[Lexicog] Re: Toolbox parsing bug

[Claire Bowern]

On Wednesday, January 4, 2012 at 10:18 PM, Mike Maxwell wrote:

> On 1/4/2012 12:06 PM, langwijmijij wrote:
> > I tried Fieldworks, but life is too short to wait more than 5 seconds
> > for each parse (with an average of 5+ morphemes per verb, I can
> > interlinearise by hand at more than double that speed). The parser is
> > incredibly slow, even on a recent machine with ample memory.
> > 
> 
> 
> Odd, I haven't heard people complain about FLEx being slow lately. 
> Which version of FLEx were you using, and which parser (XAmple or Hermit 
> Crab)?
> 
> 
> 


Can't remember, sorry. But it wasn't just the parser, it was everything about the program, switching windows, etc. There was a long discussion on the rnld email list about this a few months ago, with people saying things like "it's ok, but I have to close all other programs to run it and I have to shut down the program and restart every 20 mins".
 
> 
> > And it still had most of the difficulties that toolbox
> > has for this language (lenition causing many homophonous forms,
> > significant variation in affix forms which are ameliorated by word
> > formulas but not solved, etc).
> 
> I assume the homophonous forms can be disambiguated by part of speech of 
> the base, or by position wrt the inflectional affix template, right?
> 
> 
> 


Many, yes, but not always. For example, there are two tense suffixes, one of which is -an and the other of which is -na, and both can be realised as -n in certain circumstances. They are also immediately adjacent to one another.
> 
> (And maybe some by their morphosyntactic properties, e.g. incompatible 
> features where you have multiple exponence.) Did you assign those 
> properties to the affixes? (FLEx of course does not use word formulas, 
> although the notion of an affix template is conceptually similar.)
> 
> 
> 

I started doing so but given how slow the rest of the program was being, it didn't seem worth it to pursue it. 
> 
> > It also crashed frequently (took the best part of a day just to
> > import the toolbox lexicon).
> > 
> 
> 
> The difficulty in importing Toolbox lexicons has traditionally (I
> haven't tried it lately) been that the Toolbox lexicons have too many
> errors (missing fields, fields which are ordered wrong, data in the 
> wrong field, and so forth). I wouldn't be surprised if it took awhile 
> to straighten those things out. Did you use the "Import Wizard" 
> described here:
> http://fieldworks.sil.org/flex/lexicon/import-export/
> 
> 
> 


Yes. This was a stripped-down lexicon of just headwords, parts of speech, and glosses.

Claire
 
> -- 
> Mike Maxwell
> maxwell at umiacs.umd.edu (mailto:maxwell at umiacs.umd.edu)
> "My definition of an interesting universe is
> one that has the capacity to study itself."
> --Stephen Eastmond
> 
> 
> 


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/lexicography/attachments/20120108/176c519e/attachment.htm>