#97 Translation of the very important announcements/articles
Opened 5 months ago by amsharma. Modified a month ago

I have seen a comment in the @mattdm's post - https://communityblog.fedoraproject.org/fedoras-strategic-direction-an-update-from-the-council/ , "I think it is a pretty hard read. English is my second language, and I usually don’t have trouble understanding written English, but this piece left me puzzled with what the whole text was saying at times."

And it does make sense that if we have some very important announcements going out, they should be translated in local languages so that we can reach to the wider audience.

There are some online tools available like (https://www.deepl.com/en/translator ) but in that case too a proofread by the native speaker should be done.

Now the question is, how many languages we want to cover and how we can keep up with this effort? Any good suggestions can help.


I would say this should be done by ambassadors (or mindshare) and the i18n team, and we could first limit that for languages where the main website is as at 100% translation (not that we would prevent people from doing in anyway, but I think that translating the website is a more valuable effort than a announcement, so we should push people to that first).

One important question is whether we want to block a announce on a translation, or not, and for how long.

Metadata Update from @amsharma:
- Issue tagged with: help wanted, improvement, type - outreach

5 months ago

I would say this should be done by ambassadors (or mindshare) and the i18n team, and we could first limit that for languages where the main website is as at 100% translation (not that we would prevent people from doing in anyway, but I think that translating the website is a more valuable effort than a announcement, so we should push people to that first).

That is a good suggestions to give it to i18n team but doing it 100% for website ..hmm.. do people have that much bandwidth to accomplish that goal? I think we should take one step at a time, we can start translating the main summary of the important message in the announcement. Something is better than nothing and then check the stats, how many people are actually reading these translated web pages and if the number is significant then putting the efforts for 100% translation makes sense.

One important question is whether we want to block a announce on a translation, or not, and for how long.

Nope, I don't think that we should block anything on translation. We should not stop delivering what we can just because we are not delivering something else.

I feel very strongly that getting the D&I team into the translation business is a step in the wrong direction. We should be including and collaborating with other teams.

This specific content is an update from Matthew may be appropriate to include in the council docs to get it into the translation queue.

I agree with you @bex but how exactly we should make sure that the it is happening, atleast for the most important content.

I totally agree with @bex

@amsharma for example in cases that we see an important announcement that needs to be translated we contact the i18n/l10n teams (post on their ML) and ask for their help to translate the content. We (individually) can help with translations in our own languages, but still, we should collaborate and ask the i18n/l10n teams first - because they already have experience and know how to handle it.

Also, please check this issue we already have open: https://pagure.io/fedora-diversity/issue/33

Either we do it or get it done with other team's help, we need to have a process defined here.

Proposed process:

  1. Determine what content we feel must be multi-lingual
  2. Check with the author of that content to ensure they are ok with it (we may need to commit the content to somewhere else creating a new url/source of truth)
  3. Get the content into a traditionally transmitted place - docs.fp.o is a good start
  4. Give a heads up to the language team
  5. Watch the magic happen

@amsharma can you take point on talking to @jibecfed (or others who are active) on the translation team about this idea? Can you also locate a place on docs.ffp.o for the content you feel must be translated in this specific instance?

@bex thanks for purposing it.

I would like to propose a slight shift in this process -

  1. We have two main platforms of communication - commblog and magazine, whenever writers are pushing their content to either of these two, they should have two questions popped up -
    - Do you want your content to be translated in multi-lingual?
    - Why do you think it should be translated?

  2. Depending on the choice of the author, a mail notification should get triggered to the language team for translation.

  3. There should be "someone" (a team maybe) who should be responsible to read the reason given for translation and make the informed decision about it (translation is really required or not).

Questions -

Who are "we" in the first step of the process you mentioned?
Who are responsible for getting the content into the docs.fp.o ?

I can talk to the specific person you mentioned once we decide on the final process.
For locating the place in docs.ffp.o, I think the content should go to the project to which it belongs to.

@amsharma

I would like to propose a slight shift in this process -

We have two main platforms of communication - commblog and magazine, whenever writers are pushing their content to either of these two, they should have two questions popped up -
- Do you want your content to be translated in multi-lingual?
- Why do you think it should be translated?

Neither of these platforms are, aiui, multi-lingual. How is this translation going to happen? Why would anyone ever NOT want this translated, if translations are something we are doing here.

When you started this issue I thought you wanted specific content translated because it was super-critical, not just all content which I think is outside of hte scope and capacity of hte project.

Depending on the choice of the author, a mail notification should get triggered to the language team for translation.

There should be "someone" (a team maybe) who should be responsible to read the reason given for translation and make the informed decision about it (translation is really required or not).

Questions -
Who are "we" in the first step of the process you mentioned?

The D&I team because you are proposing this should be part of this teams function.

Who are responsible for getting the content into the docs.fp.o ?

The D&I team because you are proposing this should be part of this teams function.

I can talk to the specific person you mentioned once we decide on the final process.
For locating the place in docs.ffp.o, I think the content should go to the project to which it belongs to.

agreed - but this team will need to submit the PR if they identify the content.

ok, we need to define clearly what kind of content we want to translate and how we are deciding to choose the one to go through the process.
Should that be done using voting process within the diversity team and have a ticket opened in diversity repo every-time we identify an article or content to be translated OR that needs to have an agreement with translation team as well?

As discussed in meeting today -
https://docs.google.com/document/d/1FgLp-m-Akts1_d4oShJtbZe23X60PqUnvA8rGdp7uSs/edit?usp=sharing

We will follow this process to start with to identify an article to be translated -
Using DI repo, open a translation ticket
Having tag/label for ticket
Lazy approval should be enough

I will update the final process and checkin to our doc site and start with our first translation ticket soon.

+1 to the process discussed in the meeting.

I am nitpicking here but @amsharma when you update the docs site, can you please add that the new ticket MUST have the following details :

  1. Link to the article to be translated

  2. Languages to be translated to

  3. Has the author of the content already agreed to the translation? (YES/ NO) with relevant link to discussion/ tag the author.
    This MUST be YES for document to be approved for translation.

  4. Reason why it is important to translate this article (a few words are fine too)

Thanks @bee2502 . This is the process defined here - https://docs.google.com/document/d/1_5aB1trQ-ZGT2h4Qgmw2x7W7Y5IOOaUAhdbS_Y8N17Q/edit?usp=sharing

@jibecfed It will be great if you can please check this process from the translation team prospective and share your opinion. Thanks.

Mail sent to trans@lists.fedoraproject.org for process review.

Metadata Update from @amsharma:
- Issue assigned to amsharma

4 months ago

Metadata Update from @amsharma:
- Issue priority set to: needs review (was: awaiting triage)

4 months ago

Hello, thank you for this will to have multi-lingual content (and communicating with us through #fedora-i18n and trans mailing list)! I personally think translation are important for technology inclusion.

I'm a little bit confuse about what publication system we are talking about.
If it is documentation, it will be include in the doc translation system, and you won't have to do anything to allow it :)

If it is article publication on community blog or community magazine, I suggest to change the way to answer the issue:
everything should be translatable by any team,
diversity group decide the priorities or release criteria and communicate it to translations teams.

In any way, there is no reasons to ask an author about whether should translate or not the content as it falls under the Fedora contributor agreement, all we do and publish is under an open license.

And there is no need to justify why you want an article to be translated in one language. You may just say: we would love to see in translate at least in lang1, lang2, lang3, and if the teams can do it, they will. If other teams also can, it's good also.

To make everything translatable by any team means changing our tools, and I suggest to have a look at hugo: https://jibecfed.fedorapeople.org/blog-hugo/en/2018/11/building-a-multilingual-static-website-with-hugo/

But changing our tooling is a big work. By then, just tell us where to drop the files (opening an issue with a translated file will be fine) and what's the process.

Thanks for your response @jibecfed . This is really helpful.
So...it means if we get the content (which we want to see in different languages), we just need to get that content in the Fedora doc site and it will be translated by doc translation system without doing anything extra? If that is the case then the things are pretty simple from now on.. Please confirm. Thanks.

Sorry for this late arrival, I was not aware of this thread. Here is some minor feedback from me, a foreigner who sometimes suffers from the problems described here.

  1. Please avoid automatic translation tools. While I know there are some pairs of languages where they work fine I also know pairs where it works poorly and even cases where the automatic translations are totally unusable.
  2. A person quoted in the original post did not say that he can't understand the commented article because he could not speak English. He emphasized he usually understood English but that particular article was too difficult. (While at this: if it was too difficult for a human it would be totally impossible to process by an automatic tool.) I think this message should be said to the authors: please make sure that you don't use overly complex English phrases. I believe it is easy to foreigners to distinguish which phrases are too complex and which are not. For the native English speakers, ask yourself these questions: Can you speak any foreign language? Are you able to translate your article to that foreign language? Would you find it difficult? Would you be able to understand a similar article written in that foreign language?
  3. Proofreading of English texts written by foreigners has been mentioned here as well. I think it is a separate issue but important as well.

I am sorry for my poor English and thank you @amsharma for raising the subject.

@rluzynski Thanks for your valuable insight and I strongly agree with your second point.

I would like to explore the ideas where we can incorporate these guidelines for the writers.

Flock summary - 2019-08-10

To summarize the discussion we had at the D&I hackfest at Flock with @jibecfed and @rluzynski, here are the action items we walked away with:

  1. Better collaboration with the i18n/l10n mailing lists since this is where their community is
  2. Better engagement with the Council on how to identify important announcements and put them in a Pagure repo for the i18n/l10n teams to translate with the existing workflows

I will try to add more context here soon, wanted to get this in while it was fresh.

Login to comment on this ticket.

Metadata