Localization discussions

[Feedback needed] Updating review guidelines for localized articles

  • 12 Replies
  • Nuɖoɖo mlɔetɔ Kiki Off (Back Jun 30th)
  1. Kiki Off (Back Jun 30th)
    Kiki Off (Back Jun 30th) Post 1586
    more options

    Hi folks,

    There has been a conversation about review guidelines for localized articles recently on Matrix.

    Curently, we don't require cross-approval for non-English articles (meaning you're ok to self-approve in whatever situation). However, from the conversations we have on Matrix, some locale groups have been doing cross-review pretty pretty religously. And most of them agree that it's a good practice that should be standardized to provide quality content.

    It seems that there's an appetite to update our review guidelines for non-English articles to match what we're doing with the the en-US articles (see Article review and approval guidelines), which means forbid self-approval. However, we may still need to provide some leeway for locales that only have 1 active reviewer to do self-approval.

    Please chime in with what you think of this proposal in this thread below.

    Hi folks, There has been a conversation about review guidelines for localized articles recently on [https://matrix.to/#/!LxGoPWpREgbnMiZrmI:mozilla.org/$6EmcE32zfwG4cPgy8pPh_XDkHnmhyH7JbFEXBRUvL-U?via=mozilla.org&via=matrix.org&via=t2bot.io Matrix]. Curently, we don't require cross-approval for non-English articles (meaning you're ok to self-approve in whatever situation). However, from the conversations we have on Matrix, some locale groups have been doing cross-review pretty pretty religously. And most of them agree that it's a good practice that should be standardized to provide quality content. It seems that there's an appetite to update our review guidelines for non-English articles to match what we're doing with the the en-US articles (see [[Article review and approval guidelines]]), which means forbid self-approval. However, we may still need to provide some leeway for locales that only have 1 active reviewer to do self-approval. Please chime in with what you think of this proposal in this thread below.
  2. Mad Maks
    Mad Maks Post 37
    more options

    Thanks for starting this discussion.

    in my point of view: - best practice: do not self-approve, always have a second person proof read your translation. As good as you are you can always make a type error or translate some content in a other way then in the used translation in the product - if it is a minor change, like edit a link or remove a section to get in sync with the en_US version it is ok to do a self-approve

    I understand there are locales who has only 1 person doing the translation, and the need to do a self-review. This is not a good practice, a) if that person stops we get a lot of outdated content b)that person will get overloaded c) there will be mistakes in the translations, even the best translators make mistakes

    Remember it is important to have a good quality not only on the en-US articles but also on the localized content it is all Mozilla!

    Thanks for starting this discussion. in my point of view: - best practice: do '''not''' self-approve, always have a second person proof read your translation. As good as you are you can always make a type error or translate some content in a other way then in the used translation in the product - if it is a minor change, like edit a link or remove a section to get in sync with the en_US version it is ok to do a self-approve I understand there are locales who has only 1 person doing the translation, and the need to do a self-review. This is not a good practice, a) if that person stops we get a lot of outdated content b)that person will get overloaded c) there will be mistakes in the translations, even the best translators make mistakes Remember it is important to have a good quality not only on the en-US articles but also on the localized content it is '''all''' Mozilla!
  3. wxie
    wxie Post 58
    more options

    It's good to have a cross-reviewer, but that's a luxury for many locales. Plus, some articles are just updated with minor changes like deleting a few spaces. If all these require a cross-reviewer and cannot be self-approved, I foresee that there will be much more obsolete articles which is worse than current situation for many languages.

    It's good to have a cross-reviewer, but that's a luxury for many locales. Plus, some articles are just updated with minor changes like deleting a few spaces. If all these require a cross-reviewer and cannot be self-approved, I foresee that there will be much more obsolete articles which is worse than current situation for many languages.
  4. Bogomil Shopov - Бого
    Bogomil Shopov - Бого Post 5
    more options

    To disable the self-approve we need at least 3 active contributors for each locale. This is the challenge.

    I agree that self-approving can bring challenges already identified by @Mad Maks.

    In some cases we need to chose between self-approve with the self validation that we act as good owners of the content and not providing localized content to our consumers, which is much works that a few minor mistakes.

    To disable the self-approve we need at least 3 active contributors for each locale. This is the challenge. I agree that self-approving can bring challenges already identified by @Mad Maks. In some cases we need to chose between self-approve with the self validation that we act as good owners of the content and not providing localized content to our consumers, which is much works that a few minor mistakes.
  5. Merike
    Merike Post 14
    more options

    There's another way to get something similar to another set of eyes and refreshed attention though it's admittedly not as good as a second person. That is, require a time delay between reviewing your own changes. This gives new perspective if you do it a day or two later and truly read it over again. I've found issues reading translation changes from a feed that Estonian has on Google Groups. Feed sync takes time so delay happens naturally. Some mistakes have been my own. I haven't been able to contribute much in past few years but there would be no reviewers available beside myself for Estonian.

    There's another way to get something similar to another set of eyes and refreshed attention though it's admittedly not as good as a second person. That is, require a time delay between reviewing your own changes. This gives new perspective if you do it a day or two later and truly read it over again. I've found issues reading translation changes from a feed that Estonian has on Google Groups. Feed sync takes time so delay happens naturally. Some mistakes have been my own. I haven't been able to contribute much in past few years but there would be no reviewers available beside myself for Estonian.
  6. Kiki Off (Back Jun 30th)
    Kiki Off (Back Jun 30th) Post 1586
    more options

    Mad Maks said

    Remember it is important to have a good quality not only on the en-US articles but also on the localized content it is all Mozilla!

    Yep. Totally aggree with most of what you've said. Ideally, we should strive to cross-review all revisions (with the exception of minor edit), not only in English but also the localized versions as well.

    Ideally, for locales with only 1-2 active people, we should empower them to attract more people to participate. However, with the limited resource that we have at the moment, this is a lot harder in practice. So for now, we need to be more tolerant with other locales with less active people. As Wxie mention, it can be a luxury for some locales to do cross-review.

    ''Mad Maks [[#post-86850|said]]'' <blockquote> Remember it is important to have a good quality not only on the en-US articles but also on the localized content it is '''all''' Mozilla! </blockquote> Yep. Totally aggree with most of what you've said. Ideally, we should strive to cross-review all revisions (with the exception of minor edit), not only in English but also the localized versions as well. Ideally, for locales with only 1-2 active people, we should empower them to attract more people to participate. However, with the limited resource that we have at the moment, this is a lot harder in practice. So for now, we need to be more tolerant with other locales with less active people. As Wxie mention, it can be a luxury for some locales to do cross-review.
  7. Kiki Off (Back Jun 30th)
    Kiki Off (Back Jun 30th) Post 1586
    more options

    wxie said

    Plus, some articles are just updated with minor changes like deleting a few spaces. If all these require a cross-reviewer and cannot be self-approved, I foresee that there will be much more obsolete articles which is worse than current situation for many languages.

    I agree that we may not need cross-review for minor edits. The challenge here would be to quantify what do we mean with minor edit? Also, I think there need to be a limit for how many self-approve articles one can do at one time. I don't want people to exploit this leeway to create lots of minor edits to inflate their reputation in the community.

    ''wxie [[#post-86851|said]]'' <blockquote> Plus, some articles are just updated with minor changes like deleting a few spaces. If all these require a cross-reviewer and cannot be self-approved, I foresee that there will be much more obsolete articles which is worse than current situation for many languages. </blockquote> I agree that we may not need cross-review for minor edits. The challenge here would be to quantify what do we mean with minor edit? Also, I think there need to be a limit for how many self-approve articles one can do at one time. I don't want people to exploit this leeway to create lots of minor edits to inflate their reputation in the community.
  8. Kiki Off (Back Jun 30th)
    Kiki Off (Back Jun 30th) Post 1586
    more options

    Bogomil Shopov - Бого said

    To disable the self-approve we need at least 3 active contributors for each locale. This is the challenge.

    Excatly!

    I agree that self-approving can bring challenges already identified by @Mad Maks. In some cases we need to chose between self-approve with the self validation that we act as good owners of the content and not providing localized content to our consumers, which is much works that a few minor mistakes.

    Agree. A few minor mistakes is better than not having a localized article.

    I think another challenge if we allow self-approval like we do right now is the consentration of power on the only active reviewer. I know that in some locales, we don't have issues with new contributors. So allowing self-approve may discount the incentive to review work from the other contributors that can be a potential active members. Encouraging cross-review will force the reviewers to work with other people, which can also foster more collaboration within our community.

    ''Bogomil Shopov - Бого [[#post-86852|said]]'' <blockquote> To disable the self-approve we need at least 3 active contributors for each locale. This is the challenge. </blockquote> Excatly! <blockquote> I agree that self-approving can bring challenges already identified by @Mad Maks. In some cases we need to chose between self-approve with the self validation that we act as good owners of the content and not providing localized content to our consumers, which is much works that a few minor mistakes. </blockquote> Agree. A few minor mistakes is better than not having a localized article. I think another challenge if we allow self-approval like we do right now is the consentration of power on the only active reviewer. I know that in some locales, we don't have issues with new contributors. So allowing self-approve may discount the incentive to review work from the other contributors that can be a potential active members. Encouraging cross-review will force the reviewers to work with other people, which can also foster more collaboration within our community.
  9. Kiki Off (Back Jun 30th)
    Kiki Off (Back Jun 30th) Post 1586
    more options

    Merike said

    There's another way to get something similar to another set of eyes and refreshed attention though it's admittedly not as good as a second person. That is, require a time delay between reviewing your own changes. This gives new perspective if you do it a day or two later and truly read it over again. I've found issues reading translation changes from a feed that Estonian has on Google Groups. Feed sync takes time so delay happens naturally. Some mistakes have been my own.

    That's a fantastic idea, Merike! I can see this as a potential workaround for many locales with only 1-2 active reviewers. Thanks for sharing that idea.

    I haven't been able to contribute much in past few years but there would be no reviewers available beside myself for Estonian.

    I see that you're contributing to Pontoon too with Sander. Do you thin Sander will be interested in helping out in SUMO as well? Otherwise, is there any local language enthusiast groups that we can approach to see if they can help in some ways?

    ''Merike [[#post-86853|said]]'' <blockquote> There's another way to get something similar to another set of eyes and refreshed attention though it's admittedly not as good as a second person. That is, require a time delay between reviewing your own changes. This gives new perspective if you do it a day or two later and truly read it over again. I've found issues reading translation changes from a feed that Estonian has on Google Groups. Feed sync takes time so delay happens naturally. Some mistakes have been my own. </blockquote> That's a fantastic idea, Merike! I can see this as a potential workaround for many locales with only 1-2 active reviewers. Thanks for sharing that idea. <blockquote> I haven't been able to contribute much in past few years but there would be no reviewers available beside myself for Estonian. </blockquote> I see that you're contributing to Pontoon too with Sander. Do you thin Sander will be interested in helping out in SUMO as well? Otherwise, is there any local language enthusiast groups that we can approach to see if they can help in some ways?
  10. Bogomil Shopov - Бого
    Bogomil Shopov - Бого Post 5
    more options

    Where did we landed on this?

    Where did we landed on this?
  11. Kiki Off (Back Jun 30th)
    Kiki Off (Back Jun 30th) Post 1586
    more options

    Bogomil Shopov - Бого said

    Where did we landed on this?

    We haven't agree on anything at this point. Also, I want to make sure that we hear from the people who are most impacted by this change. So I doubt that we'll make a decision around this very soon.

    ''Bogomil Shopov - Бого [[#post-86933|said]]'' <blockquote> Where did we landed on this? </blockquote> We haven't agree on anything at this point. Also, I want to make sure that we hear from the people who are most impacted by this change. So I doubt that we'll make a decision around this very soon.
  12. ErlingR
    ErlingR Post 55
    more options

    Sorry for engaging late in this thread! It is an important subject that has been raised.

    For my locale (Norwegian) there are few active contributors: 2 leaders, 2 reviewers and 4 active contributors (last 90 days). The two leaders/reviewers are the same persons, and I am one of them. I was recruited by the other person nearly three years ago, and since then I think I have done most of the initial translation / revision work for Norwegian KB articles myself, and approved the translations. The only other translations that I have been asked to review, have turned either to be misunderstandings about what should be done when translating an article (often the English text has been promoted as the Norwegian article), or trolls.... I have only given positive feedback to those translators, for taking part in the Mozilla localization work, and offered advice on how to proceed correctly, but have never gotten any replies to that.

    I had a brief discussion with the person that promoted me to reviewer status about auto-approval of translations, and he stated that at that time (March/April 2021), it was only me and him that did translation work. So we saw no other way to allow me to approve my own article translations.

    Of course I see the pitfalls of auto-approval; it is against all quality assurance principles I have come across . . in an ideal world with plentiful of resources to engage in the KB translation work. In the real world I would argue that auto-approval is the lesser of two evils. In the real world one has to choose: Articles with possible errors vs. none or outdated localized articles. And the latter ones obviously have erroneous content.

    Anyway, I will watch this thread to see where we land.

    Sorry for engaging late in this thread! It is an important subject that has been raised. For my locale (Norwegian) there are few active contributors: 2 leaders, 2 reviewers and 4 active contributors (last 90 days). The two leaders/reviewers are the same persons, and I am one of them. I was recruited by the other person nearly three years ago, and since then I think I have done most of the initial translation / revision work for Norwegian KB articles myself, and approved the translations. The only other translations that I have been asked to review, have turned either to be misunderstandings about what should be done when translating an article (often the English text has been promoted as the Norwegian article), or ''trolls''.... I have only given positive feedback to those translators, for taking part in the Mozilla localization work, and offered advice on how to proceed correctly, but have never gotten any replies to that. I had a brief discussion with the person that promoted me to reviewer status about auto-approval of translations, and he stated that at that time (March/April 2021), it was only me and him that did translation work. So we saw no other way to allow me to approve my own article translations. Of course I see the pitfalls of auto-approval; it is against all quality assurance principles I have come across . . '''''in an ideal world with plentiful of resources to engage in the KB translation work'''''. In the real world I would argue that auto-approval is the lesser of two evils. In the real world one has to choose: Articles with possible errors vs. none or outdated localized articles. And the latter ones obviously have erroneous content. Anyway, I will watch this thread to see where we land.
  13. Kiki Off (Back Jun 30th)
    Kiki Off (Back Jun 30th) Post 1586
    more options

    ErlingR said

    Of course I see the pitfalls of auto-approval; it is against all quality assurance principles I have come across . . in an ideal world with plentiful of resources to engage in the KB translation work. In the real world I would argue that auto-approval is the lesser of two evils. In the real world one has to choose: Articles with possible errors vs. none or outdated localized articles. And the latter ones obviously have erroneous content. Anyway, I will watch this thread to see where we land.

    Thanks for sharing your experience, Erling. I can totally understand the struggle with not having enough people to help with cross-review. But I appreciate your understanding of the importance of that process.

    At the moment, we haven't land on any sort of decision yet. This discussion is started because I wanted to hear first from our localizers what they think about this idea.

    ''ErlingR [[#post-87054|said]]'' <blockquote> Of course I see the pitfalls of auto-approval; it is against all quality assurance principles I have come across . . '''''in an ideal world with plentiful of resources to engage in the KB translation work'''''. In the real world I would argue that auto-approval is the lesser of two evils. In the real world one has to choose: Articles with possible errors vs. none or outdated localized articles. And the latter ones obviously have erroneous content. Anyway, I will watch this thread to see where we land. </blockquote> Thanks for sharing your experience, Erling. I can totally understand the struggle with not having enough people to help with cross-review. But I appreciate your understanding of the importance of that process. At the moment, we haven't land on any sort of decision yet. This discussion is started because I wanted to hear first from our localizers what they think about this idea.