Community mailing list archives

Re: Reflection about documentation.

Akretion, Raphael Valyi
- 08/13/2014 13:41:55

It happens that earlier in the thread I mentioned that even built more effectively, open source ERP still has a cost and that knowledge was part of the revenue model of OpenERP SA, at least the one of the past.

One has to notice that knowledge revenue model was not at all just about selling documentation (this never happened even) or training.

Instead people have to remind themselves that OpenERP SA is making a large majority of its revenue through a partner network and the revenue they generate by paying an annual fee and the Odoo Enterprise their customer buy eventually.

Well, the fact that knowledge was initially scarce certainly fueled a bit that model.
** Please notice that I'm not defending that knowledge should be scarse at all, just explaining a few things.**

Many partners payed a partnership fee because of the bundled training (or discount on training). For instance in Brazil we we used to deliver the certified training, I see that the sales of partnerships dropped the last month after some mismatch between the global certification program and the implementers local expectations. So certainly OpenERP selling partnerships was also related to selling knowledge.

Also, knowledge being scarse and Odoo being a highly moving target. This had a relatively high entry barrier. With a high entry barrier, only very dedicated professionals would eventually succeed doing it. And again a large part of these dedicated professionals happened to pay a partnership fee.

So of course, the trend for the future is a democratization of Odoo and this is absolutely great.
But this is not true that the knowledge part of OpenERP/Odoo revenue model was just about the direct trainings revenues.
That is the large availability of a great documentation has a cost that should be absorbed, definitely social aspects can decrease that cost. And the large availability of Odoo knowledge will probably even weaken the existence of a paid partnership network or of Enterprise services. So for OpenERP SA it's important that the scalability (not the technical scalability) of its SaaS service is able to replace the old revenue model and / or that the cost of maintaining/evolving Odoo is lowered so much (by good coding practises for instance) that it can deal with the absence of revenue growth during the replacement of the old revenue model.

Also, David opposed the "healthy ecosystem" vs the "revenue model". Well, as a product I would say OpenERP/Odoo already won the battle. It may not be clear to everybody on every markets yet, but even if there are some frustrated expectations, it is quite clear that the product will do great anyway, at least better than these proprietary ERP and even other open source ERP's, even with the current bad documentation.
On the contrary, I'm personally much more worried about the revenue model of the editor as an open source editor rather than anything else. Because for me the biggest threat to Odoo today is not that the product will fail, it is that it would be pressured toward a proprietary business model and/or eventually fragmented with such pressure. Again hoping to be wrong but we are never too picky.

"The problem is that it's a huge overhead to do a pull request just to fix a typo. It has to be super easy for documentation writers to contribute. We need to support the "Edit in Github" button."

-> Github editable wiki is called the Gollum wiki my friend and this is open source and that can even connect to Odoo with my work:
we use it as our primarily wiki at Akretion and are happy with it.


On Wed, Aug 13, 2014 at 1:09 PM, Fabien Pinckaers <> wrote:
Hi everyone,

It's clear that documentation is a key topic to onboard new users,
developers and designer.

Despite huge and continuous investments on building and adapting
documentations, we always struggled to get something great. Lots of
books have been written, content has been adapted to every new version,
lots of different documents / platforms, ...

We mostly failed in:
- focusing everyone to work on the same direction rather than having
  every contributor that do stuff in his own way.
- getting the right tool/platform; loved by contributors, clear for
  reader, accessible for developers

The challenges are not the same according to the type of doc:

We have 2 issues with the functional doc:
1/ Odoo is evolving very quickly. Something that should be explained
   in a version will be improved in a future version. It's like
   building on quicksand; every year you have to rewrite a large part
   of the doc.
2/ The documentation to cover is huge, because Odoo is huge. You can
   write around 1000 pages (current size of the doc) and it will not
   answer 50% of user's questions.

For technical and designer doc, we don't have such issues. The framework
do not change too much year over year.

Our priorities

The designer and developer documentation is a key priority. We started
to rewrite all the documentation to adapt it to v8. You can follow the
work in progress in this branch: 8.0-newnewdoc-xmo

We plan to write, for developers:
  1/ A tutorial
  2/ A reference guide
  3/ An API guide

For designers:
  4/ A designer tutorial

We expect to release at least 1/, 4/ and partially 2/ within around 4 weeks.

For the functional doc, we decided we will put most of our focus on
making the software easier through improved usability, on boarding tips
and configuration/deployment tools for every app. Example with this
mockup: (click on the top green zone)

In other words, we will invest time so that users do not need a
documentation any more. (a status we will never reach but will tend too)
I think the pay off will be better if we put 500 man*days on usability
tools rather than 500 man*days in writing documentation or a balance
like 400 man*days tools+100 man*days documentation.

So, functional documentation is not our priority and we know we will not
be able to create a complete documentation by our own. But we would be
very happy to help the community succeed in building something if you
are ready to seriously contribute.

So, the only way to build a great documentation is to have:
  1/ a lot of contributions (like on
  2/ with a good coordination / leadership to align everyone on the
     same direction.

We can help for 2/ and a bit of 1/. But we will succeed on 1/ only if we
succeed to create a strong translator group.

To help you measure the effort we put in the documentation, some numbers:
  - Sources v7 (.py): 858.000 words
  - Doc v7 :          209.000 words

And these 200k words have to be reviewed every year, at every new
version. This is why we really have to choose our priorities. It's a
super big task.

Moving forward

For the functional doc, we tried to build how-tos based on the Q&A
platform. Here is what have been written after 30 man*days by
experienced people:

With the feedback of this approach, I am still not convinced it's the
way to go. A traditional Sphinx + clean CSS + github would do the job

We would like to have the documentation on, not on a separate
website. So, we will use the sphinx template of XMO (one based on the
bootstrap documentation). We implemented a clean CSS rendering and we
can connect it to transifex through github (though I think translation
is not the priority, to not spread efforts too much)

I prefer the bootstrap template over Readthedoc because of the Afix
module that keep it available so that you don't always have to scroll to
the top to navigate the ToC.

So, next steps:

1/ We will finish our technical doc and communicate on it as soon as
   it's ready to be published

2/ We created the "Community Translators" mailing list to start
   organising discussions and responsabilities amongst translators
      Join here:
      Write to:

3/ We will move our efforts on the technical doc to a separate
   repository with public write access. We initially put it in a
   odoo/odoo branch fork. But, for access rights, it's better to have
   all docs in a separate repository.

4/ Can someone contribute to restructure the existing documentation
   into a new and clean structure? If someone is ready to do this
   (around 20 hours of work), we can discuss together on how to do it.

5/ Once we will have the structure of the documentation, we will be
   able to assign responsibilities on writing / improving chapters.

6/ We will clean the platform and add missing features from readthedoc
   but we will keep the design of bootstrap. (the only things we
   need is to connect transifex to the github repo and the Edit in
   Github button)

The focus for the new documentation should be on a tutorial approach (to
help new users configure/deploy/use) and not a reference guide that
answers all questions (it's just too big).

Side note

I read in the thread that some people think having a bad documentation
was a strategy to sell services. (revenue model) To be clear: it's not
our business model and we think it's a stupid approach. Our approach is
to build an exceptional product and help to on board new users (through
better UI, docs, ...) We never had plans to sell docs or support services.

We sell Odoo Enterprise: upgrade to benefit from new features, bugfix
guarantee and optional hosting. We also sell trainings, but it's not key
to our business model. We sell trainings because people need them
(requirement to have new partners/users) but scalability of trainings
are such that it's not good enough to sustain our R&D efforts.

We have plans to sell support contracts but I am convinced that the
better the documentation is, the more support contract we will sell.
(because Odoo will have more users with a better doc)

Fabien Pinckaers
Odoo Founder

Phone: +
Twitter: @fpodoo

Instant Demo:

Post to:

Raphaël Valyi
Founder and consultant
+55 21 3942-2434