A spell-checker for Sami

Intro

The Sami Parliament in Norway plans to make spell checkers for North, Lule and South Sami, within the eSápmi program.

The Sami Parliament plans

Sjur Moshagen's origina 2003 plan
Sjur Moshagen's revised 2004 plan

Our lexica seen in a spellchecker perspective

At the moment, our lexica (*-sm?-lex.txt) are not normative, but descriptive, recognising both standard and substandard spelling conventions. If our lexica shall be used for spellcheckers, it must distinguish between standard and substandard forms, in the way described in Beesley's and Karttunen's book.

Different technologies

The spellchecker plans involve different technologies. Basically, the distinction goes between automata and transducers, i.e. in this context, between systems unaware and systems aware of grammatical information.

transducer-based spell checkers

hmm...

automaton-based spell checkers

Some applications will need an ispell-type interface. This should be made, either by using the original ispell method, or by compressing a fullformlist to an automaton automatically.

Trond Trosterud

Last modified: Sat Mar 27 23:55:19 2004