meeting_2007-01-23

Meeting with Polderland 23.1.2007

Participants:

  • Peter Beinema
  • Sjur Moshagen

Agenda

  • Since last time
  • Possible issues
  • Next meeting

Since last time

Polderland:

  • did poll technical contacts, no answer yet
  • mklex: beta test results:
    • some things we can't / won't deal with (multiword expr, words starting with hyphen, words containing '+')
    • repaired for problem with initial non- A-Za-z, requires rerun of test
  • investigating possibilities for 2 extra word/noun classes:
    • proper noun (as second part of compound), making it possible to limit the 1st parts
    • word that can follow genitive stem (as second part of compound), making it possible to limit the 1st parts too
  • words in lexicon starting with hyphen:
    • is not even processed "correctly": hyphen is skipped prior to checking so, if lexicon and text contain "-abcde"(without quotes), the speller will flag "abcde" and suggest "-abcde", resulting in "--abcde" (and then "abcde" is flagged again). Has to do with handling of characters such as " ' _ - etc.

Divvun:

PLX conversion made good progress last week as well, noun compounding now ok, and all POSes except numbers should be ok. We hope to deliver the first large-scale PLX lexicon today or tomorrow. It will not contain derivations, though, which will account for a large portion of the total size.

Possible issues

Name "prefixes"

Some nouns are common as prefixes to names, mainly words for North, South, etc. Even though these are regular nouns, when prefixing proper nouns they should be capitalized, and there needs to be a hyphen between the prefix and the second part. Example (spelling error -> correct form):

  • davvi-Norgga -> Davvi-Norgga (= North(ern) Norway)

That is:

name-prefix + name => upper case + hyphen

Thus, to correctly handle these cases, we need to identify names as different from other nouns, such that we can direct the upppercased and hyphenated prefixes to the correct second parts.

Compound where first part is genitive:

similar issue as with name-prefix. Proposed solution: new word (sub)class that can act as compound following a genitive only

Hyphen as prefix

In constructions of coordinated compounds with common first part (YX and YZ => YX and -Z: bussbillett og -sjåfør) it is customary to replace the common part with a hyphen. We could reduce the size of the noun lexicon by 50% if we can specify the hyphen as a prefix, instead of, as now, pregenerate all nouns in a version with a prefixed hyphen.

Decision: such constructions will be handled automatically, and should not be included in the lexicon.

Next meeting

Next Tuesday (30.1.) at the usual time.

TODO:

  • get back on linguistic issue regarding proper nouns vs. common nouns
    • done, see above
  • get back on linguistic issue re. hyphen as prefix
    • hard-coded, the speller will accept both with and without hyphen
  • check if North Sámi hyphenation can be disabled when processing Lule Sámi (PLD)
    • can be done with a simple work-around
  • make complete PLX data set (Tomi)
    • approaching: -)
  • get language codes to work with Mac Office 2004 (and check MacOffice 2007) ( Polderland)
  • deliver mklex + hyphen script
  • try to find proper compiler version for Adobe Indesign (old version will probably do)
  • try to get an answer to the language codes in MS Office for Mac question from other sources (Sjur)
    • done, responded via e-mail