[developers] CSLI vs MRS profiles

[Alexandre Rademaker]

Hi,

A couple of months ago, I was editing the page http://moin.delph-in.net/MatrixMrsTestSuite and Stephan corrected my mistake and explained to me the origin of both datasets:

> the MRS test suite is something that ann and dan cooked up over the course of five or so weeks while dan was visiting cambridge, 2001 or 2002, i would say.  except for some reuse of Abrams and Browne, I doubt there is any overlap in actual sentences with what was originally called the HP test suite.  the latter was created to explore variation in syntactic structures and lives on in the DELPH-IN universe under the name CSLI test suite (since around 1994).  the MRS test suite, on the other hand, exemplifies basic semantic constructions.  so, in my view it is misleading to say it was derived from the HP data, but dan was of course centrally involved in both efforts.

I am playing with these profiles again, trying to have both translated to Brazilian Portuguese (not the European Portuguese translation we have in the wiki). But I have a question 

Why 7 sentences are different if we compare http://moin.delph-in.net/MatrixMrsTestSuite to http://svn.delph-in.net/erg/trunk/tsdb/gold/mrs/? Diff below. Does anyone remember what happen? Should we update the wiki?

The page http://moin.delph-in.net/MatrixMrsTestSuite is now marked as Immutable, but I believe the sentence 

> Currently, there are test suites for the following languages (included in the [incr tsdb()] software package) 


is misleading. We don't have a [incr tsdb()] package with profiles. The page http://www.delph-in.net/itsdb/ has a link to http://lingo.stanford.edu/ftp/latest/ but it was not working. I found three profiles from the root of LOGON tree:

% find . | grep mrs/item.gz
./lingo/erg/tsdb/gold/mrs/item.gz
./lingo/terg/tsdb/gold/mrs/item.gz
./dfki/gg/tsdb/gold/mrs/item.gz

So the profiles are not included in [incr tsdb()], they are maintained with grammars, right? We don’t have profiles with all the languages listed in the wiki, only two.

We have also some discussions in https://delphinqa.ling.washington.edu/t/matrix-mrs-test-suite/484. In the forum, I suggested other changes in the wiki:

> The wiki is confusing since on the main page we have translations for Japanese and a relevant discussion about the structure of the set only at the bottom of the page. Moreover, in http://moin.delph-in.net/MatrixMrsTestSuiteEn links are all broken. Does anyone know what is this old server that the links like http://cypriot.stanford.edu/~bond/mrs-en060524/11.html point to? How can I help to make them work again?


Any comment? 


Best,


% diff matrix-en.sent mrs-en.sent
20c20
< Cats bark.
---
> Cats go.
22c22
< Some bark.
---
> Some went.
53c53
< Chased dogs bark.
---
> Chased dogs go.
55c55
< That the cat chases Browne is old.
---
> That the cat chases Browne is obvious.
62,64c62,64
< Browne's barks.
< Twenty three dogs bark.
< Two hundred twenty dogs bark.
---
> Browne's goes.
> Twenty three dogs go.
> Two hundred twenty dogs go.
79c79
< Abrams promised Browne to bark.
---
> Abrams promised Browne to go.
97c97
< The cats found a way to bark.
---
> The cats found a way to go.




--
Alexandre Rademaker
http://arademaker.github.io