(/tokenisation) scheme
to add a pair of state queues. You can't use most of the more
complicated search algorithms without a notion of cost/distance. If
you can think of one, by all means, mention it.

>
> Do you think there would be any real downsides on doing all the work in the
> compiler?
>

Doing it as an extension to the existing compiler is only marginally
more work than the compiler for the searching analyser. In either
case, you have to expand the potential forms at compile time; the
difference is, because the regular analyser needs lexical forms, you
would need to collect those. This is why I recommended building a
second, in-memory transducer, for analysis-as-generation. All of the
runtime downsides are exclusive to splitting analysis into two stages.

-- 
<Sefam> Are any of the mentors around?
<jimregan> yes, they're the ones trolling you

------------------------------------------------------------------------------
Learn Graph Databases - Download FREE O'Reilly Book
"Graph Databases" is the definitive new guide to graph databases and their
applications. Written by three acclaimed leaders in the field,
this first edition is now available. Download your free book today!
http://p.sf.net/sfu/13534_NeoTech
_______________________________________________
Apertium-stuff mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/apertium-stuff

Reply via email to