openerp-community team mailing list archive

[Raphael Valyi <rvalyi@xxxxxxxxx>]

On Mon, Jan 20, 2014 at 7:13 PM, Kurt Haselwimmer <kurt@xxxxxxxxxxxxxx>wrote:

> Dear Raphael,
>
> [...]
>
> Say I create a new feature for the webstore, say even - the shipment
> tracking feature I described - whereby I can store parameter trackno,
> together with a courier choice, such that the URL
> www.courier.com/tracking.asp?id=%s where %s is trackno, so that I can
> simply link through to a courier's website to see how far my parcel has got
> yet, and even if the customer has received it - this is not a USD5000
> development - but I bet others would be prepared to pay $50 for this simple
> and useful functionality. Why then can I not publish the code on the
> launchpad, thereby keeping the AGPL purists happy, but ALSO offer a service
> to install it on customer's servers for $50 ? Just because the code is
> published freely and for nothing, it does not seem to logically exclude the
> possibility of also charging a fee to help those people who don't want to
> have to spend days trying to find the exact piece of code on the launchpad
> that they should be using.
>

Honestly that  doesn't work. So many pure open source project just don't
pay the bill.
Did you hear last week OpenBSD could shut down its integration servers
because the energy bill was to high?
It can exceptionally work but not enough to make it a valid business model.
What makes a valid business model is pay to get some code, either for the
1st time, either in a recurring way if the license allows it.

Worst than that: what we see that a lot in the OpenERP ecosystem is: the
"leeches". That is the guy who is going to publish the extension you
developed is in fact your competitor who didn't invest a penny on the
development but instead put all the money on the marketing to fool people
and make them believe they are the providential guys with their extensions
that in fact they didn't develop. And this totally suck out the initiative
to invest in new development.
This is typically a Prisoner's Dilemma game
http://en.wikipedia.org/wiki/Prisoner's_dilemma

Now in OpenERP you still have thousands of modules are absolutely free of
charges. Not all are good. But there are hundreds of very good ones, much
better than the paid and closed ones of open core platforms.

Consider these cases where the OpenERP copyleft model rocks:

A) you have a simple project where you can use all the free features. Then
it cost you zero development, or just the consulting or migration work.

B) you have a very complex project where you want to do what nobody does
Then you can reuse all the rich peer-developed free eco-system you wouldn't
have in an open core model.
And for your very specific features, you should still pay as you would
anyway in an open core model

Now I agree, there is an in between case between A) and B) where probably
an open core or even proprietary models would be more competitive in the
short term. Well open source doesn't fit all use cases. May be tomorrow we
will have 100% open source computers. As for now we buy computers that
still have non open source piece because the 100% open source ones aren't
mature yet, just may be like OpenERP isn't mature yet for any user case.

Well, no model is perfect.
But it's important to have in mind the shortcomings of open core models
before judging.  We typically see this in an open core model such as
Magento:
many closed source modules are of very poor quality past the surface,
because nobody peer-review them, no module reuse an other one from another
editor...
So you have plenty of useless gadget modules, lot's of marketing noise
around them preventing you to get the "signal" about what really fits your
need (you know the buy button of WinRar..). Ambitious projects should buy
many of these badly integrated independent modules and for the most advance
needs they should still pay to get their feature developed properly because
the modules of the market place are too broken to built upon.

Also people buy many times the same development (because it's not
copyleft). And at the end of the road, everything goes to trash when the
platform starts to looks old and cannot update because it's not a community
open source project and the investment cycle on which it has been built is
already burnt away, without becoming a free public commodity such as
copyleft projects.

The reason it is relevant right now - is because there are large numbers of
> small changes and customisations that webstore designers want to have, and
> would be prepared to pay for, but which may never get developed because you
> are suggesting that the AGPL does not allow it. I think a clarification of
> this point would be really helpful for me, and perhaps for anyone else new
> to this particular form of license.
>

This is probably yet another of the reasons Magento or Prestashop may not
be wiped off f the map by OpenERP ecommerce as soon as it's said ;-)
An other possibility is OpenERP SA or some other integrator take the
speculative risk to develop some of these extensions upfront to attract
people to their platform or services. But as I said, upfront investment to
reach market position has less fuel than an open core model so in any case,
many things will only come with time.


NB: and yeah, capitalism has taken over open source, that's a fact now.


Regards



-- 
Raphaël Valyi
Founder and consultant
http://twitter.com/rvalyi <http://twitter.com/#!/rvalyi>
+55 21 2516 2954
www.akretion.com