l10n ideas and discussions

0 views
Skip to first unread message

David Tenser

unread,
Oct 1, 2007, 7:56:37 PM10/1/07
to
There are a number of discussions circulating about l10n (localization)
for Sumo and how to handle that as the site grows. I would like to hear
your thoughts on the topics. Some ideas have popped up here at the
Mozilla offices, and others have been discussed about during phone
meeting, but we need to get these ideas out for everyone to read and
comment on.


What we've established so far:

* Generic url format for the KB should be
https://support.mozilla.com/en-US/kb/ArticleName
https://support.mozilla.com/sv-SE/kb/ArticleName

* If a user navigates to e.g.
https://support.mozilla.com/sv-SE/kb/ArticleName and a Swedish
translation of ArticleName doesn't yet exist, the en-US version would be
displayed by default, with a <div> box at the top of the article
informing the user that a Swedish translation doesn't exist and that
he/she is welcome to do the translation.

* Since TikiWiki supports different locale fallbacks than just en-US, we
should obviously support that as well. For visitors not logged in, en-US
would be the default.


Ideas that are not yet formalized:

* When someone makes a significant update to an en-US version of a kb
article, we need a mechanism to "notify" the localized versions of that
same page that there has been significant updates. Ideally, this would
be a checkbox that you check when making changes to a document saying
e.g. "This is an update that needs to be translated". What that checkbox
would do is actually marking all localized versions as "Possibly
outdated". The localizers could then check the changes in the en-US page
and, upon finishing updating the page, mark the page as "up to date".

* If an important update actually happens on a localized version first,
we don't have a nice way of handling it, but I think just sending an
e-mail or adding a note on the en-US page discussion thread that "it
needs an update".

* For the localized start pages, it might make sense to link to third
party sites that have better content. The long term goal is to have
everything hosted on support.mozilla.com, but promoting the major
contributors' sites first would be a good start. Some locales are
actually ad sponsored and it could be hard to convince them to move
their content.


What we're still pondering about:

* How about in addition to the discussion thread for every article, we
could have some sort of Todo-list, where you could say e.g. "Fix
paragraph 2, wrong file path on Linux (don't have Linux here)" or
whatever. The alternative would be to use bugzilla for these kinds of
things, with the new Knowledge Base Content component of Sumo. Thoughts?


David

David Tenser

unread,
Oct 1, 2007, 8:30:11 PM10/1/07
to
Other things to think about is what to do with Forum and Live Chat.
We'll probably not be able to gather enough contributors to make
localized Live Chat work, but we still might want to translate the
launch page for it?

For forums, maybe we want to set up forums for the most popular locales?
We should probably start with English only though.

David

Jesper Kristensen

unread,
Oct 2, 2007, 11:23:12 AM10/2/07
to
David Tenser skrev:

> There are a number of discussions circulating about l10n (localization)
> for Sumo and how to handle that as the site grows. I would like to hear
> your thoughts on the topics. Some ideas have popped up here at the
> Mozilla offices, and others have been discussed about during phone
> meeting, but we need to get these ideas out for everyone to read and
> comment on.
>
>
> What we've established so far:
>
> * Generic url format for the KB should be
> https://support.mozilla.com/en-US/kb/ArticleName
> https://support.mozilla.com/sv-SE/kb/ArticleName

I like to see pretty urls. How would pages that are not articles look
like? I still see some problems with these urls:

* Article names should be localized. I see no reason for having a
localized article with an English name. I don't know the software used.
It might be easier to track things for localized articles that are
translations of English articles if they have the same name. But I think
having a localized name gives more value than being able to track changes.

* I think the locale code for Swedish is sv and not sv-SE, but I might
be wrong.

* Where is Firefox??? I know that mozilla.com is the home of Firefox
Corporation ("Mozilla Corporation") and that a new Thunderbird
Corporation ("MailCo" or unnamed) is being created. This problem might
have nothing to do with Sumo but with a wrong choice of domain name for
the home of Firefox. Suggestions:
support.firefox.com/en-US/kb/ArticleName or
support.mozilla.com/en-US/firefox/kb/ArticleName

>
> * If a user navigates to e.g.
> https://support.mozilla.com/sv-SE/kb/ArticleName and a Swedish
> translation of ArticleName doesn't yet exist, the en-US version would be
> displayed by default, with a <div> box at the top of the article
> informing the user that a Swedish translation doesn't exist and that
> he/she is welcome to do the translation.
>
> * Since TikiWiki supports different locale fallbacks than just en-US, we
> should obviously support that as well. For visitors not logged in, en-US
> would be the default.

I assume you mean "en-US would be the default fallback".

What happens if the article is not localized into en-US?

Why not use the user's Accept-Language header?

>
>
> Ideas that are not yet formalized:
>
> * When someone makes a significant update to an en-US version of a kb
> article, we need a mechanism to "notify" the localized versions of that
> same page that there has been significant updates. Ideally, this would
> be a checkbox that you check when making changes to a document saying
> e.g. "This is an update that needs to be translated". What that checkbox
> would do is actually marking all localized versions as "Possibly
> outdated". The localizers could then check the changes in the en-US page
> and, upon finishing updating the page, mark the page as "up to date".

I assume by your post that the software used has some way of knowing if
article X is a translation of article Y. But is it possible to change
that relationship? Like: "Article X is now no longer a translation of
article Y, but an independent article" or "Article X is no longer an
independent article, but a translation of article Y."

Do you have any idea of how much of the localized content are going to
be translations of en-US and how much is not? If we move our content for
my locale (da, Danish) to Sumo, none of it would be translations, at
least in the beginning. The ratio might change later, but I guess it
won't change much.

>
> * If an important update actually happens on a localized version first,
> we don't have a nice way of handling it, but I think just sending an
> e-mail or adding a note on the en-US page discussion thread that "it
> needs an update".
>
> * For the localized start pages, it might make sense to link to third
> party sites that have better content. The long term goal is to have
> everything hosted on support.mozilla.com, but promoting the major
> contributors' sites first would be a good start. Some locales are
> actually ad sponsored and it could be hard to convince them to move
> their content.

Implementing this in a good way is a prerequisite for moving our
existing Danish content into Sumo. Of course I can't say if anyone out
there would make new Danish content for Sumo if this point is not done well.

To be more concrete: Creating a Sumo forum in Danish would be a pain in
the *** for Danish contributors as long as Thunderbird, Sunbird etc. is
not part of Sumo. On the other hand, I don't see a problem in spreading
kb content out on different sites, and I see a benefit in having it
hosted at Sumo. But that would require a solution where the forum
(hosted by ourselves) and the kb (hosted by Sumo) would feel tightly
integrated. I also think we would like to make other possibilities than
the kb more visible on the front page for our locale, therefore
flexibility is a key issue here.

pascal

unread,
Oct 2, 2007, 1:03:50 PM10/2/07
to
David Tenser a écrit :

> What we've established so far:

> would do is actually marking all localized versions as "Possibly
> outdated". The localizers could then check the changes in the en-US page
> and, upon finishing updating the page, mark the page as "up to date".

Will it provide a diff to the localizer? Is there a way to link a
specific en-US revision of a document to a translated revision? That
would be useful.

> * For the localized start pages, it might make sense to link to third
> party sites that have better content. The long term goal is to have
> everything hosted on support.mozilla.com, but promoting the major
> contributors' sites first would be a good start. Some locales are
> actually ad sponsored and it could be hard to convince them to move
> their content.

Is "moving their content" equivalent to "closing their sites"? Sounds
scary the way you put it. You should also define what "everything
hosted" means for you, community portals are rarely limited to
Firefox/thunderbird but also give support to other gecko-based software
like Nvu/Kompozer, Seamonkey, Camino, Sunbird... What is the scope of Sumo?

Pascal

David Tenser

unread,
Oct 2, 2007, 2:07:12 PM10/2/07
to
Jesper Kristensen wrote:
> David Tenser skrev:
>> There are a number of discussions circulating about l10n
>> (localization) for Sumo and how to handle that as the site grows. I
>> would like to hear your thoughts on the topics. Some ideas have popped
>> up here at the Mozilla offices, and others have been discussed about
>> during phone meeting, but we need to get these ideas out for everyone
>> to read and comment on.
>>
>>
>> What we've established so far:
>>
>> * Generic url format for the KB should be
>> https://support.mozilla.com/en-US/kb/ArticleName
>> https://support.mozilla.com/sv-SE/kb/ArticleName
>
> I like to see pretty urls. How would pages that are not articles look
> like? I still see some problems with these urls:

Example
https://support.mozilla.com/en-US/forum/

So the "/kb/" is just for knowledge base articles.

> * Article names should be localized. I see no reason for having a
> localized article with an English name. I don't know the software used.
> It might be easier to track things for localized articles that are
> translations of English articles if they have the same name. But I think
> having a localized name gives more value than being able to track changes.
>

There are both pros and cons as you say. According to Nelson Ko of
TikiWiki, it is actually possible to have both. See
https://bugzilla.mozilla.org/show_bug.cgi?id=388841#c16
I will talk to Nelson today about this and other stuff.

> * I think the locale code for Swedish is sv and not sv-SE, but I might
> be wrong.

Bad example then. :)

>
> * Where is Firefox??? I know that mozilla.com is the home of Firefox
> Corporation ("Mozilla Corporation") and that a new Thunderbird
> Corporation ("MailCo" or unnamed) is being created. This problem might
> have nothing to do with Sumo but with a wrong choice of domain name for
> the home of Firefox. Suggestions:
> support.firefox.com/en-US/kb/ArticleName or
> support.mozilla.com/en-US/firefox/kb/ArticleName

Currently there are no plans to do support/help for other products. If
we do end up with a similar support site for Thunderbird, setting up a
/firefox/ sub-folder and server redirects will be an easy task.


>>
>> * If a user navigates to e.g.
>> https://support.mozilla.com/sv-SE/kb/ArticleName and a Swedish
>> translation of ArticleName doesn't yet exist, the en-US version would
>> be displayed by default, with a <div> box at the top of the article
>> informing the user that a Swedish translation doesn't exist and that
>> he/she is welcome to do the translation.
>>
>> * Since TikiWiki supports different locale fallbacks than just en-US,
>> we should obviously support that as well. For visitors not logged in,
>> en-US would be the default.
>
> I assume you mean "en-US would be the default fallback".

Right!

>
> What happens if the article is not localized into en-US?

That would only happen if the user is trying to access a locale-specific
page in en-US, in which case we would show a 404 I guess. Not that I
think many users would manually enter URL names like that.

>
> Why not use the user's Accept-Language header?

That's a good idea and we might do exactly that. I need to talk to
Nelson to see what's possible there.

>
>>
>>
>> Ideas that are not yet formalized:
>>
>> * When someone makes a significant update to an en-US version of a kb
>> article, we need a mechanism to "notify" the localized versions of
>> that same page that there has been significant updates. Ideally, this
>> would be a checkbox that you check when making changes to a document
>> saying e.g. "This is an update that needs to be translated". What that
>> checkbox would do is actually marking all localized versions as
>> "Possibly outdated". The localizers could then check the changes in
>> the en-US page and, upon finishing updating the page, mark the page as
>> "up to date".
>
> I assume by your post that the software used has some way of knowing if
> article X is a translation of article Y. But is it possible to change
> that relationship? Like: "Article X is now no longer a translation of
> article Y, but an independent article" or "Article X is no longer an
> independent article, but a translation of article Y."

I don't know. Is that something that would be common usage you think?

>
> Do you have any idea of how much of the localized content are going to
> be translations of en-US and how much is not? If we move our content for
> my locale (da, Danish) to Sumo, none of it would be translations, at
> least in the beginning. The ratio might change later, but I guess it
> won't change much.

I'd be very interested in hearing the reason of that. Why wouldn't the
Danish users benefit from having the current KB translated to Danish?
What kind of content would exist if not what's currently in the en-us kb?

>
>>
>> * If an important update actually happens on a localized version
>> first, we don't have a nice way of handling it, but I think just
>> sending an e-mail or adding a note on the en-US page discussion thread
>> that "it needs an update".
>>
>> * For the localized start pages, it might make sense to link to third
>> party sites that have better content. The long term goal is to have
>> everything hosted on support.mozilla.com, but promoting the major
>> contributors' sites first would be a good start. Some locales are
>> actually ad sponsored and it could be hard to convince them to move
>> their content.
>
> Implementing this in a good way is a prerequisite for moving our
> existing Danish content into Sumo. Of course I can't say if anyone out
> there would make new Danish content for Sumo if this point is not done
> well.
>
> To be more concrete: Creating a Sumo forum in Danish would be a pain in
> the *** for Danish contributors as long as Thunderbird, Sunbird etc. is
> not part of Sumo. On the other hand, I don't see a problem in spreading
> kb content out on different sites, and I see a benefit in having it
> hosted at Sumo. But that would require a solution where the forum
> (hosted by ourselves) and the kb (hosted by Sumo) would feel tightly
> integrated. I also think we would like to make other possibilities than
> the kb more visible on the front page for our locale, therefore
> flexibility is a key issue here.
>

Absolutely. And I'd be interested in hearing your thoughts on how to
move forward with the Danish locale. I can definitely see us linking to
your content and forums.

David

David Tenser

unread,
Oct 2, 2007, 2:31:02 PM10/2/07
to
pascal wrote:
> David Tenser a écrit :
>
>> What we've established so far:
>> would do is actually marking all localized versions as "Possibly
>> outdated". The localizers could then check the changes in the en-US
>> page and, upon finishing updating the page, mark the page as "up to
>> date".
>
> Will it provide a diff to the localizer? Is there a way to link a
> specific en-US revision of a document to a translated revision? That
> would be useful.

The localized versions of an article are linked somehow to the English
version, so having this ability shouldn't be a huge problem. And you're
right, it would be helpful. One more thing to bring up in today's
meeting with Nelson. :)

>
>> * For the localized start pages, it might make sense to link to third
>> party sites that have better content. The long term goal is to have
>> everything hosted on support.mozilla.com, but promoting the major
>> contributors' sites first would be a good start. Some locales are
>> actually ad sponsored and it could be hard to convince them to move
>> their content.
>
> Is "moving their content" equivalent to "closing their sites"? Sounds
> scary the way you put it. You should also define what "everything
> hosted" means for you, community portals are rarely limited to
> Firefox/thunderbird but also give support to other gecko-based software
> like Nvu/Kompozer, Seamonkey, Camino, Sunbird... What is the scope of Sumo?
>

Right, so this is a delicate topic indeed.

What we need to do is to get an overview of the existing
community-driven support channels out there for all locales and then
decide a good approach from there. For some locales with near-complete
coverage of our content, it might make sense to just promote those sites
and link to them from Sumo. For others, it might be small enough to just
"move over" if the contributors want it. For some, we might be able to
host the content on our site as well as keeping the community site and
see how that goes. But the long term goal for Sumo should be to provide
complete support for Firefox in all of our locales.


The scope of Sumo is just Firefox.

Jason Barnabe (np)

unread,
Oct 2, 2007, 2:58:59 PM10/2/07
to
David Tenser wrote:

> Jesper Kristensen wrote:
> >
> > Do you have any idea of how much of the localized content are going to
> > be translations of en-US and how much is not? If we move our content for
> > my locale (da, Danish) to Sumo, none of it would be translations, at
> > least in the beginning. The ratio might change later, but I guess it
> > won't change much.
>
> I'd be very interested in hearing the reason of that. Why wouldn't the
> Danish users benefit from having the current KB translated to Danish?
> What kind of content would exist if not what's currently in the en-us kb?

I think the confusion here may be in the word "translate". You can
have the same content in different languages without having to
translate it.

For example, if there's already Danish content on "Lost Bookmarks",
you would hook that up to the English "Lost Bookmarks"; you wouldn't
translate the English article into Danish. The two would generally
have the same content (and once they were hooked up could gain some
useful information from each other), but they wouldn't be translations
of each other. I think the majority of content would be like this.

Thinking of different versions as "translations" leads to thinking
that the only role of non-English contributors is to translate what
the English contributors have written. While I think that's true with
Firefox l10n, I don't think it's the case here. Content can start in a
language other than English.

Jesper Kristensen

unread,
Oct 2, 2007, 3:59:14 PM10/2/07
to
David Tenser skrev:

> Currently there are no plans to do support/help for other products.

Exactly. Why does a website for Firefox users not have Firefox in its name?

>> Do you have any idea of how much of the localized content are going to
>> be translations of en-US and how much is not? If we move our content
>> for my locale (da, Danish) to Sumo, none of it would be translations,
>> at least in the beginning. The ratio might change later, but I guess
>> it won't change much.
>
> I'd be very interested in hearing the reason of that. Why wouldn't the
> Danish users benefit from having the current KB translated to Danish?
> What kind of content would exist if not what's currently in the en-us kb?

I am not talking about Danish users, but the Danish community. We
already have some support articles in Danish that we might be able to
copy, but these articles are not a translation of any English articles.
Replicating the amount of content in en-US is something that I don't
think is possible at this time.

Example:
http://support-stage.mozilla.org/kb/lost+bookmarks
http://mozilladanmark.dk/wiki/Firefox/Bogm%C3%A6rker_forsvundet

These two articles are on the same subject. But as you might see from
the size of each article, the Danish article is not a translation of the
English one. One day someone might have time to do a full translation,
but no one has done it yet, and I guess no one will in the near feature.
The Danish article tries to cover the most essential part, and in edge
cases, users have to refer to the English article. It would be great to
have a localization of the English kb into Danish, but there are no
resources for that at this time.

Another example of what we focus on in the Danish kb is something like this:
http://mozilladanmark.dk/wiki/Firefox/TDC_Digital_signatur
This article is about how to obtain and install the digital signature
for use on Danish government websites - content that probably won't
exist in English.

David Tenser

unread,
Oct 2, 2007, 4:03:58 PM10/2/07
to

So I think you're right about l10n having generally the same content and
that we could gain useful information from each other. I don't
necessarily agree that there will be a lot of different versions of a
page for different locales, but sure that could happen. However, since
we're going to benefit from the page ratings of the pages, we should be
able to see if our content is good or bad, and I'm assuming most of our
weekly metrics will be based on the en-US content.

What that means is that the en-US content will be subject to continuous
incremental improvements to a greater extent than the
localized/translated content. Therefore, as the content quality
increases, it would be wise for localizers to stay on track with how
en-US presents the content to its readers.

There might still be cases where a localized version is measurably
better than the en-US version, in which case we should simply update the
en-US version accordingly.

I'm all for content freedom when it comes to l10n, as long as we can
agree on a common goal: to leverage the best possible content to our users.

David

David Tenser

unread,
Oct 2, 2007, 4:13:51 PM10/2/07
to
Jesper Kristensen wrote:
> David Tenser skrev:
>> Currently there are no plans to do support/help for other products.
>
> Exactly. Why does a website for Firefox users not have Firefox in its name?
>

Actually the name of the site is "Firefox Support Home Page" at the
moment. Are you referring to the url?

>>> Do you have any idea of how much of the localized content are going
>>> to be translations of en-US and how much is not? If we move our
>>> content for my locale (da, Danish) to Sumo, none of it would be
>>> translations, at least in the beginning. The ratio might change
>>> later, but I guess it won't change much.
>>
>> I'd be very interested in hearing the reason of that. Why wouldn't the
>> Danish users benefit from having the current KB translated to Danish?
>> What kind of content would exist if not what's currently in the en-us kb?
>
> I am not talking about Danish users, but the Danish community. We
> already have some support articles in Danish that we might be able to
> copy, but these articles are not a translation of any English articles.
> Replicating the amount of content in en-US is something that I don't
> think is possible at this time.
>
> Example:
> http://support-stage.mozilla.org/kb/lost+bookmarks
> http://mozilladanmark.dk/wiki/Firefox/Bogm%C3%A6rker_forsvundet
>

Yes, and I don't see that as a problem right now. Maybe not ever. As
long as we can get good content on the site, it's irrelevant whether or
not it's a sentence by sentence translation or not. However, the article
should still technically be connected to the en-US article so we can say
that "we have a Danish version of LostBookmarks".

> These two articles are on the same subject. But as you might see from
> the size of each article, the Danish article is not a translation of the
> English one. One day someone might have time to do a full translation,
> but no one has done it yet, and I guess no one will in the near feature.
> The Danish article tries to cover the most essential part, and in edge
> cases, users have to refer to the English article. It would be great to
> have a localization of the English kb into Danish, but there are no
> resources for that at this time.

I actually understand Danish (I'm Swedish). :)

And we should take advantage of the different versions of pages because
we can actually measure which one of them is the best using the page
rating metrics.

>
> Another example of what we focus on in the Danish kb is something like
> this:
> http://mozilladanmark.dk/wiki/Firefox/TDC_Digital_signatur
> This article is about how to obtain and install the digital signature
> for use on Danish government websites - content that probably won't
> exist in English.

So content that doesn't apply to other locales don't have any
restrictions at all as far as translation goes. Content that helps the
users of that locale should just be there. :)

David

Jesper Kristensen

unread,
Oct 3, 2007, 3:00:20 AM10/3/07
to
David Tenser skrev:

> Actually the name of the site is "Firefox Support Home Page" at the
> moment. Are you referring to the url?

Yes.

> Yes, and I don't see that as a problem right now. Maybe not ever. As
> long as we can get good content on the site, it's irrelevant whether or
> not it's a sentence by sentence translation or not. However, the article
> should still technically be connected to the en-US article so we can say
> that "we have a Danish version of LostBookmarks".

Ok, I can see your point.

Axel Hecht

unread,
Oct 3, 2007, 6:23:36 AM10/3/07
to

What's "long term" here?

Localized content on SUMO and local forums etc is going to be really
interesting to new and emerging localizations and their communities. For
those, there is no migration effort, neither on the side of the content
authors, nor on the side of users. I have no idea what they would do for
Thunderbird, though. Not that I know what I will do with/for
Thunderbird, honestly. Did you talk to David Ascher on this yet?

That said, you laid out many good reasons for local communities to keep
their sites. First and foremost, support is a people business. The
support sites are going to be where the people are. Moving just the
Firefox part over to SUMO might work for those communities that are
already structured that way, but AFAICT, that's a slim portion of the
world. For most communities, their sites cover all the various Mozilla
products and projects, and the community is not going to leave those
behind. Assuming that they'd support SUMO, too, they had to be working
on two sites, which is not working out that well in general. I don't see
any existing community moving over lightly, or soon.

This is just the people/community part. Migrating existing content is
another. This is not just a question of KB, but forum content, too. For
both, we'll be in the very same licensing mess that we're in with mz. We
were able to solve it there, but localized content is living under
different legal systems, and I know that we didn't dare to merge the Fx
and Tb forums in Germany just for licensing reasons.

Another thought on the Firefox-centric thing. I'm not convinced that
there is a globally-right answer on whether SUMO should be just Firefox
or Mozilla-project-wide. If we need to have one answer, SUMO might just
not be the right pick for everybody. If we can have more than Firefox on
SUMO for some, we probably end up with more for all.

Bottom line, to me, SUMO is an opt-in thing on a global scale, and we'll
need to constantly re-evaluate those decisions for each reason. And it
will change in terms of years rather than months.

Axel

David Tenser

unread,
Oct 3, 2007, 1:59:30 PM10/3/07
to

We are currently working on the technical requirements for l10n, which I
think will take less than a month to complete for the web developers.
Once that's completed, we will start to encourage contributors to
localize the content. I don't have any estimations on when the big
locales are "complete", but I would say it depends heavily on our
success in gathering community attention to the project.

We could probably get a good head start by getting help from the
in-product localizers. The plan is to move that content to Sumo anyway,
so hopefully we can get that going as soon as the technical requirements
are implemented.

>
> Localized content on SUMO and local forums etc is going to be really
> interesting to new and emerging localizations and their communities. For
> those, there is no migration effort, neither on the side of the content
> authors, nor on the side of users. I have no idea what they would do for
> Thunderbird, though. Not that I know what I will do with/for
> Thunderbird, honestly. Did you talk to David Ascher on this yet?

I have scheduled a meeting with him when he's here in Mountain View next
week to discuss the situation. I can definitely see us sharing the
support platform in the future.

You raise a valid point about the l10n strategy, to approach the new and
emerging locales first. The more content we get over on Sumo, the more
attractive the site will look for other locales.

>
> That said, you laid out many good reasons for local communities to keep
> their sites. First and foremost, support is a people business. The
> support sites are going to be where the people are. Moving just the
> Firefox part over to SUMO might work for those communities that are
> already structured that way, but AFAICT, that's a slim portion of the
> world. For most communities, their sites cover all the various Mozilla
> products and projects, and the community is not going to leave those
> behind. Assuming that they'd support SUMO, too, they had to be working
> on two sites, which is not working out that well in general. I don't see
> any existing community moving over lightly, or soon.

I could give a number of good reasons to have their content hosted on
Sumo as well:

* The content would get official recognition, which would benefit both
the readers and authors.
* They would automatically benefit from technology to collect metrics
such as page ranking, traffic and bounce rate.
* The content would integrate with other languages and the readers would
have easy access to not-yet-translated content.
* The site would get more exposure and would likely get more complete
content faster.
* It would blend in well with the visual style of the rest of the
Firefox end-user documentation.

All in all, moving their content to Sumo would really help users of
Firefox to find it.

>
> This is just the people/community part. Migrating existing content is
> another. This is not just a question of KB, but forum content, too. For
> both, we'll be in the very same licensing mess that we're in with mz. We
> were able to solve it there, but localized content is living under
> different legal systems, and I know that we didn't dare to merge the Fx
> and Tb forums in Germany just for licensing reasons.

I'm not sure what you mean with migrating existing forum content. The
strategy for Sumo is to use Forum as a fallback if the requested content
isn't already available in Sumo (see
http://wiki.mozilla.org/images/9/9d/Support-Firefox_Support_Work_Flow.pdf).


>
> Another thought on the Firefox-centric thing. I'm not convinced that
> there is a globally-right answer on whether SUMO should be just Firefox
> or Mozilla-project-wide. If we need to have one answer, SUMO might just
> not be the right pick for everybody. If we can have more than Firefox on
> SUMO for some, we probably end up with more for all.

I think the approach from Mozilla on this is that we need to do
something about the Firefox support situation because Firefox is used by
over 100M users. So the focus is clearly on Firefox for now, until we
can get the site up and running fully with a vibrant community around
it. What you're saying, though, is that getting a vibrant community
around it in part could depend on whether to support just Firefox or
Mozilla in general.

>
> Bottom line, to me, SUMO is an opt-in thing on a global scale, and we'll
> need to constantly re-evaluate those decisions for each reason. And it
> will change in terms of years rather than months.

Absolutely. And at this point it's better to move forward than to think
about the best solution. It's easier to steer the wheel when the vehicle
is moving. :)

David

Jesper Kristensen

unread,
Oct 4, 2007, 6:16:29 AM10/4/07
to
David Tenser skrev:

> I could give a number of good reasons to have their content hosted on
> Sumo as well:
>
> * The content would get official recognition, which would benefit both
> the readers and authors.
> * They would automatically benefit from technology to collect metrics
> such as page ranking, traffic and bounce rate.
> * The content would integrate with other languages and the readers would
> have easy access to not-yet-translated content.
> * The site would get more exposure and would likely get more complete
> content faster.
> * It would blend in well with the visual style of the rest of the
> Firefox end-user documentation.
>
> All in all, moving their content to Sumo would really help users of
> Firefox to find it.

Some of these points are not valid, and I hope they ar not going to be
valid. Currently mozilla-europe.org links to the local community support
sites. And I hope mozilla.com will do the came when localization of that
site is up and running. If a community site does a better job in a
locale than Sumo does in that locale, mozilla.com should link to the
community site for that locale.

The point about metrics and integration of content are however valid.
Ant there is another point too: Local communities doesn't have to set up
and host the site themselves.

> I'm not sure what you mean with migrating existing forum content. The
> strategy for Sumo is to use Forum as a fallback if the requested content
> isn't already available in Sumo (see
> http://wiki.mozilla.org/images/9/9d/Support-Firefox_Support_Work_Flow.pdf).

Interesting graphic. While this may work for large locales like en-US,
it might not be a good work flow for smaller locales. When one en-US
user has a problem, at least 1000 others are likely to have the same
problem, and a kb would work better than a forum. However when one user
of a small locale has a problem, that particular problem might (or might
not) only be experienced by a few other people in the same locale, and
it might be easier to give these three people a short individual answer
than creating/translating a large kb article on the subject. Therefore
it might be desirable to vary the visibility of forums and chat between
locales.

> What you're saying, though, is that getting a vibrant community
> around it in part could depend on whether to support just Firefox or
> Mozilla in general.

I think this is very likely to be true.

Majken Connor

unread,
Oct 4, 2007, 3:14:49 PM10/4/07
to Jesper Kristensen, support-...@lists.mozilla.org
but what are the odds that users in any given locale are looking for support that is locale specific?  Any of the problems that people are hitting in en-us are going to be problems in all the other locales, and as more and more users from each locale switch to firefox, the likelihood of them needing any one of the documents that are already written in english goes up as well.

Certainly in locales with fewer users it's less work for the localizers to field the questions one at a time, but is that better for the users?  Do the users prefer having to ask a question and wait for an answer, but get a personal response, or do they prefer to be able to find the answer already written down somewhere?

David Tenser

unread,
Oct 4, 2007, 3:33:14 PM10/4/07
to
Jesper Kristensen wrote:
> David Tenser skrev:
>> I could give a number of good reasons to have their content hosted on
>> Sumo as well:
>>
>> * The content would get official recognition, which would benefit both
>> the readers and authors.
>> * They would automatically benefit from technology to collect metrics
>> such as page ranking, traffic and bounce rate.
>> * The content would integrate with other languages and the readers
>> would have easy access to not-yet-translated content.
>> * The site would get more exposure and would likely get more complete
>> content faster.
>> * It would blend in well with the visual style of the rest of the
>> Firefox end-user documentation.
>>
>> All in all, moving their content to Sumo would really help users of
>> Firefox to find it.
>
> Some of these points are not valid, and I hope they ar not going to be
> valid. Currently mozilla-europe.org links to the local community support
> sites. And I hope mozilla.com will do the came when localization of that
> site is up and running. If a community site does a better job in a
> locale than Sumo does in that locale, mozilla.com should link to the
> community site for that locale.

Sure, as long as the community sites are _better_ than Sumo, we'll link
to them. The goal is to do what's best for our users.

>
> The point about metrics and integration of content are however valid.
> Ant there is another point too: Local communities doesn't have to set up
> and host the site themselves.

Good. Noted. :)

David

David Tenser

unread,
Oct 4, 2007, 4:00:54 PM10/4/07
to
Majken Connor wrote:
> but what are the odds that users in any given locale are looking for
> support that is locale specific? Any of the problems that people are
> hitting in en-us are going to be problems in all the other locales, and
> as more and more users from each locale switch to firefox, the
> likelihood of them needing any one of the documents that are already
> written in english goes up as well.
>
> Certainly in locales with fewer users it's less work for the localizers
> to field the questions one at a time, but is that better for the users?
> Do the users prefer having to ask a question and wait for an answer, but
> get a personal response, or do they prefer to be able to find the answer
> already written down somewhere?
>

I agree with Majken here.

Reply all
Reply to author
Forward
0 new messages