Back to LanguageTool Homepage - Privacy - Imprint

A few questions for my master on LanguageTool and Breton


(Novram) #1

Hello everybody,
Firstly I really want to thank you for the great job you're doing!
I'm using LanguageTool every day, and I think it's a really useful software!
My name is Pierre Morvan and I'm now in my second year of master in Breton in Rennes.
I've chosen to study the ability of LanguageTool for correcting the mutation in the Breton language.
It's very fascinating, but I'm not a software programmer.
My work is only to analyze what LT can correct or what LT can't.
I've been working on it for a year seeing what can be done for making LT better than it is now on that point of grammar,
I've actually classified everything in a spreadsheet document.
I have a few questions to ask to you for my memory...
1. Does LT work on mobile phones?
2. Does it work on Microsoft Office?
3. Is LanguageTool one of the grammarchecker for Wikipedia?
4. How can we do for ad words to the word corpus?
Thanks a lot :slight_smile:


(Daniel Naber) #2

Hi Pierre,

thanks for your interest in LT.

Here's a list of integrations: http://wiki.languagetool.org/software-that-supports-languagetool-as-a-plug-in-or-add-on

@Dominique_PELLE will take suggestions for Breton, I think.

Regards
Daniel


(Dominique PELLÉ) #3

Not sure.

Yes, according to
http://wiki.languagetool.org/software-that-supports-languagetool-as-a-plug-in-or-add-on
but I've never tried LT with word myself.

Do you mean a checker used when editing Wikipedia articles?
I don't think that LT is used for that, but I'd love to be proven wrong.
Daniel Naber made a presentation at FOSDEM 2014
about detecting grammar mistakes in Wikipedia a few years ago.
See:

https://archive.fosdem.org/2014/schedule/speaker/daniel_naber/

LT in Breton uses 2 dictionaries as explained here:

  • the Hunspell dictionary "Difazier Hunspell an Drouizig" (v0.13) is for spell
    checking only. It is not used for grammar checkin. LT uses this dictionary
    as-is. It makes no additions. Now if some Breton words are missing in that
    dictionary, we could add them in a text file LT.
  • the dictionary from Apertium (see https://en.wikipedia.org/wiki/Apertium)
    is used for tagging words, i.e. to know whether each word is noun, an adjective
    or a verb (etc), if it is a noun, whether it is masculine/feminine, singular/plural,
    it's mutated form, etc. I added extra tags from what's available in Aperitum,
    in the following script, in order to know whether plural
    nouns are for persons or for things, since it matters for Breton mutations.
    Using LT, I detected a few gender mistakes or missing tags that were
    causing false errors in the Breton LT. I reported them and Fulup Jakez
    quickly fixed them upstream in Apertium, I have not done that in a long time now,
    but the quality of the tagging dictionary is quite good now I would say,
    so false errors caused by the tagging dictionary are rare.
    See this Perl script used to transform the Breton Apertium dictionary
    into a dictionary suitable for LT:

(Michael) #4

The LanguageTool.org website does work on mobile phones, at least on mine (Samsung Galaxy with Chrome Browser)


(Novram) #5

Firstly Thanks a lot for all your answer, it will really help me :slight_smile:
A few others questions that I had thought in it: :wink:

a. How many grammar rules does the German and the English have for the moment in LT?

b. How do you get LT to make corrections when there are homograph issues, for example in English?

c. How does LT work when correcting a text?

d. And how could we make it so that the software would recommand specific prepositions (after specific verbs or nouns)? And how could we add a way to recommend different types of prepositions depending on the meaning? For instance in English, it could offer the choice between "come from" and "come to".

e. Otherwise I was thinking, would it be possible to add the synonyms and antonyms of words?


(Daniel Naber) #6

You can find an overview at https://languagetool.org/languages/

I assume you mean homophones. There are several approaches, you can manually write a rule or use the ngram data.

Please see my link above to the wiki.

Not easily, it would be a new feature. I'm also not sure if that's not a bit out of scope.


(Novram) #7

Hello everyone, how are you doing?!
Firstly could it be possible for me to quote what you had said and put it into my dissertation?

I went on a lot of others grammar checker website as Antidote, Cordial, Ginger, Grammarly, Grammalecte, ... so-and-so, while I was reading some others questions come into my mind, here they are:

f. How can we add expressions and phrases? (I've seen that on Antidote and Cordial)

g. Would there be a way to include a dictionary with terminology, the IPA and the likes in LT? (Antidote and Cordial)

h. Is there a way to include different language registers? (Antidote and Cordial)

i. Is there a way to set a system as can be found in many proof-readers, with the software offering options to complete a word we started writing?

j. How were the different English dialects included in the correcter?

k. Is there a way to have a plug-in allowing going from one English dialect to another? And if it has already been done, would it be applicable to another language such as Breton?
Maybe this might be helpful: https://github.com/SubtitleEdit/plugins/tree/master/AmericanToBritish/DLL

l. Have you ever encountered a correcter that allowed to know the type and function of words when right-clicking on them? (Not seen anywhere that)

m. Is it possible to create a corpus from several journals, newspapers, books and the like in order to give a natural context for words and give actual examples of their use?
(Cordial has done it for French, 5 000 books of classical literature has been incorporated the user text is compared to these text)

n. Is the correcter able to notice when one way of saying something is over-used, and do you think such a system can be developed? (Maybe it's the case now? I don't know :slight_smile: )

o. Is the tatoeba corpus included or LT go on Tatoeba website to check only? Will it bee includ with the LT software that is installed onto one's computer?

p. Where could be found the payable version of LanguageTool? I can't find it any-more.

q. Did you know about SpaCy?

Thanks a lot :slight_smile:


(Novram) #8

Hello,
I would like to know when the pay version of LT has been published?
Thanks a lot? :slight_smile:


(Daniel Naber) #9

https://languagetoolplus.com has been online since 2017-09-20


(Novram) #10

Hello @Dnaber, thanks a lot for your quick answer :slight_smile:


(Novram) #11

Hello folks,
How do you do.
A simple question this time.
Is LT running on Emacs always?
Thanks a lot


(Daniel Naber) #12

Have you checked http://wiki.languagetool.org/software-that-supports-languagetool-as-a-plug-in-or-add-on?


(Novram) #13

Yes I've checked it before asking.
I was asking that because a friend of mine wanted to install it on Eclipse it was impossible for him...
I don't know why... So I wanted to know if it was possible to use LT on Emacs to say to him if he wanted he could install Emacs instead ; and having LT :slight_smile:
Thanks for your quick answer Daniel.
Pierre