To reply to some questions above:
Can we please get a higher level of quality from the article writers and especially from the people that are reviewing the articles? ... however when you approve English articles, you are affecting a lot of people
To some, that remark may sound like an insult to anyone contributing, but OK. :) As said earlier, the number of active contributors for TB articles is quite low. And even for FF artciles and paid staff contributors for them, things aren’t always perfect, but I’m sure everyone is doing their best. Regarding “affecting a lot of people”: I think it’s exactly what we want - both for users and localizers, since it’s about vital changes in info that affect all of us. It ’s never meant to annoy people - on the contrary.
That article had four approved versions in the first six days of May and eight approved versions in April...
That’s been defended above.
Another example:
https://support.mozilla.org/en-US/kb/organize-your-messages-using-filters/history
Three versions approved in less than a month.
Basically this was done because despite of the comment (“may need clarification”) intended to not have it approved, someone else approved and marked RFL. The second edit fixed that, the third could be treated as minor since it involved a few words, but has a huge impact on localizations, i.e. not al changes looking minor are as minor as they may look, please keep that in mind.
Besides, unlike the cosmetic changes done, sporting two extra approvals, there is a serious error in the article which has not been corrected. Do you guys read the comments to the articles?
Even though this may involve some changes considered to be “uncertain” sometimes, I find it much more disturbing to see people replying this way instead of applying such changes themselves. Again, the KB is a wiki where everone can apply/suggest changes, whereas it’s best if the person spotting the errors does so rightaway since anyone else would need to jump in and freshly focus on what’s wrong, unless the one making the comment isn’t sure of it and would like to discuss it first. About not reading the comments: frankly, I don’t. I have enough on my plate, and I only see comments for article threads replied to earlier (or for en-US articles edited, and therefore following) - I’m not subscribed to all article’s comments by default, and I’m not into following forum content and other related discussions, nor the many other linked-to-other-discussions. So in short: article discussions in the KB contributor’s forum may be overlooked, and if one’s sure about something being wrong, they’d better apply their edit instead of just notifying “someone” and let it linger, since proper content could be seen as the notifying contributor’s responsibility just as much as for any other.
When approving an article, you should think of it as a printed book. Once it is released, you cannot make corrections, thus you better be sure the product is fine before you send it to the book printer.
That’s been pointed out - the KB is a wiki, and prone to errors. Even professional content writers or translators don’t always do things right at first instance. I’m not sure if you ever contributed to the in-content Help for FF/TB back in the days before Sumo, but compared to that, the current KB is a heaven. Nothing so bad as to find out about (English or localized) content not being accurate and wait for another or even 2 releases to get it fixed, amongst new edits. Something tells me you would rather go back to those days, but I can guarantee you won’t like it (and if you do, try localizing current SM Help for instance. :) )
This is something I’d like to see more input from localizers on. On IRC, I was asked by a localizer to mark more edits as ready for localization. It seems that each localizer has his/her preference about when to mark an edit as ready for localization, so if we can get more localizers coming in, it will help establish a consensus.
That localizer was me. And why? Because and as pointed out above, I’ve see too many unapproved and forgotten edits, as well as changes considered to be minor but hugely affecting localizations. In other words, I’m not sure why en-US content writers/editors/reviewers should a) decide what’s best to mark RFL (or rather, they’re sometimes wrong), and b) (more important), what’s behind the thought of not doing so rightaway? Is such a change not important enough? Are other than English locales not important enough? Do you think users feel comfortable checking the en-US version as well when localized instructions don’t seem accurate, or even understand English, if they don’t trust their localized content in the first place? How would they know if their localized content is incorrect? Ever thought of users relying on KB content and starting forum topics or refering to other sources because of incorrect or unclear KB content? I can give a number of examples of recently not-marked-RFL content changes right here but won’t (unless asked), but it’s an ongoing issue that I heavily dislike (sorry Chris).
Thunderbird articles are sometimes approved without checking the changes thoroughly. Thus, some revisions are approved although there’s something wrong with them, and then of course they need to be corrected.
You might have a point here, depending on who’s approving. I can only speak for myself but when approving, I have probably and at least checked them thoroughly for nonsense, although I might fail when it comes to some TB behavior. So a) that’s why I do my best checking by asking other TB contributors on irc if possible, and b) that’s where others may come in allright. I trust on them to apply more changes a.s.a.p. then, which would show up on the general TB dashboard’s pending revisions that I do check occasionally (not their discussions or the article discussions topic.)
>> This is largely why the ready for localization feature was created - so localizers don't get notified of every edit. How about this solution: by default, we don’t mark edits as ready for localization, then wait a certain amount of time before retroactively marking the latest edit as ready for localization.
That would help a lot. That would actually be what I asked for: Don’t release the book, until you are sure it is fine :-)
I strongly disagree. For starters, we (localizers) are not babies, so we don’t need to get protected for too much workload, nor do we need our food to get “prechewed”. As said, localizers are basically the majority of proofreaders themselves, so they can check and improve, and if there is anything to improve, it affects all locales. Remember there’s also less assumed one-time l10n work involved than writing en-US content in some way, so just do the l10n, think along and check, and feed back if needed. If there’s a new edit approved and marked RFL within let’s say a few days after, that should only mean someone as well as yourself (or rather: the complaining localizer) didn’t check for (or wasn’t quite familiar with) proper en-US content, and everyone should be happy seeing such a new edit, which is likely very tiny as well and hence does not even cause much workload. Besides, postponed approvals are being forgotton about unless properly watched (which basically is already true in some cases), and finally, users are offered improper content meanwhile, which is probably worst of all. All this because some localizers don’t like notifications or workload.
Simple solution is to author the localized version articles from scratch! Word your articles as complete as you think they should be from the start, and don’t worry about what the "other guys" are doing.
IMO, all KB articles should be written by either the developers of whatever is being documented or by technical writers. Relying upon mere users is what begins the circle jerk of edits & re-edits. And the whole approval process is like "too many cooks spoiled the broth", plus the process from the start to approval / posting is too long in too many cases.
I might be wrong but Sumo currently has no developers also being involved in writing new article content as well, in particular for TB. Some efforts have been made to attrack content writers since Roland left, but yet (almost) to no avail. Regardless of TB or FF, anyone doing l10n is a proofreader and is free to adapt a current revision, as long as they don’t start rewriting entire sections after many previous edits and optimizations. For the process, I don’t think there’s much wrong with that - good edits could ideally be approved and marked RFL within a day at most. Too many cooks? I don’t see any sign of that, but of course anyone would like to see less of them, providing there were a few full-time content writers involved. Only if that was true, this discussion might not exist, though I think you would see much more RFL approvals, or none at all.
(see next post)
To reply to some questions above:
<blockquote>Can we please get a higher level of quality from the article writers and especially from the people that are reviewing the articles? ... however when you approve English articles, you are affecting a lot of people</blockquote>
To some, that remark may sound like an insult to anyone contributing, but OK. :) As said earlier, the number of active contributors for TB articles is quite low. And even for FF artciles and paid staff contributors for them, things aren’t always perfect, but I’m sure everyone is doing their best. Regarding “affecting a lot of people”: I think it’s exactly what we want - both for users and localizers, since it’s about vital changes in info that affect all of us. It ’s never meant to annoy people - on the contrary.
<blockquote>That article had four approved versions in the first six days of May and eight approved versions in April...</blockquote>
That’s been defended above.
<blockquote>Another example:
https://support.mozilla.org/en-US/kb/organize-your-messages-using-filters/history
Three versions approved in less than a month.</blockquote>
Basically this was done because despite of the comment (“may need clarification”) intended to not have it approved, someone else approved and marked RFL. The second edit fixed that, the third could be treated as minor since it involved a few words, but has a huge impact on localizations, i.e. not al changes looking minor are as minor as they may look, please keep that in mind.
<blockquote>Besides, unlike the cosmetic changes done, sporting two extra approvals, there is a serious error in the article which has not been corrected. Do you guys read the comments to the articles?</blockquote>
Even though this may involve some changes considered to be “uncertain” sometimes, I find it much more disturbing to see people replying this way instead of applying such changes themselves. Again, the KB is a wiki where everone can apply/suggest changes, whereas it’s best if the person spotting the errors does so rightaway since anyone else would need to jump in and freshly focus on what’s wrong, unless the one making the comment isn’t sure of it and would like to discuss it first. About not reading the comments: frankly, I don’t. I have enough on my plate, and I only see comments for article threads replied to earlier (or for en-US articles edited, and therefore following) - I’m not subscribed to all article’s comments by default, and I’m not into following forum content and other related discussions, nor the many other linked-to-other-discussions. So in short: article discussions in the KB contributor’s forum may be overlooked, and if one’s sure about something being wrong, they’d better apply their edit instead of just notifying “someone” and let it linger, since proper content could be seen as the notifying contributor’s responsibility just as much as for any other.
<blockquote>When approving an article, you should think of it as a printed book. Once it is released, you cannot make corrections, thus you better be sure the product is fine before you send it to the book printer.</blockquote>
That’s been pointed out - the KB is a wiki, and prone to errors. Even professional content writers or translators don’t always do things right at first instance. I’m not sure if you ever contributed to the in-content Help for FF/TB back in the days before Sumo, but compared to that, the current KB is a heaven. Nothing so bad as to find out about (English or localized) content not being accurate and wait for another or even 2 releases to get it fixed, amongst new edits. Something tells me you would rather go back to those days, but I can guarantee you won’t like it (and if you do, try localizing current SM Help for instance. :) )
<blockquote>This is something I’d like to see more input from localizers on. On IRC, I was asked by a localizer to mark more edits as ready for localization. It seems that each localizer has his/her preference about when to mark an edit as ready for localization, so if we can get more localizers coming in, it will help establish a consensus.</blockquote>
That localizer was me. And why? Because and as pointed out above, I’ve see too many unapproved and forgotten edits, as well as changes considered to be minor but hugely affecting localizations. In other words, I’m not sure why en-US content writers/editors/reviewers should a) decide what’s best to mark RFL (or rather, they’re sometimes wrong), and b) (more important), what’s behind the thought of not doing so rightaway? Is such a change not important enough? Are other than English locales not important enough? Do you think users feel comfortable checking the en-US version as well when localized instructions don’t seem accurate, or even understand English, if they don’t trust their localized content in the first place? How would they know if their localized content is incorrect? Ever thought of users relying on KB content and starting forum topics or refering to other sources because of incorrect or unclear KB content? I can give a number of examples of recently not-marked-RFL content changes right here but won’t (unless asked), but it’s an ongoing issue that I heavily dislike (sorry Chris).
<blockquote>Thunderbird articles are sometimes approved without checking the changes thoroughly. Thus, some revisions are approved although there’s something wrong with them, and then of course they need to be corrected.</blockquote>
You might have a point here, depending on who’s approving. I can only speak for myself but when approving, I have probably and at least checked them thoroughly for nonsense, although I might fail when it comes to some TB behavior. So a) that’s why I do my best checking by asking other TB contributors on irc if possible, and b) that’s where others may come in allright. I trust on them to apply more changes a.s.a.p. then, which would show up on the general TB dashboard’s pending revisions that I do check occasionally (not their discussions or the article discussions topic.)
<blockquote>
>> This is largely why the ready for localization feature was created - so localizers don't get notified of every edit. How about this solution: by default, we don’t mark edits as ready for localization, then wait a certain amount of time before retroactively marking the latest edit as ready for localization.
That would help a lot. That would actually be what I asked for: Don’t release the book, until you are sure it is fine :-)
</blockquote>
I strongly disagree. For starters, we (localizers) are not babies, so we don’t need to get protected for too much workload, nor do we need our food to get “prechewed”. As said, localizers are basically the majority of proofreaders themselves, so they can check and improve, and if there is anything to improve, it affects all locales. Remember there’s also less assumed one-time l10n work involved than writing en-US content in some way, so just do the l10n, think along and check, and feed back if needed. If there’s a new edit approved and marked RFL within let’s say a few days after, that should only mean someone as well as yourself (or rather: the complaining localizer) didn’t check for (or wasn’t quite familiar with) proper en-US content, and everyone should be happy seeing such a new edit, which is likely very tiny as well and hence does not even cause much workload. Besides, postponed approvals are being forgotton about unless properly watched (which basically is already true in some cases), and finally, users are offered improper content meanwhile, which is probably worst of all. All this because some localizers don’t like notifications or workload.
<blockquote>Simple solution is to author the localized version articles from scratch! Word your articles as complete as you think they should be from the start, and don’t worry about what the "other guys" are doing.
IMO, all KB articles should be written by either the developers of whatever is being documented or by technical writers. Relying upon mere users is what begins the circle jerk of edits & re-edits. And the whole approval process is like "too many cooks spoiled the broth", plus the process from the start to approval / posting is too long in too many cases.</blockquote>
I might be wrong but Sumo currently has no developers also being involved in writing new article content as well, in particular for TB. Some efforts have been made to attrack content writers since Roland left, but yet (almost) to no avail. Regardless of TB or FF, anyone doing l10n is a proofreader and is free to adapt a current revision, as long as they don’t start rewriting entire sections after many previous edits and optimizations. For the process, I don’t think there’s much wrong with that - good edits could ideally be approved and marked RFL within a day at most. Too many cooks? I don’t see any sign of that, but of course anyone would like to see less of them, providing there were a few full-time content writers involved. Only if that was true, this discussion might not exist, though I think you would see much more RFL approvals, or none at all.
(see next post)
