Meeting setup

Date: 09.01.2006
Time: 09.30 Norw. time
Place: Wherever we are :-)
Tools: iChat, SubEthaEdit

Agenda

Opening, agenda review
Reviewing the task list from two weeks ago
Documentation - divvun.no
Corpus gathering
Corpus infrastructure
Linguistics
name lexicon infrastructure
Other issues
Summary, task lists
Closing

1. Opening, agenda review, participants

Opened at 09:45.

Present: Børre, Sjur, Trond

Absent: Maaren, Saara, Thomas, Tomi (sick leave until Jan. 17th)

Main secretary: Børre

Agenda accepted as is.

2. Reviewing the task list from the last meeting

Børre

Gather public texts, preferrably also parallel ones
- Not done
Contact Odin editor to ask for source (and parallel) documents
- Not done
Continue converting text from input format to our xml
- Not done
Document the corpus infrastructure
- Done
review code and documentation for corpus xsl files under version control
- Not done
review the convert2xml.pl script, what works well and what doesn´t.
- Done
comment review template made by Saara
- Not done
fix bugs!
- Not done
Misc.
- Have moved some documentation from the gt/doc tree to xtdoc/giellatekno and some of the content to xtdoc/sd as it seemed to fit better there. Have worked on translating these documents.
- Had a look at different formats for new testament. Me and Trond found that the paratext format is the best for making parallell corpora.

Maaren

work with risten.no
review corpus usage documentation, and the usage of the corpus
discuss with relevant people regarding seminar on proofing tools, normativity and SGL in February/March, including place.

Saara

Look at the corpus infrastructure issue
Convert the name lexicon from present format to xml and test it
- not done
discuss the new lexicon format in the newsgroup
Language detection for the corpus files
- Implemented with heuristics for Sámi. Not yet in use, since there is still problems with Finnish: it is not properly distinguished from Sámi languages. I have to look for larger text material for the language model for Finnish.
Review the hyphenation detection.
- In progress.. There is not enough hyphenated texts in our corpus, I will probably test with Finnish.
Catxml review: look at the ccat tool
- Installed and documented, can be reviewed as well.
Review the handling of xsl-files in corpus infrastructure
- Started major update for all the xsl-files. xsl-processing is not yet included in the web upload script, and the xsl-files do not correspond to the current corpus.dtd. The basic functionality is there however, so finishing and testing the scripts takes two weeks (part time work).
Do some testing for bug #211
- not done
add version control of the corpus xsl files to the (upload) processing
- related to xsl-files, see above.
xml2lexc update to handle complex names: construct entries like we have now from the different parts of a complex name entry
update preprocess to handle inflected forms of complex names
fix bugs!

Sjur

Follow up the lawyer treatment of the contracts
- not done
Lule Sámi twol problems, with Thomas and Trond
- nothing
follow up on voice group-chat not working to Sámediggi
- Test Marratech when the new Marratech server is in place
  - still waiting
project planning with Trond, continued
- also look at the development processes - specification and testing
  - nothing this week
Follow up on place names from Norge Digitalt
- waiting for Bjørn Olav Megard
- should call EDD at UiO: they’re planning a national place name lexicon
Evaluate SFST as speller (and analyzer) lexicon
- more thorough analysis than was possible in Guovdageaidnu
write a background document on the corpus contracts
- nothing
continue proper name lexicon work and discussion
- nothing
public tender
- updated the selection criteria with the latest remarks
- finally found and contacted a firm that helps with public tenders and the legal and formal requirements associated with such - now waiting for a price quote from them
smj G3 issue with Thomas and Trond
- nothing
sme G3 issue with Thomas and Trond
- nothing
fix bugs!
- nothing
other:
- risten.no had crashed again just before Xmas, but it wasn’t noticed until now (it looked like it was still running until New Year). The last backup was dated 22. dec, which means that some work had been lost. I have a set of logs that can be analysed. risten.no is back up now, but searches are only done on sma for unclear reasons. I’m still working on it.

Thomas

Sick leave.

Tomi

Sick leave.

Trond

Follow up the lawyer treatment of the contracts ** Done, got feedback from university lawyer, he gave green light (with no comments)
project planning with Sjur, continued
- Nothing with Sjur, done some on disamb.
smj G3 issue with Sjur and Thomas
- Not done.
sme G3 issue with Sjur and Thomas
- Not done.
document corpus infrastructure, your part
- Not done.
review corpus usage documentation
- Not done.
discuss the new lexicon format in the newsgroup
fix bugs!
- Some postings, yes.

3. Documentation

Reviews

Sjur: What is the reasoning for moving the documents around? Børre: Some documents sitting in the gt/doc tree seemed to fit best in the xtdoc/giellatekno tree.

Add documentation on our corpus infrastructure and our corpus work in general (Børre, Tomi, Trond). For the basic corpora, we need 2 additional types of documentation, or doc for 2 target groups:

For the users/linguists: What corpus are found, how do I use them (this info is now scattered) (Part of the HOWTO USE is documented in the catxml docu, what documents are found where etc + an overall documentation is not written, since the corpus is so sparsely populated)
1. catxml done, which is what is needed mostly. Do we need more?
2. Review: Thomas, Maaren, Linda, Ilona, Trond

Saara has installed ccat. It is ready for a first round of reviews.

Findings so far: Basic text presentation works fine, but the text-type options does not.

4. Corpus gathering

Contracts

Next step:

wait for comments from the lawyers - remind them of the task
1. Trond is done (no comments, ok by UiT), Sjur still to do it
possibly update contracts with remarks from lawyers
start using them!

Collecting

Nothing new, now hampered by the lawyers checking the final version of the contracts.

We want both parallell and errouneous (unproofed) text files. What we need is a direct contact with Odin, in order to have as good coverage as possible.

TODO: Trond, and then Børre to call Ove Sæth to re-establish contact.

Bible texts

We have received Norwegian texts and a contract draft from Bibelselskapet, in essence requiring a separate contract for internet use of extracts of the text. The texts are in the paratext format, an international standard for the bible in electronic form. Most translations are available in this format.

TODO: Trond will accept the contract as is, and then negotiate a separate contract with them for use with the online, searchable parallel corpus when it is ready.

Example of the paratext format

The linebreak in v1 is ours, in the original it is all on one line

\s Gud skaper verda
\p
\v 1 I opphavet skapte Gud himmelen og jorda. \f + \fr 1,1 \fk skapte \ft Det hebr. verbet som vert nytta her,
     står alltid med Gud som subjekt.\f* \x - \xo 1,1 \xt 2,4; Job 38; Sal 8; 33,6; 104; Joh 1,1.3; Kol 1,16\x*
\v 2 Jorda var aud og tom, og mørker låg over havdjupet. Men Guds Ande sveiv over vatnet.
\v 3 Då sa Gud: «Det verte ljos!» Så vart det ljos. \x - \xo 1,3 \xt Sal 33,9; 2 Kor 4,6\x*
\v 4 Og Gud såg at ljoset var godt, og han skilde ljoset frå mørkret.
\v 5 Gud kalla ljoset dag, og mørkret kalla han natt. Og det vart kveld, og det vart morgon, fyrste dagen.
\p

5. Corpus infrastructure

Task list:

Include the xsl files under version control
1. RCS version control is almost finished, but an issue with access control is still open. Discussed a bit in the meeting, but nothing conclusive. We’ll continue the discussion in the newsgroup.
Incorporate language detection as part of the corpus processing (Saara)
1. Almost finished. Some heuristics regarding other Sámi languages in the same document to be added.
we need a way to deal with hyphenated documents (documents with (manually) inserted hyphenation marks) in catxml/preprocess.
1. done, needs review (Saara)
we need to review whether only automatic hyphen detection is good enough, or whether manual post-processing in some form is needed. Delayed until we have some results to base the review on.
1. Acceptable results: 90% of all real hyphens correctly tagged.
CGI-admin script to add xsl-file to a corpus file that doesn’t have one (Tomi)
1. Saara will review the existing code, consult Tomi, and try to make a script to help Børre utilize the template and the infra in place.

6. Linguistics

North Sámi

three-part compounds issue still open
- Thomas is finished with three-part compounds for North Sámi.
  - compilation time has increased to 10m for the sme parser as a whole. This is probably linked to the three-part issue, but we are not sure about that.
  - it overgenerates wildly. This is bad for several of our applications.
- compound shortening issue is waiting for the SGL to make a normative decision
  - we have to wait until SGL´s meeting protocol (???) is ready
- linguistic analysis/discussion to continue in the newsgroup
- we should include the members-to-be of the Sámi Giellalávdegoddi (SGL) in this discussion, but how? We need a common meeting / normativity seminar / seminar on proofing issues and on language technology in a broader sense with them. Then we also need to open the linguistic issues newsgroup we talked about earlier, and set up their computers and teach them to use the newsgroup. This seminar needs to include the Lule Sámi section of SGL (others are welcome too).
  - Laila has informed the members of the SGL about this. This is very okei - open the linguistic issues newsgroup for the members of SGL, but we have to know how to do it. Seminar also very okei, but we have to wait until the new SGL is elected.
- Timetable:
  - The Sámediggiráđđi is going to appoint new members in January (Johan Mikkel Saara is asking people)
  - Contact the SGL when it is elected, and ask them to arrange a normativity seminar, with invited keypersons. Have to wait until it is elected. Initially we planned for February, but it is too early, because SGL is not elected yet. Divvun wants the seminar as early as possible.
number project still open (see below)
diphthong simplification/G3 issue should be carried over from Lule Sámi.
- TODO:
  - Trond, Thomas and Sjur to discuss this this month.
  - Writing the rules (copy from smj, adjust) (Sjur, Thomas and Trond)
  - Bug 77 - clearify whether háliid is acceptable - it is found, thus we need to be able to analyse it, but only as !SUB:

Lule Sámi

Open tasks:

Derived G3 that looks like G2 are still open.
- Thomas, Trond and Sjur should meet again to finish the rest.

Numerals

The following North Sámi linguistic issues should be settled before going into the numeral project:

Three-part compounds
Diphthong simplification
Derivation

These issues are recently done in Lule Sámi, and it is more efficient to complete them in North Sámi directly thereafter instead of beginning a new topic

Numeral treatment is on different level in the existing sme and smj parsers, but the issue itself is common to the two langauges, and should therefore be treated in parallel.

Numerals in North Sámi: Inventory is listed elsewhere.

Numerals in Lule Sámi: There are 70 lines of code setting up the structure for case inflection of basic numerals.

7. Name lexicon infrastructure

Complex names

Task list for this issue:

make sure xml2lexc can handle complex names in ways compatible with our present tool chain (=reconstruct the lexc format we have now) (Saara)
- Tomi is working on a C++ version of the xml2lexc tool. He will put it into CVS soon (it is actually catxml, can easily be adopted to the xml2lexc requirements)
the resulting file format should be identical to our present prop-name file (=lexc), that can then be converted to our new xml format using the same script as for the regular names (Tomi or Saara, but only when the technical details are settled)
The core issue: The preprocess script can handle uninflected (A + B) compounds, but not (A + inflected B). Saara will add the analyzer as part of the preprocess, to be able to handle inflected cases as well.

XML format

We had our meeting, and the result was pretty close to the structure of the existing risten.no term base. Now the conversion script to xml needs to be updated. The new format is documented in the newsgroup.

Tasks:

make a test lexicon for evaluating the format, set up the editing, and test it (Børre)
update conversion from lexc to xml to reflect new xml format
testing of conversion
continue the discussion of the name lexicon format (Saara, Tomi, Sjur, Trond)
implement a prototype in eXist
eXist as editor:
1. develop the needed XQueries and interface
2. synchronisation between risten.no and
3. test whether eXist as editor is actually working well

8. Other

Seminars

SGL/normativity seminar
- all members = potentially/likely all languages
- date? As early as possible, but not likely before the end of February
- place? Maaren will investigate
project meeting
- date: 23.-27. januar.
- place: Tromsø/Helsinki?
- still too many open questions re place+date, to be determined in a separate meeting later this week, when all are available.

Technical issues

The mac os / perl bug (at least Trond and Sjur has it, [Bugzilla #211|http://giellatekno.uit.no/bugzilla/show_bug.cgi?id=211]):
- utf8 “\xC4” does not map to Unicode at /Users/trond/gt/script/preprocess line
  1. This msg did not show up in 10.3 (perl 5.8.1), but does so in 10.4 (perl 5.8.6). It is probably a perl - OS mismatch. (Trond, Saara, Tomi)
  - 10.4 introduced support for locales in the shell (10.3 and earlier didn’t know about locales)
- Test: the result of the last line should indicate whether this is a problem in cat or a Perl/OS mismatch.

   preprocess file_name.txt OK
   cat file_name.txt | preprocess !!
   catxml file_name.xml | preprocess ??

After some heavy investigation, we got no further. There is no difference between different bash versions (2.0x, 3.0x), neither between different locales as long as they are UTF-8, and whether it is stored in LANG or LC_ALL.

One new insight though: both cat and print gives the same errors, thus indicating that the error is NOT strictly reltated to cat.

Bug fixing

29 open bugs (and 25 risten.no bugs)

9. Summary, task list