"Mikel L. Forcada" <[email protected]> čálii:

[...]

> What I would like to see is a proper lightning-fast finite state 
> implementation of VISL CG3 (or at least, of a large subset of the kind 
> of rules supported by VISL CG3). Is there a GSoC idea to finish this up? 
> Who would be able to work on it?

There already was one :-)

Dávid Nemeskey did some great work:
- http://wiki.apertium.org/wiki/User:David_Nemeskey/GSOC_proposal_2013
- http://wiki.apertium.org/wiki/User:David_Nemeskey/GSOC_progress_2013
- http://wiki.apertium.org/wiki/User:David_Nemeskey/GSOC_2013_literature_review
- http://www.aclweb.org/website/old_anthology/C/C14/C14-1073.pdf

but "the speed of our parser falls short of that of VISL CG-3 — with the
exception of the execution of very small grammars".

VISL CG-3 has seen quite a lot of tuning for speed over the years (and
in fact does use things like tries as part of the implementation,
although the whole rule application is not finite-state), and I suppose
that's the reason for the title of the paper.

The paper mentions one unimplemented speed-up idea, but sounds a bit
complicated; I think the low-hanging fruit has been picked already.


-Kevin

Attachment: signature.asc
Description: PGP signature

------------------------------------------------------------------------------
Transform Data into Opportunity.
Accelerate data analysis in your applications with
Intel Data Analytics Acceleration Library.
Click to learn more.
http://makebettercode.com/inteldaal-eval
_______________________________________________
Apertium-stuff mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/apertium-stuff

Reply via email to