mix & mash

2011 Winners

Congratulations to our 2011 Winners

Newbie Mashup

  • orcon
    Sponsored by:

    Orcon

"Maori Dictionary for Android" by Marielle Lange

http://widged.com/labs/mixmash11/

Newbie Large

Description

It is lovely to see a mobile app built from CC-licensed Te Reo information, and that it came from someone new to public mashups was a double delight. Award sponsored by Orcon

Six months of awesome Orcon broadband goes to the best data mashup by someone who’s never made one before. The winner is selected by our Lead Mashup Judge Nat Torkington.

Selected Judges' Comments:

"Simple, open UI, immediately useful. Well thought out design, from automation to clean up of data sources, to consideration of multiplatform support. Source code and binary distribution - what more could you ask for?"

"Excellent to see some openly licenced material for the Maori language. This dictionary is extremely well polished and is a huge asset to learners of te reo everywhere. From the code base, it is clear that a huge deal of effort was required to transform the input data into something useful in an application. This has been coupled with an attractive application design. I really liked that the source data is available for download in other applications as well. That means that the clean data made for this project will be available for reuse."

Marielle's description of the entry:

Port of Williams Maori Dictionary for Mobile devices. The dictionary, published in 1957 is still seen as one of the best and most complete reference. The mobile version was made possible by a scanned version released of William's dictionary under a CC-BY-SA license thanks to NZETC. The motivation for this is simply that I wanted an excuse to publish an android application. As migrant, my exposure to Maori has been minimal and I found disappointing that no free resources existed on the Android store. That seemed like a project small enough to be done in a maximum of two weeks (what I had available). It also presented a rare opportunity to refresh some old skills in linguistic text analysis. The work includes tools for parsing the xml generated by the OCR tools containing not always reliable information. I used a mixture of regular expressions and manual updates (made quicker by the set up of a custom textmate bundler) to come up with xml tags that better capture linguistic information. I used the experience I had as a psycholinguistic, in a past career. A flexible parsing tool turns the xml into a sqlite database. The database serves as datastore for mobile applications. The largely reworked version of the dictionary is made available as a mobile application, targeting Android devices for now. The application is written in Flex. In theory, in the future, I should be able to target other devices (iPhone) with the same codebase.

Sources