year-suffix heads-up

[Frank Bennett]

Following up on the discussion in this thread ...

http://forums.zotero.org/discussion/12953/some-citations-crash-firefox/

... I have done some playing around, and discovered that there is a
glitch in the handling of year-suffix disambiguation in citeproc-js.
It's a small thing internally, although the surface symptoms are
pretty annoying. Should have a fix in a day or so.

Testing with large numbers of identical entries also raises a question
over performance, which I think can be improved significantly. More
on that in the next few days as well.

Frank

[Frank Bennett]

A little progress on this. I've turned up three buggish things: (1)
failure to continue to treat cites as ambiguous in subsequent
transactions once year-suffix is applied to them; (2) failure to
discriminate between ambiguous subsets within a single "pool" of cites
that are ambiguous in their base form; and (3) testing of all cites in
a maxed-out (i.e. year-suffixed) ambiguous subset against new
partners, even though they are known to be identical.

I've gotten issue (1) sorted out. It's clear what needs to be done to
address issues (2) and (3), and a fix should be ready within the next
couple of days. Once that's done, we'll be ready to revisit the
torture trial linked above in Zotero trunk.

Frank

>
> Frank

[Frank Bennett]

After working with the code some more, seeing clearly what needs to be
done for issues (2) and (3), and beginning the necessary work, I have
concluded that the time to reimplement the disambiguation code for
citeproc-js is now. The current code is simply too badly arranged to
be easily controllable.

The requirements and coding strategy are clear, but it will take a
week or two to get the code in place. When finished, even test cases
of the kind referred to under the link above (52 works by the same
author in the same year, in an author-date style) will disambiguate at
good speed (1-3 seconds or so), and the improved clarity of the code
should give us all a boost in confidence that the thing will work
reliably, and as intended.

More later, when I have something to show.
Frank

>
> Frank
>
>
>
> > Frank