@fabien,


> Sometimes I feel like I don't understand how you think  Of course our 
> employees are part of the success of OpenERP


The way I reason is simple: you hired all those guys while you had no money to 
attract the best talents because like all of us, you were doing integration 
with a half mature product (eg tough profitability).
Now you have those 80 guys, but your mission has suddenly changed: we don't 
need you to develop screens and reports massively anymore, we need Tiny to be a 
real editor, we need you to rather enhance the framework, develop a powerful 
widget system or develop a decent report engine (or just jumpstart on 
report_relatorio it will be enough).
By that time, you lost the Tryton devs, you even lost Antony Lesuisse which 
thank you god is back now you have some decent work condition to offer now you 
have the money...
80% of your guys were working on integration, fine but was that such a 
resounding success? We know feedback from Indian employees as well as Tiny 
support customers, or Odoo maintainers and I'm sorry but I don't call that a 
clear success (but I wont make the offense to detail that publicaly) you can be 
proud to be part of. Let's say we have a different judgement, fine.


@fabien, @amit

Now, don't get me wrong, I don't want to offend people. You know what: I 
consider when I had less than 2 years of work experience I WOULD NEVER HAVE 
BEEN A VALID OPENERP INTEGRATOR NOR EDITOR BECAUSE I DIDN'T HAD ENOUGH SKILLS 
MYSELF. And even now, I would never claim I hold enough Python coding skills to 
develop the core server all by myself. Now, I perfectly see what bad code is 
(specially large methods, un-explicit variable names, no comment, bad 
dependency management, sub-optimal use of the type system, bad inversion of 
control etc...)

So, in 50% of the time, it's not because the guys aren't smart enough, it's 
rather just that they have not enough experience for the incredibly hard 
challenge to develop the next generation ERP while sticking to mission critical 
coding standards.
Both of you know that SAP or Openbravo hire their core product employees much 
more experienced guys than Tiny used to hire, specially the average guy in 
India.
Where are the domain experts that are required to build an ERP? Who at Tiny 
holds say an accounting diploma? (On the contrary you have those guys on the 
expert mailing lists you weren't even reading)...


@amit, I will never say that 100% of what Tiny did in India is crap. But sorry, 
I still go for a solid 50% and unfortunately this is way to much for the open 
source world. For instance the overall idea and even impl of the web-client is 
great. Now, can you imagine how much will cost you the choice of Mochikit? How 
much will cost you the refactoring now that Mochikit won't follow the new 
browsers (Webkit ones based already have issues) neither HTML5...
What is the long term return/re-usability of:
Wiki, OpenETL, ERPMatrix, server-sa, all the one shot specialized modules that 
are polluting extra addons, Dia stuff, training module, dm module, mrp_repair 
when you have no module for basic RMA and mrp_repair has not been designed to 
be compatible with such an eventual module...
I mean all that is not what we expect from an ERP editor which is firstly a 
decent accounting (and we are still far from it, especially in India BTW).

And how you want to talk individual productivity?
All right, here is the product quality the 2 Tryton core developers can achieve:

http://hg.tryton.org/hgwebdir.cgi/1.0/trytond/file/e89c7f614279/trytond/osv/orm.py

http://hg.tryton.org/hgwebdir.cgi/1.0/modules/account/file/ddba56e4fd64/account.py

vs what quality the 80 employees of OpenERP can achieve (low code quality but 
even spend time/money to develop modules to attest that low quality and dare to 
call that "official quality certification" when partner module have way better 
results than Tiny's one...):

http://bazaar.launchpad.net/~openerp/openobject-server/5.0/annotate/head:/bin/osv/orm.py

http://bazaar.launchpad.net/~openerp/openobject-addons/5.0/annotate/head:/account/account.py

Seriously...


And here is how that Indian """quality team""" find it natural to fix a web 
client flaw by hardcoding different behaviors right into the account module 
according to the client:
http://bazaar.launchpad.net/~openerp/openobject-addons/5.0/revision/2498
And this is only reverted because us, poor mortal integrators shout on Twitter 
to try to reach Tiny which was not listening at all...
Seriously, that kind of commit would be enough to be fired in other companies 
that edit such mission critical software...

So I mean I totally respect that everyone starts by being a noob. And yes there 
has been some great stuff and yes in India too. But now, not facing the 
problems (and there are a lot) is the best way to not fixing them...

Hope I'm talking to responsible people able to let their attitude behind to 
spot the problems and fix them rather than just ignore them or fake to ignore 
them.

------------------------
Raphaël Valyi

CEO and OpenERP consultant at
http://www.akretion.com




-------------------- m2f --------------------

--
http://www.openobject.com/forum/viewtopic.php?p=51471#51471

-------------------- m2f --------------------


_______________________________________________
Tinyerp-users mailing list
http://tiny.be/mailman2/listinfo/tinyerp-users

Reply via email to