Translating MediaWiki into Three Saami Languages

Kimberli Mäkäräinen

Wikimedia Finland

Wikimedia Norway

Anarâškielâ servi

AvoinGLAM

The 9 living Saami languages

1. Southern Saami

2. Ume Saami

3. Pite Saami

4. Lule Saami

5. Northern Saami

6. Skolt Saami

7. Inari Saami

8. Kildin Saami

9. Ter Saami

Source: https://commons.wikimedia.org/wiki/File:Corrected_sami_map_4.PNG

5. Sámegiella— Northern Saami (se)

• By far the most widely spoken Saami language
• Northern Saami people
• many non-Saami learning the language
• many Saami who speak one of the other languages also learn Northern Saami
• Spread across 3 countries:
• Norway
• Sweden
• Finland
• In all 3, it is used on government websites to some degree
• If Saami is one of the languages that can be selected on websites, it is commonly the only Saami language available
• this makes it easier to find equivalents when translating MediaWiki
• caveat: it is often only partially equivalent to what is offered in other languages
• The user interface on multiple non-government sites has also been �(relatively recently) translated into Northern Saami

5. Sámegiella — Northern Saami (se)

• Wikimedia Norway received funding to translate
• MediaWiki (NUUG)
• the Wikipedia apps for Android and iOS (the Saami Parliament of Norway)
• The National Library of Finland supported the translation work by sharing its translation memory with us
• Multiple translators from different dialects and 2 countries worked on the translations
• some differences in opinion on terminology
• settings: heivehus vs. ásahus (heivehus means “app” on the Finnish side of the border)
• Word order and choices were influenced by the majority language(s) �in these countries to some degree
• Older and dialectal words were repurposed for the translations: gilkor

​

​

​

6. Sääʹmǩiõll — Skolt Saami (sms)

• Approximately 350 speakers
• Majority of the speakers are Skolt Saami
• Recent development: people who are not Skolt Saami learning the language
• Large generational gap due to forced assimilation
• Mainly spoken within Finland’s borders
• Estimates of 10–12 speakers in Norway
• Estimates of 15–20 speakers in Russia
• Currently used in minute amounts on government websites in Finland
• With the exception of a few recently updated websites, usually hidden behind Northern Saami (aka Skolt speakers need to know Northern Saami to find stuff in their own language!)

6. Sääʹmǩiõll — Skolt Saami (sms)

• No translation project
• No funding
• 1 translator
• (fortunately) not too many differences in opinion on terminology because of it
• Nearly impossible to find terminology for MediaWiki translations because it is not commonly used in the UIs Northern Saami appears in
• 2 websites have their UI in Skolt Saami
• lack of modern vocabulary in reference material
• lack of quick linguistic support
• Special characters can cause multiple issues when it is finally included somewhere
• few fonts work with the language (and Montserrat is not one of them).
• “To/For the Finnish National Agency for Education” in Skolt Saami:

Mättʼtõshalltõʹsse

​

7. Anarâškielâ — Inari Saami (smn)

• Approximately 450 speakers
• Majority of the speakers are Inari Saami
• Recent development: people who are not Inari Saami learning the language
• Only spoken within Finland’s borders*
• Currently used to some degree on government websites in Finland
• With the exception of a few recently updated websites, usually hidden behind Northern Saami
• Used more than Skolt Saami, but less than Northern Saami
• Has its own Wikipedia, so the translations spread through that and other platforms that use MediaWiki
• Samsung’s UI has extremely limited support for Inari Saami (and Northern Saami, but not any of the other Saami languages)

​

* Also spoken by people living outside of Finland nowadays, for example Inari Saami who have moved abroad

7. Anarâškielâ — Inari Saami (smn)

• No translation project
• Applied for funding; did not get any

​

• Multiple translators,
• both native and non-native
• Not easy to find terminology for MediaWiki translations, but relatively easy for the translators to find linguistic support when unknown terms pop up
• the language community has an established group on Facebook where people often ask about unknown words, terms, phrases, etc.
• Not commonly used in the UIs Northern Saami appears in, but it appears on more sites than Skolt Saami does

Issues, suggestions, and stats

Issues with the source text

• Ambiguous strings in English, e.g.
• When are edit, display, or search nouns and when are they verbs?
• If verbs, what do you mean by it? To edit, (they) edit, or edit!
• Single-word strings ending in -ing like saving, checking: noun or verb?
• Editors: people or programs?
• block registered users: block as in a command or is this part of another string and it is the infinitive form (you do not have permission to block registered users)?
• Change: change what? Passwords? Languages? Something else?
• Depending on what you’re changing, the verb can be different. Don’t make translators guess.
• Great variation in the spelling of words at times; translation memories do not always recognize them
• Non-standard English, not so much in terms of vocabulary, more so with the grammar

Suggestions to make translators’ lives easier (and improve the linguistic quality of your software)

• Variables
• Be clear what is behind it: $1 page categorization
• What is going to be coming in the place of $1? A verb, a noun?
• For languages that use grammatical cases like the Saami languages, the cases might not even be the same depending on the word or phrase used in $1:
• $1 = Show (verb) -> page categorization = Čájet + siidu(id) kategoriserema (accusative case)
• $1 = Like (verb) -> page categorization = Liiko + siidu(id) kategoriseremii (illative case)
• $1 = Fear (verb) -> page categorization = Bala + siidu(id) kategoriseremis (locative case)
• Make sure your translators know how your variables are going to be put to work

​

Suggestions to make translators’ lives easier (and improve the linguistic quality of your software)

• Don’t assume other languages work like English (or Spanish, German, French, etc.)
• We have different word orders, your nouns can be our verbs, etc.
• Prepositions
• Not all languages have prepositions and the ones that do do not necessarily have a 1-to-1 equivalent.
• “Of” has to be one of the most annoying strings, especially when it is on its own.
• Gender
• Not all languages have gender; don’t force it on languages that don’t.
• Yes/No
• Seems simple, no? But no, it’s not. For some languages, you need to know what it is referring to since yes is indicated by repeating the verb in the question and no by negating that verb, so there is not one standard way of saying yes or no.

​

​

​

Suggestions to make translators’ lives easier (and improve the linguistic quality of your software)

• Number
• The English/Spanish/German/French 1 -> singular, 2 and everything else -> plural is not the only system.
• Translators need to know things like:
• What the possible range of numbers will be
• 1 result, 2 results, 8 results, 1000 results
• How they and any words attached to them will be used:
• 1 result vs. 6 results -or- Show 1 result vs. Show 6 results
• 0 is not always easy to translate using a number
• It could be more fluent in some languages to say “not even a single something or another” instead of “0 something or another”:

​

​

Suggestions to make translators’ lives easier (and improve the linguistic quality of your software)

• Terminology!
• An up-to-date glossary of the concepts used in your software is a must.
• Terms and definitions for these concepts? Add them!
• Terms you prefer to be used (or not to be used)? Say so!
• You use the terms differently than (nearly) everyone else? Tell the translator(s)!
• 10/10 translators would recommend Jon Harald’s terminology gadget in TranslateWiki.
• Translation memory/memories
• Abbreviations
• do not assume translators will know what they mean
• Notes
• add enough notes that the translators do not have to constantly be wondering what the string is referring to.
• If possible, add screenshots to the notes.
• Keep both notes and screenshots up-to-date, please!

​

Usage elsewhere

• MediaWiki translations are being reused to translate equivalent strings in other FOSS projects such as:
• KIWIX
• OpenStreetMap
• In addition, the translated UI strings are also available on non-Wikimedia sites that are powered by MediaWiki such as
• The Helsinki Term Bank for the Arts and Sciences (HTB) (https://tieteentermipankki.fi/)
• Lokalhistoriewiki (https://lokalhistoriewiki.no/wiki/The_Local_History_Wiki)
• Nordterm (http://www.nordterm.net/)

​

​

Späʹsseb! Takkâ! Giitu!

Vääʹld õhttvuõđ:

Kimberli Mäkäräinen

User:Yupik

Language-specific translation statistics: https://codelookup.toolforge.org

​

CC BY-SA 4.0