Delphine Lebédel [:delphine - Off until Sept 19] Reporter
	Description • 10 years ago

Please create Ewe Dictionary and Worlist on FxOS

	Delphine Lebédel [:delphine - Off until Sept 19] Reporter
	Comment 1 • 10 years ago

[Blocking Requested - why for this release]:
Meeting with Bus Dev this morning confirmed that this would be needed in 2.0 and onwards
thanks

	Wesly Huang (EPM)
	Comment 2 • 10 years ago

[Triage]

BD has confirmed the need, recorded in Program Management mana page.

	Francesco Lodolo [:flod]
	Comment 3 • 10 years ago

Gaia::Keyboard is the right place for this

	Delphine Lebédel [:delphine - Off until Sept 19] Reporter
	Comment 4 • 10 years ago

Do we know more about for when this is needed exactly?
No one is assigned or working on this. Just want to make this clear if this is urgent....

	Delphine Lebédel [:delphine - Off until Sept 19] Reporter
	Comment 5 • 10 years ago

Hey Kevin (sorry, I'm pinging you a lot): we do not have an identified community for Ewe. Do you think you could help with this please? thanks

	Wesly Huang (EPM)
	Comment 6 • 10 years ago

Hi Delphine:

We need information from BD for this, currently partner's R&D team is working on other region's SW so I assume this some early information via business discussion. You can check bug 1129845 (Mozilla only) for the regions partner is working on currently.

@Tony and Nisha: 

Would you help provide the information? thank you.

	Kevin Scannell
	Comment 7 • 10 years ago

I can probably do something for Ewe, although being able to consult a native speaker would help a lot. I'll wait to proceed until there's confirmation that this is really needed.

	Delphine Lebédel [:delphine - Off until Sept 19] Reporter
	Comment 8 • 10 years ago

Great, thanks a bunch Kevin.
Maybe someone from Rubric team is also an Ewe native speaker? :ni on Peiying to see if she can plug in someone to collaborate with on this!

	Peiying Mo [:CocoMo]
	Comment 9 • 10 years ago

+ Devon and Ian from Rubric.

Hi Ian and Devon, 

As mentioned in email before, we need help with dictionaries in some African languages.  Kevin is our key contributor in the area. If bug is not your preferred way to discuss this, we can start email threads to see how we go about with this.

Thank you,
Peiying

	Devon
	Comment 10 • 10 years ago

Hi Peiying,

Thank you for adding us to the circulation list.

Kevin,

Can you provide more info as to what exactly you need.  I have attached the Ewe glossary that we have already created with main key terms for the FXOS component.

Thank you,
Devon

	Kevin Scannell
	Comment 11 • 10 years ago

I can generate an Ewe frequency list from texts on the web, but this approach can be problematic in a lot of different ways (texts with missing diacritics, variant orthographies, dialects, ambient French/English boilerplate text...)  For the other autocorrect dictionaries I've created, I was able to rely on an existing spellchecker to help deal with these issues; there's no such spell checker for Ewe.   

I could use help from a speaker of the language to (1) decide which texts in my corpus are ok to use and (2) help "clean up" the resulting freq list - probably on the order of 10,000 words based on the amount of text I see now on the web, maybe 20 hours of work for someone with a good knowledge of the orthography.

	Devon
	Comment 12 • 10 years ago

Hi Kevin,

I will consult our Ewe linguists to confirm if there are online dictionaries available to start with and will provide you with feedback ASAP.

Thanks you,
Devon

	Devon
	Comment 13 • 10 years ago

Hi Kevin,

Our Ewe linguists have confirmed that there aren't any good online dictionaries available.  In the meantime I have contacted this https://glosbe.com/ee/en Polish community site to confirm whether it would be possible to provide an extract of the content of their ee->en dictionary in CSV format.

I'll let you know as soon as I receive a response.

Thanks, 
Devon

	Kevin Scannell
	Comment 14 • 10 years ago

Thanks Devon, getting permission to use that dictionary would be a big help, and would make the checking of the web-crawled word list much easier/faster.

	Josh Cheng [:josh]
	Comment 15 • 10 years ago

[Blocking Requested - why for this release]:
Per comment 6 denominate this from 2.0+

	aappleton@mozilla.com
	Comment 16 • 10 years ago

In response to the needinfo.....

Personally, I am not aware of a specific requirement for a launch in Ghana which might necessitate the need for Ewe. We are talking to operators that have a footprint in Ghana, but there is no launch currently in discussion.

Redirecting the question to Nisha, as she is BD responsible for these Operators and will have a better idea of when a launch might happen.

	Wesly Huang (EPM)
	Comment 17 • 10 years ago

[Triage]

For partner working with FxOS 2.0, TAM has communicated to them to handle languages other then what are already supported in FxOS 2.0 by their own, so this will not block 2.0.

De-mon it from 2.0.

	Peiying Mo [:CocoMo]
	Comment 18 • 10 years ago

Kevin, our l10n agency also suggests to check Pootle's current localized strings from FFOS 2.0 to help generate this list.  Will it help?

	Kevin Scannell
	Comment 19 • 10 years ago

Yes, every bit helps, especially well-edited text that uses the correct orthography.  But unless we can source a good dictionary/word list (see Devon's comment 13), we really need to get someone to edit (or at least sanity-check) the frequency list I generate from the web.

	Delphine Lebédel [:delphine - Off until Sept 19] Reporter
	Comment 20 • 10 years ago

Kevin: I think Devon/Rubric can help review the frequency list you generate from the web. Please let us know if you need anything more to move forwards with this :)
thanks!

	ian.henderson
	Comment 21 • 10 years ago

(In reply to Kevin Scannell from comment #19)
> Yes, every bit helps, especially well-edited text that uses the correct
> orthography.  But unless we can source a good dictionary/word list (see
> Devon's comment 13), we really need to get someone to edit (or at least
> sanity-check) the frequency list I generate from the web.

Let's try the same approach as my Comment 12 on Bug 1121730

	Kevin Scannell
	Comment 22 • 10 years ago

I'm gathering a bunch of Ewe web text as we speak.  Can you email me the PO files please? I'll combine them with what I have so far and start generating word lists for editing. Thanks.

	ian.henderson
	Comment 23 • 10 years ago

(In reply to Kevin Scannell from comment #22)
> I'm gathering a bunch of Ewe web text as we speak.  Can you email me the PO
> files please? I'll combine them with what I have so far and start generating
> word lists for editing. Thanks.

Do you have access to this location? http://mozilla.locamotion.org/ee/firefoxos_2.0/

	Kevin Scannell
	Comment 24 • 10 years ago

I'm not able to export a zip of all of the PO files, which is what I need to do what you're suggesting.

	ian.henderson
	Comment 26 • 10 years ago

(In reply to Kevin Scannell from comment #24)
> I'm not able to export a zip of all of the PO files, which is what I need to
> do what you're suggesting.

No problem. I have uploaded the FFOS zip for you.

	Kevin Scannell
	Comment 27 • 10 years ago

Thanks. I combined the translations with a couple million words of Ewe from the web (mostly religious material).  I'll attach a file with an initial word list for editing.  This contains about 2500 pairs of words that differ only in their use of special characters, diacritics, and so on.  So for example you'll see the pair:

boboe 2
bɔbɔe 1887

I don't know any Ewe but I'd guess that the second spelling with the open o's is the correct one, and that "boboe" is simply a rendering of this word on an ASCII keyboard.  The numbers are the frequency counts - these often give good hints as to what the preferred spelling is.

What would be ideal is if some speakers could go through this list and simply delete the "incorrect" words.  I'll use the result (1) as the beginning of a clean word list for autocorrect (2) to improve the statistical filters I use to remove bad texts, giving better frequency counts.

	Delphine Lebédel [:delphine - Off until Sept 19] Reporter
	Comment 29 • 10 years ago

Not sure what the steps are going forwards anymore for this :)
ni on Bruce so he can advise on how to get this implemented. thanks!

	Dwayne Bailey
	Comment 30 • 10 years ago

(In reply to Kevin Scannell from comment #24)
> I'm not able to export a zip of all of the PO files, which is what I need to
> do what you're suggesting.

Kevin if you can give a Pootle login I can give you the needed rights that should allow you to export any of the POs across the site.

	Kevin Scannell
	Comment 31 • 10 years ago

Thanks Dwayne - I'm "scannell" on http://mozilla.locamotion.org/

	Dwayne Bailey
	Comment 32 • 10 years ago

(In reply to Kevin Scannell from comment #31)
> Thanks Dwayne - I'm "scannell" on http://mozilla.locamotion.org/

Enabled.  Looks like you had those rights anyway, but now its explicit across the whole site.

	ian.henderson
	Comment 35 • 9 years ago

Hello Kevin,

Some more content ready:

http://mozilla.locamotion.org/ee/mozilla_lang/main.lang.po
http://mozilla.locamotion.org/ee/mozilla_lang/mozorg/home/index.lang.po
http://mozilla.locamotion.org/ee/mozilla_lang/firefox/os/index.lang.po
http://mozilla.locamotion.org/ee/mozilla_lang/firefox/os/devices.lang.po
http://mozilla.locamotion.org/ee/mozilla_lang/firefox/os/faq.lang.po
http://mozilla.locamotion.org/ee/mozilla_lang/firefox/partners/index.lang.po
http://mozilla.locamotion.org/ee/mozilla_lang/firefoxos/firefoxos.lang.po
http://mozilla.locamotion.org/ee/mozilla_lang/legal/index.lang.po
http://mozilla.locamotion.org/ee/mozilla_lang/tabzilla/tabzilla.lang.po

https://localize.mozilla.org/ee/masterfirefoxos/