User:ArcticSeeress/sandbox/3

Giellatekno, Centre for Saami language technology at the University of Tromsø (Giellatekno, Romssa universitehta sámi giellateknologiija guovddáš, Giellatekno, Senter for samisk språkteknologi i Tromsø), or simply Giellatekno, is a language technology research group at the Department of Language and Linguistics at the University of Tromsø led by linguists Trond Trosterud and Lene Antonsen.

Early History
The first work on a grammatical model for Northern Sámi began in 1993 with a draft based on Pekka Sammallahti's analyses of the language. The work was continued in 2000 at the University of Tromsø. In 2001, a three‑year language technology project was started at the Faculty of Humanities at the University of Tromsø, with the primary goal being to create a programme for morphological parsing for Northern Sámi. The work was based on Kimmo Koskenniemi's morphological research, including his 1983 doctoral thesis. By 2002, the tool had been uploaded to the university's website. Trosterud stated on NRK Sámi Radio that they were also working on spell checking and translation software. Trosterud had been working alone on Sámi language technology at the university until the Divvun group was established in 2004 by the Sámi Parliament of Norway. From 2004 to 2006, the Research Council of Norway would fund their second project, which included a disambiguator and a searchable electronic corpus for Northern Sámi. Trosterud's work on spell checking was used for Greenlandic with the Kukkuniiaat software. In September 2007, the University of Tromsø advertised open positions for Sámi linguists and programmers/computational linguists for a language technology project in the Sámi‑language newspapers Áššu and Min Áigi. The first Sámi-language spell checker was released in December 2007 in collaboration between the Sámi Parliament and the University of Tromsø.

By 2008, there were three fixed positions in the group, which were occupied by project leader Trond Trosterud and his doctoral advisees Lene Antonsen and Ciprian‑Virgil Gerstenberger. Trosterud had previously worked at a subcontractor for Microsoft developing Norwegian, Swedish and Finnish spell checkers in the 1990s.

In November 2008, Antonsen announced at a conference that they were creating "many types" of language software at the University of Tromsø, among which were tools for teaching Sámi and spelling. These were to be released after Christmas according to Antonsen, and were intended to be "just as good as a teacher". The Giellatekno project was introduced by Trond Trosterud at a seminar during Sámi Week in Tromsø on 2 February 2009, where he talked about their goals to create programmes for Sámi languages and other languages in the north, such as spell checkers and software for textual analysis and language education.

Learning programmes
On 5 February, the day before the Sámi National Day, the team working at Giellatekno released the learning application Oahpa! , which contained five different programmes for learning Northern Sámi: Numra (numbers), Leksa (vocabulary), Morfa (morphology), Vasta (question and answer) and Sahka (dialogue). Upon its release, computers were set up at the University of Tromsø to let visitors test the programme. The application was designed for both children and adults, and allowed for users to learn either of two dialects from Kárášjohka and Guovdageaidnu. It was implented in the Python-based framework Django, with database support from MySQL. Similar software was also planned for other Sámi languages, including Lule, Southern, and Kildin Sámi, in addition to the Komi and Nenets languages. On the same day, they also released the first digital Sámi-language dictionary, Vuosttaš Digisánit (lit. 'The First Digi-Words'), which had been in the works since 2007.

In November 2011, the learning app Oahpa Åarjel! was released for Southern Sámi, with translations to Swedish provided by the Sámi Language Centre in Sweden. Users could also use Norwegian, Finnish, and English.

The University of Tromsø published the online Kven learning programme Meiđän joukko through the Giellatekno website in 2019. It contained 14 chapters of text, audio, and exercises. The work on the project started in 2016, with support from 20 people in "some way or another". The website was still not finished, however, so they included a way for users to send feedback through the website. The courses were primarily designed for students of Kven at the University of Tromsø, but it was open for anyone to use. It was designed to be easy to use, with professor Leena Niiranen hoping it would inspire people to learn the language.

Dictionaries
The first South Sámi online dictionary, "Åarjelsaemien-daaroen digibaakoeh" (lit. 'South Sámi-Norwegian Digi-Words'; also titled Voestes Digibaakoeh ), was released through Giellatekno's website on 5 February 2010, the day before the Sámi National Day. It contained six thousand lemmata, including a thousand place names from the Norwegian Mapping Authority and the Sámi Parliament of Sweden. It translated words primarily into Norwegian, but some Swedish translations were also available. Although the dictionary was not as comprehensive as physical alternatives, it did provide conjugations for words, making it the first of its kind. With it allowing users to be able to find the dictionary forms of conjugated words, it was more accessible for those with only passing knowledge of the language's grammar trying to translate material from it. The developers received support from the Sámi Parliament's Divvun group in developing the dictionary. Giellatekno stated on their website that tools like this were a "condition for languages to surive in a modern society". A report by Synovate for the Ministry of Education and Research claimed that the dictionary was complicated and not user-friendly based on comments from teachers at the secondary school in Kárášjohka.

In September 2010, Giellatekno released a Norwegian-Kven and Kven-Norwegian dictionary in cooperation with the Kven language institute Kainun institutti. It was built upon a glossary collected by Terje Aronsen, and contained 4,371 words by October. Trosterud said that it was not complete, as it was still missing inflected forms.

The first Lule Sámi dictionary to be released online was revealed at a language conference held in Jokkmokk, Sweden, on 17 February 2012. Published through Giellatekno's website, it contained around 18,500 words collected by Anders Kintel, a Sámi-language professor at the University of Nordland. Kintel had previously published a Lule Sámi dictionary, but the rights to it were purchased by the Sami Parliament to make it free for the public to use. It was described by its developers as a "temporary version", with a complete version in the works.

On 9 February 2013, Giellatekno released the dictionary website Neahttadigisánit, which contained about 11,000 words in total for Sámi, Kven and Finnish dictionaries. In June, Ruijan Kaiku reported that Kainun institutti was working "around the clock" to fill out the newly published dictionary with words. The Kven dictionary was only one-way, from Kven to Norwegian, with the other direction needing more time in development. The Kven dictionary was presented at the Nordic Lexicography Conference in August. While they were working on Neahttadigisánit, Giellatekno was also giving techincal assistance to Kainun institutti and Tromsø University's Finnish Department with their work on tools for analysis and spell checking in Kven. By March 2014 the analytic tool was completed and already in use, but a spell checker would take longer. They stated that it could take a year to create the programme, but that a user-friendly version would likely take longer to develop. An "acceptable" spell checker for Kven was made available December 2015. The software infrastructure used for it would also be available for other languages in the Arctic, including in Russia and Canada.

In 2014, Giellatekno released several Skolt Sámi dictionaries, which translated words between the language and English, Russian, Finnish, or Norwegian.

The Oulu University-developed Skolt Sámi online dictionary was released through Giellatekno's website in August 2020. It allowed for translation into Finnish or Northern Sámi, and was released with over 19.000 words. The dictionary received a physical print in March 2021.

Digital keyboards
On 15 December 2014, Giellatekno released free infrastructure for creating mobile keyboards in partnership with Divvun and Australian programmer Brendan Molloy. This was used to create free keyboards for six Sámi languages on Android and iPhone devices; prior, Sámi speakers had to switch between different keyboards if they wanted to write their own language. Jan Tore Sanner, Minister of Local Government and Modernisation, described his hope that the software would encourage young people to use Sámi as their daily language, and Sámi Parliament President Aili Keskitalo expressed joy at overcoming the technical hurdle of writing in Sámi on social media, calling it a "Christmas present for all". The infrastructure was made available for use with other languages, with Ávvir reporting that the software would eventually receive updates with new features. Users nevertheless experienced technical difficulties with the software, such as it slowing down or crashing, with Keskitalo stating that she wasn't "entirely satisfied with the mobile solution". Another app had been released a month prior for iPhones, but it was poorly received due to it costing money, while Norwegian keyboards were free.

In 2017, Giellatekno and Divvun released their own computer keyboards for Windows and Mac, which would be able to write in the Southern, Lule, and Inari Sámi languages. A Skolt Sámi keyboard was also in the works. Previously, people have had to use Northern Sámi keyboards, which one developer described as "not working well enough for some". They also released a keyboard for Northern Sámi that replaced the letter Ŧ of the standard Sámi keyboard with Y due to it being more common in loan words and names.

SámiSarepta: speech synthesis
NRK reported on 23 November 2013 that the Sámi Parliament of Norway had started work on speech synthesis software for Northern Sámi to be developed by Acapela, Divvun, and Giellatekno. It was tentatively scheduled for release in October 2014, but by 11 February 2015, it was not available for consumers, and a release date was not certain. However, the following day, the Norwegian Directorate for Education and Training released the software, entitled SámiSarepta, coinciding with a seminar at Tromsø University the same day. It was released without the involvement of the Sámi Parliament, whose president, Aili Keskitalo, expressed bewilderment towards the Directorate's decision to launch it. The Sámi Parliament later released the programme for a second time in May.

SámiSarepta provided two voice outputs, female and male, which were named Elle and Biera. It was described by forskning.no as the first text-to-speech software for a minorty language. The programme was designed for assisting in the education of blind, autistic, and other disabled children, but the developers described its purpose as general use, and as such was also usable by others with communication disorders. SámiSarepta was developed as a shell script programme, customisable to the needs of the user. Speech synthesis software for other Sámi languages was also planned, but their corpora were not big enough to facilitate it.

Machine translation
Giellatekno released the translation software Apertium into open beta in May 2012. The programme had been in development since 2010, but received a full release in February 2017. When it was released, it recognised around 97 percent of Northern Sámi words in running text, but could give poor results if typos were present. It only provided translations from Northern Sámi to Norwegian and not the other way around to prevent abuse; Trosterud stated in an interview that he feared the possibility of "ninety percent of all [Sámi] text on the internet being made by our machine translation" due to people publishing poorly translated material without understanding the language. Users found that the software did not always provide grammatically correct Norwegian sentences, but they still considered it to be a good start.

Work with Wikimedia
In October 2020, during Sámi Language Week, Ávvir reported that Giellatekno had started working with the National Library of Norway and Wikimedia Norway to standardise Sámi place names on the internet. The goal was to upload Giellatekno's lists of inflectional paradigms for place names to Wikidata, the National Library's database Språkbanken, and the Norwegian Mapping Authority's Place Name Register for easier access, and to increase the use of local names. In 2022, Wikimedia Norway announced during Sámi Language Week that they would be translating the software language on Wikipedia into Northern Sámi, with practical assistance from Giellatekno. Before then, only 22 percent of it had been translated, with most of the software being in either Norwegian or English. Giellatekno leader Trond Trosterud was awarded "Wikipedian of the Year" by Wikimedia Norway in 2023 for his work on the Norwegian versions and Northern Sami version of Wikipedia.

Other
Giellatekno and Divvun jointly won one of two Gollegiella Awards in 2012 for their work with Sami language proofreading software and dictionaries. This was the second time the University had won the Gollegiella Award, the first being in 2006. The award was worth 120.000 kr, or €12.000. The other award of the year went to Kildin Sámi activists Nina Afanasyeva and Aleksandra Andreevna Antonova. The awards were given out in Oslo at a meeting between the Sámi Parliaments of Norway, Sweden and Finland.

The Sámi Parliament of Norway arranged an information meet at the 18th session of the United Nations Permanent Forum on Indigenous Issues in April 2019 with the University of Tromsø's Divvun and Giellatekno departments. They were invited to present at the event by the United Nations during their International Year of Indigenous Languages. They presented their work on Sámi language technology, and how it could work with other indigenous languages.