[developers] ERG demo webpages
Stephan Oepen
oe at ifi.uio.no
Tue Apr 29 13:29:42 CEST 2008
thanks for the feedback on the ERG web demo, ann.
> I don't think it's a good idea to allow non-sentences through in the
> demo as is happening currently. The rubric implies one has to enter
> a sentence, and when one just gets MRSs (as by default) the
> non-sentence nature of some of the analyses is very unclear.
>
> the sample `Just try pressing return in this window' probably ought
> to have some punctuation, but maybe rethink the example given what
> gets generated? `Try pressing return in this window justly.'
i will leave these for dan to decide. right now, i dropped the `just'
from the default example, but maybe we can find an even better sample
sentence?
regarding degrees of robustness in the web demo, i would like to have a
toggle, where `robust' mode could then comprise fragment roots plus PoS
tagging and generic lexical entries; `precise' mode, on the other hand,
could be for in-depth linguistic inspection. i still hope to implement
this before the summit this year.
> the generate button is active on the generator output but nothing
> will generate
that was a nice little bug: MRSs on generator results were not external
(compliant with the SEM-I, i.e. post-VPM); now corrected.
in this latest face-lift i also addressed a few issues dan had reported
earlier, viz. (a) the choide of tree comparison mode is back; (b) it is
possible to generate from results even if only tree display was active;
(c) tree nodes are now mouse sensitive. finally, i added LaTeX output
for trees and MRSs, but i still need to find a good way of distributing
the corresponding style files.
best wishes - oe
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
+++ Universitetet i Oslo (IFI); Boks 1080 Blindern; 0316 Oslo; (+47) 2284 0125
+++ CSLI Stanford; Ventura Hall; Stanford, CA 94305; (+1 650) 723 0515
+++ --- oe at ifi.uio.no; oe at csli.stanford.edu; stephan at oepen.net ---
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
More information about the developers
mailing list