Mātāpuna Dictionary Database System The Open Source Multi-user Web-based Dictionary Writing System Dave Moskovitz DWS 2004, Brno.

Mātāpuna Dictionary Database System

The Open Source

Multi-user

Web-based

Dictionary Writing System

Dave MoskovitzDWS 2004, Brno

Mātāpuna – Dave Moskovitz – www.thinktank.co.nz

Outline

Background info

Design criteria

Functions

Database structure

Future development

Lab

Call for collaboration


Background – New Zealand / Aotearoa / Māori

4m people; 268,000 km2

15% Māori; 1 in 4 of those speak Māori

Median age 22; median income NZD14,000(compared to 35 and 18,500 for pākehā)

Māori is an official language, polynesian language group

Uses standard roman character set with macrons


Background – New Zealand in the Pacific


Background – The Mātāpuna project

First monolingual dictionary of Māori, written from Māori cultural perspective

Target of 20,000 entries (1 entry = 1 definition)

Designed for language learners with some proficiency

3+ year project under auspices of Te Taura Whiri i te Reo Māori / The Māori Language Commission

4 writers, one editor, one lexicographer, one project manager, admin support, one geek


Background – The Mātāpuna team

Sharon ArmstrongWiha Te Rakihawea

Te Waireka WalkerPou Temara Phil Matthews

Ruka Broughton

Hēni Jacobs

Not in photo: Te Awanuiārangi Black, Te Haumihiata Mason, Dave Moskovitz


Background – Dave

BA (Hons) Comp Sci Univ. California Berkeley

Began PhD in Applied Linguistics – NZ Sign Language phonology

25 years in IT industry

Background in Application Development, Systems Architecture, System Performance, Internet

3rd lexicography project, after Dictionary of NZ Sign Language and Oxford NZ Dictionary

Open Source bigot


Background – Software

Free Software – Open Source – GPL

Uses Linux, Apache, mod_perl, Postgres, runs on any old hardware (eg Pentium 600)

Browser based

About 4,000 lines of Perl code

Won Computerworld excellence award for use of IT in Government


Open Source is Good for Lexicography

Free

Market is too small to support proprietary software

Everyone’s needs are unique – and you can modify the source code to suit

Open source programmers not hard to find

Low risk and futureproof: no vendor lock-in

Everyone helps each other

Software is open, but data is not (necessarily)


Design Criteria

Easy to use by untrained lexicographers

Support workflow and management as well as entry

End-to-end processing

Produce printed output as well as web access

Multiuser

Multilingual interface, easy to add languages

Unicode-based, allows any character set to be used


Sample Output


Functions

Add

Search

Edit

Corpus search

Reports


Functions - Add


Functions - Search


Functions - Edit


Functions – Corpus search


Functions – Reports


Functions – Validation

Field-based, including:- orthography- punctuation- blank- undefined word / not in defining vocab- synonym rules


Functions – Workflow

Basic workflow:Add → Self check → Editor 1 → Editor 2

Editor can make minor changes, or send the entry back to the owner

Owner is notified of any changes by email

You can always view the history of an entry


Functions – Synonym handling

Entries allow for synomym ‘families’

Master – slave (tuakana – teina) relationship

Masters can’t have masters and slaves can’t have slaves

Slave definitions printed from master

All cross-references managed


Functions – Multilingual interface

186 text snippets

Can add additional languages


Functions – Multilingual interface


Database Structure

wordclass

category

examplesource

headword

qastatus

hwarchive

matapunauser

activityjournal


Future development

Multiple citations

Bilingual / multilingual

More corpus material (and better corpus performance)

Advanced search

Better user administration

XML / SGML export

More languages

… what do you want or need ????


Lab – Words from the Olympics

15 users, 15 categories of words

Rawiri is the editor

Practise entering definitions, linking synonyms, playing with major and minor senses, searching, breaking validation rules …

Be nice to Rawiri, he can send work back to you to get fixed



This is Free Software

Use it and contribute enhancements

It’s robust and capable of producing a major lexicographical work

We are interested in your feedback and participation



Contact:

Dave MoskovitzThinktank Consulting LimitedPO Box 15-212Wellington, New Zealand

[email protected]

+64 27 220 2202

Mātāpuna Dictionary Database System The Open Source Multi-user Web-based Dictionary Writing System Dave Moskovitz DWS 2004, Brno.

Documents

mtpuna dave moskovitz

dave moskovitz slide

mtpuna project

collaboration slide

brno slide

geek slide

pacific slide

government slide