Re: [libreoffice-design] Topic for design team to investigate: Content hosting consolidation on ask.libreoffice.org

[Maarten Brouwers (murb)]

Hi Cor,

> My impression is not that people in this discussion think that it will
> be easy per see.

I got a bit different impression, and still see the extensive requirements doc 
(also the new one) highly contrasting with what I read as the idea of ‘just 
start iterating from ask.libreoffice.org ’ (I 
know, no-one said that as simply). Sure you can do your bookkeeping in Writer, 
no need for Calc (or Base), but I just wonder if you should even considering 
starting with Writer.

> Since there are some advantages on Askbot, and the fact that boosting
> the existing extensions site to a certain level somehow didn't work, we

I’d suggest we first start with ’somehow’, much of what is specced is already 
working to some extend.

- Lack of integration with LibreOffice? -> Create an API on top of the current 
database and integrate it with LibreOffice (another thing that I guess would be 
hard to realize with a foundation based on Askbot)
- Lack of downloaders? -> 
   - Maybe the site wasn’t looking professional enough? 
   - Lack of discoverability, where to find a link to the extensions-site 
(Under “Discover?", that’s where you start before you’ve downloaded 
LibreOffice, not when you’re using it)
   - Maybe the screenshots were too small to even get an idea of what they’re 
getting?
   - Maybe the focus on ’new’ instead of ‘popular' made people wary of the 
quality of what is on offer?
   - Maybe the majority of the users don’t even care as much about extensions 
as much as we’d hope they do, and users don’t expect to find templates there as 
well?
- Lack of contributors? -> All of the above + Maybe a function of the lack of 
downloaders? (no reward in creating & maintaining stuff that is not being used 
as much as you’d hoped for?)
- Lack of moderators? -> All of the above?

I really doubt that OAuth or not has anything to do with it. Localisation, 
probably yes.

g.,


Maarten
-- 
To unsubscribe e-mail to: design+unsubscr...@global.libreoffice.org
Problems? https://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: https://wiki.documentfoundation.org/Netiquette
List archive: https://listarchives.libreoffice.org/global/design/
Privacy Policy: https://www.documentfoundation.org/privacy

Re: [libreoffice-design] Topic for design team to investigate: Content hosting consolidation on ask.libreoffice.org

[Maarten Brouwers (murb)]

Hi Bjoern,

> That -- and the fact that we have 45.000 accounts on it -- are the reason I
> brought up Askbot. Content and accounts are the things that are hard earned 
> on a
> platform, maintainers and moderators are a bit easier, but still hard. The
> pure technical platform is the easiest part.

I wonder how important having existing accounts is for an extension site. I’d 
like to be convinced otherwise, but I think that a usable submission & 
maintenance process will convince people to contribute their extensions. A more 
prominent call for volunteers could help registering the few people needed for 
occasional quality assurance (which with all security threats these days is not 
something you want to do with simply up and down votes [1]). Registering for 
one more account is the least part of creating a good extension / template 
(although it would be a nice to have if you wouldn’t have to). Big exception 
where every hassle of creating accounts might hold back user contributions back 
is feedback from users (star-ratings / comments / up/down-votes). Maybe the ask 
libreoffice org platform could help there (like discourse can be used to allow 
for discussion on pages), but I strongly believe that getting the UX right for 
a good extension website is the hard part, which also requires quite some 
technical skill if you build something like that [2] from scratch (building a 
template site from a Q/A platform is basically building from scratch, support 
for templates is just scratching the surface of what is needed [again, 2]).

g.,


Maarten

[1] I wonder to what extend installing extensions can be called safe in 
LibreOffice. I was wondering if I could make a nice looking extension that does 
something useful and simultaneously sends some user data to a server?
[2] Euh, this: https://nextcloud.documentfoundation.org/s/E5RX5xK6jxQPLdK





-- 
To unsubscribe e-mail to: design+unsubscr...@global.libreoffice.org
Problems? https://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: https://wiki.documentfoundation.org/Netiquette
List archive: https://listarchives.libreoffice.org/global/design/
Privacy Policy: https://www.documentfoundation.org/privacy

Re: [libreoffice-design] Topic for design team to investigate: Content hosting consolidation on ask.libreoffice.org

[Maarten Brouwers (murb)]

Forgot that Bjoern actually mentioned some of the somehows

Technical infrastructure was lacking for contributors:
> [1] Seen e.g. by:
>a/ repeated cries for help on twitter, some examples: (both technical 
> issues)
>   https://twitter.com/nimbleslides/status/1020899933161848832 
>  
>   https://twitter.com/FloraCanou/status/1020517686453735424 
> 
and lack of categorisation-ability which makes it useless for sharing certain 
templates:
>b/ currently no presentation templates being offered there, despite this
>   being the most common use case:
>   
> https://extensions.libreoffice.org/templates?getCategories=Presentation&getCompatibility=any
>  
> 
>   and massive template collections actually being hosted elsewhere e.g.:
>   https://github.com/dohliam/libreoffice-impress-templates 
> 

Those look mainly like technical problems to me. I’m not sure Plone is the 
right basis (don't know it), but maybe helping the current maintainer out and 
addressing the right issues may be a more fruitful attempt to keep a good 
extension site running than starting from scratch and rediscovering the same 
old problems. As said, I’m happy to assist him/her, but just redrafting 
requirements without any observation-based ideas about what makes a great 
template site (and I just don’t buy the idea that is mainly about the 
user-accounts) doesn’t end really well anywhere in my book. Maintaining 
software doesn’t become easier when you build a complex tool (an extensions 
website) on something that isn’t meant for it (a Q/A site).

Yours,


Maarten



-- 
To unsubscribe e-mail to: design+unsubscr...@global.libreoffice.org
Problems? https://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: https://wiki.documentfoundation.org/Netiquette
List archive: https://listarchives.libreoffice.org/global/design/
Privacy Policy: https://www.documentfoundation.org/privacy

Re: [libreoffice-design] Topic for design team to investigate: Content hosting consolidation on ask.libreoffice.org

[Maarten Brouwers (murb)]

Hi Bjoern,

> Stating that a classic CMS is by definition better suited for the task at hand
> than something more interactive like discourse is a foregone conclusion. Given
> that the lack of community involvement is the core reason for many limitations
> of the current setup, its not a valid one.

I think the quickest route to success is adopting a proven framework for 
hosting extensions. If you want to build something unique start from scratch, 
I’d use a generic framework (again, not sure about Plone, I couldn’t find the 
repo for the current site), but really never a framework made for a very 
specific function. I’ve been there (helping others professionaly) and pulling 
my hair out as a developer when I was asked to add some new functionality that 
didn’t align with the core functionality that that specific framework offered. 

> 
>> - Lack of integration with LibreOffice? -> Create an API on top of the
>> current database and integrate it with LibreOffice (another thing that I
>> guess would be hard to realize with a foundation based on Askbot)
>> [...]
>> - Lack of moderators? -> All of the above?
> 
> Please dont guess or speculate on this, you will likely get it wrong. This is
> something that needs to be evaluated by trying.

Sure, these are questions. But I’d always want to try the simplest route that 
also is promising for future development and do basic experiments in the 
current situation when possible that will learn us something more about the 
requirements. You’ve been suggesting starting with a MVP on ask.libreoffice.org 
. Sure, try the templates function, you always 
learn, but I can assure you it will be a deadlock when you want to grow it into 
a full fledged extensions site. Look at the growing list of ideas. One day you 
want features comparable to other platforms and I can’t see how a Q/A platform 
will get us there.

>> I really doubt that OAuth or not has anything to do with it. Localisation, 
>> probably yes.
> 
> Well, where do the 45.000 accounts on askbot come from? The site is far from
> being perfect, but still doing better that all other forums we have. OAuth 
> etc.
> certainly has a role in that, as does gamification, badges and social media
> integration.

Sure, but again, I doubt it matters to an extensions website. Mozilla has no 
OAuth, and still has imho a great extensions website. In an ideal world 
everyone contributes and consumes, but in reality it is just a small 
percentage. Especially when it comes to more complicated stuff like building 
extensions and quality themes.

> For the community the decision should go with the team that gets the most
> content, interaction and users on their platform. 

Interaction & users should imho not be the primary goal for an extensions 
website, great extensions should be. If clearly communicated, I strongly 
believe the 1% of users that is willing to volunteer will step up, but until 
the moment you opened the thread there wasn’t really a way I could see this was 
something I could help out, and I’m subscribed to a list oriented toward 
LibreOffice’s end-user experience (this list). 

g.,


Maarten


-- 
To unsubscribe e-mail to: design+unsubscr...@global.libreoffice.org
Problems? https://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: https://wiki.documentfoundation.org/Netiquette
List archive: https://listarchives.libreoffice.org/global/design/
Privacy Policy: https://www.documentfoundation.org/privacy

Re: [libreoffice-design] Topic for design team to investigate: Content hosting consolidation on ask.libreoffice.org

[Bjoern Michaelsen]

Hi,

On Sun, Oct 14, 2018 at 03:54:13PM +0200, Maarten Brouwers (murb) wrote:
> Sure, but again, I doubt it matters to an extensions website. Mozilla has no
> OAuth, and still has imho a great extensions website. In an ideal world
> everyone contributes and consumes, but in reality it is just a small
> percentage. Especially when it comes to more complicated stuff like building
> extensions and quality themes.

Mozilla has an full discourse forum for extensions:

- https://discourse.mozilla.org/c/add-ons

which allow signing on with Firefox/Github/Google and plain email.

> Interaction & users should imho not be the primary goal for an extensions
> website, great extensions should be.

There is nothing an extension website can do to improve the content uploaded to
it as long as the upload is resonably simple. ... Except building a broad
community (users) and easy feedback, discussion and cooperation (interactions).

Best,

Bjoern

-- 
To unsubscribe e-mail to: design+unsubscr...@global.libreoffice.org
Problems? https://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: https://wiki.documentfoundation.org/Netiquette
List archive: https://listarchives.libreoffice.org/global/design/
Privacy Policy: https://www.documentfoundation.org/privacy

Re: [libreoffice-design] Topic for design team to investigate: Content hosting consolidation on ask.libreoffice.org

[Bjoern Michaelsen]

Hi Maarten,

On Sun, Oct 14, 2018 at 01:55:38PM +0200, Maarten Brouwers (murb) wrote:
> Maintaining software doesn’t become easier when you build a complex tool (an
> extensions website) on something that isn’t meant for it (a Q/A site).

Its really simple:

Dont tell others "Do not consider askbot/nextcloud/discourse".
I will not tell others "Do not consider plone" either.

Decisions will have to be made on the result in the end, which are: content,
users, interactions. However a team delivers that from the tech side is
irrelevant to the decision[1]. I will gladly support a platform that I dont like
technically, if it delivers results.

Best,

Bjoern

[1] Modulo some annoying things like e.g. legal or security considerations.

-- 
To unsubscribe e-mail to: design+unsubscr...@global.libreoffice.org
Problems? https://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: https://wiki.documentfoundation.org/Netiquette
List archive: https://listarchives.libreoffice.org/global/design/
Privacy Policy: https://www.documentfoundation.org/privacy

Re: [libreoffice-design] Topic for design team to investigate: Content hosting consolidation on ask.libreoffice.org

[Bjoern Michaelsen]

Hi,

addendum:

On Sun, Oct 14, 2018 at 02:14:36PM +0200, Bjoern Michaelsen wrote:
> For the community the decision should go with the team that gets the most
> content, interaction and users on their platform. I dont care if someone 
> thinks
> the platform was not made for the task -- it there is a team that makes the
> experience awesome and grows content, users and interaction on it.

the caveat being that in general having more platforms is worse than having
one, as it splits the community and causes friction. So we shouldnt just let
everyone run with their proposal and then end up with five extension websites,
which are all badly maintained by too small teams.

But: For making that call, a team argueing "we can grow content, users accounts
and interactions with $foo" ~always trumps "I think platform $bar was
originally made for the task."

Best,

Bjoern

-- 
To unsubscribe e-mail to: design+unsubscr...@global.libreoffice.org
Problems? https://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: https://wiki.documentfoundation.org/Netiquette
List archive: https://listarchives.libreoffice.org/global/design/
Privacy Policy: https://www.documentfoundation.org/privacy

Re: [libreoffice-design] Topic for design team to investigate: Content hosting consolidation on ask.libreoffice.org

[Bjoern Michaelsen]

Hi,

On Sun, Oct 14, 2018 at 01:24:10PM +0200, Maarten Brouwers (murb) wrote:
> I got a bit different impression, and still see the extensive requirements
> doc (also the new one) highly contrasting with what I read as the idea of
> ‘just start iterating from ask.libreoffice.org ’
> (I know, no-one said that as simply). Sure you can do your bookkeeping in
> Writer, no need for Calc (or Base), but I just wonder if you should even
> considering starting with Writer.

Stating that a classic CMS is by definition better suited for the task at hand
than something more interactive like discourse is a foregone conclusion. Given
that the lack of community involvement is the core reason for many limitations
of the current setup, its not a valid one.

> - Lack of integration with LibreOffice? -> Create an API on top of the
> current database and integrate it with LibreOffice (another thing that I
> guess would be hard to realize with a foundation based on Askbot)
> [...]
> - Lack of moderators? -> All of the above?

Please dont guess or speculate on this, you will likely get it wrong. This is
something that needs to be evaluated by trying. To get integration with
LibreOffice right, the API or hosting is the smallest problem.

The hard problems are:
1/ verification of content (Codereview by humans)
2/ signing of content (ensuring that what was review is what is installed)

Again: Hosting and even integrating this is easy, compared to the social task
of verification, feedback and interaction. As such, the platform underneath is
not much of a relevant destinction -- unless on the latter. Reviewing or even
casual reviewing is the hard part. It requires manpower => which requires an
active community => which requires a platform that encourages interaction.

> I really doubt that OAuth or not has anything to do with it. Localisation, 
> probably yes.

Well, where do the 45.000 accounts on askbot come from? The site is far from
being perfect, but still doing better that all other forums we have. OAuth etc.
certainly has a role in that, as does gamification, badges and social media
integration.

Finally, it should be obvious that is better to do custom development on our
specific (extension hosting) needs on a platform that provides broad generic
features (social media integration, gamification, oauth) than using a specific
platform for "extension hosting" and trying to add a lot of missing generic
features to it. The reason is doing custom development for social media
integration, gamification, OAuth will drown us in constant maintainance of those
custom build features.

Anyways: The start of this thread was "please consider more than one platform"
and the repeatedly given rationale is that bringing together users and content
are harder than bringing together moderators and maintainers and that is harder
than customizing a base technology. This remains universally true and as such
please stop speculating about the base technologies in a way that tries to
suggest to only use one technology.

For the community the decision should go with the team that gets the most
content, interaction and users on their platform. I dont care if someone thinks
the platform was not made for the task -- it there is a team that makes the
experience awesome and grows content, users and interaction on it.

Best,

Bjoern

-- 
To unsubscribe e-mail to: design+unsubscr...@global.libreoffice.org
Problems? https://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: https://wiki.documentfoundation.org/Netiquette
List archive: https://listarchives.libreoffice.org/global/design/
Privacy Policy: https://www.documentfoundation.org/privacy

Re: [libreoffice-design] Topic for design team to investigate: Content hosting consolidation on ask.libreoffice.org

[Cor Nouws]

Hi Maarten,

Maarten Brouwers (murb) wrote on 14-10-18 11:57:

> I wonder how important having existing accounts is for an extension
> site. I’d like to be convinced otherwise, but I think that a usable
> submission & maintenance process will convince people to contribute
> their extensions. A more prominent call for volunteers could help
> registering the few people needed for occasional quality assurance
> (which with all security threats these days is not something you want
> to do with simply up and down votes [1]).

I thought of this subject too, but have the idea that at the moment
there is no quality/safety check either?
Anyway, good that you mention, because having something in place (by
moderation maybe) looks wise.

> to do with simply up and down votes [1]). Registering for one more
> account is the least part of creating a good extension / template
> (although it would be a nice to have if you wouldn’t have to). Big
> exception where every hassle of creating accounts might hold back
> user contributions back is feedback from users (star-ratings /
> comments / up/down-votes). Maybe the ask libreoffice org platform
> could help there (like discourse can be used to allow for discussion
> on pages), but I strongly believe that getting the UX right for a
> good extension website is the hard part, which also requires quite
> some technical skill if you build something like that [2] from

My impression is not that people in this discussion think that it will
be easy per see.
Since there are some advantages on Askbot, and the fact that boosting
the existing extensions site to a certain level somehow didn't work, we
can take this as an opportunity to look at the situation from scratch:
what can we do in a simple way to help people make templates &
extensions available easily.

> scratch (building a template site from a Q/A platform is basically
> building from scratch, support for templates is just scratching the
> surface of what is needed [again, 2]).

Depends. If you want to do versioning of extensions, it will definitely
be hard. If you skip that, leave it to the author on how to make that
clear (or not), it may be much easier. Tags to help finding stuff of
course is a must. But I think I start to repeat what already has been
expressed before.

Best,
Cor

-- 
Cor Nouws
GPD key ID: 0xB13480A6 - 591A 30A7 36A0 CE3C 3D28  A038 E49D 7365 B134 80A6
- vrijwilliger https://nl.libreoffice.org
- volunteer https://www.libreoffice.org
- Member Board The Document Foundation
- http://www.nouenoff.nl / https://www.mijncloudoffice.nl

-- 
To unsubscribe e-mail to: design+unsubscr...@global.libreoffice.org
Problems? https://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: https://wiki.documentfoundation.org/Netiquette
List archive: https://listarchives.libreoffice.org/global/design/
Privacy Policy: https://www.documentfoundation.org/privacy