← Back to team overview

openerp-community team mailing list archive

Re: proposal to discuss forking of OpenERP GTK client

 

2012/10/11 Fabien Pinckaers <fp@xxxxxxxxxxx>

> Hello,
>
> We would be glad if someone wants to become the official maintainer of
> the GTK client. We can transfert the rights to the current trunk to a
> team or ~openerp-community.
>
> But in my opinion, there is nothing in the GTK client that the web
> client can not do:
>   - the web client is faster (especially if not standalone in v7.0)
>

This is subjective and also depends a lot on browser config/crap in my
opinion, the gtk suffers less from the enviroment because it has its own
"runtime"


>   - there is nothing that the web client can not do; as en example, the
>     v7.0 POS does barcode, touchscreen, non connected mode, ...
>

Printing the current workflow state of an object, proper handling of pdf
files (webclient still messes up mimetype which makes chrome not recognize
it as pdf).

  - the web client has shortcuts (in v7.0)
>

Not even sure what you mean here ?


>   - the v7.0 web client is so good that nobody will want to use GTK
>     anymore
>

Ask anyone who has to enter a bank statement if they prefer web or gtk,
good luck with that !


> But, more importantly, the web client is modular. It means that if your
> accountant or warehouse user need something, you can develop it easily.
> A community member, Colin MacMillan, developed tabs for the web client.
> It took him 16 minutes of development:
> https://twitter.com/ColinMacMillan/status/236788938331791360
> As it's a module, you can use tabs or the breadcrumb at your own choice.
>
>
I understand the modularity is a big plus but let's be honest these could
quite easily be made as plugins
for the client if the plugin system would be more robust and developed. You
could even stream the required plugins
from the server to the gtk client and "patch" your gtk client depending
where you login.


> The reason why we want to stop our efforts on the GTK client is that we

prefer to develop one UI which is great, rather than two UI that are
> halfly great.
>

Can agree on that and market pressure will probably tilt in favor of the
web.
But when you lose your "ace" desktop client you become one more python
framework
and I don't think I have to tell you how much competition is there, not to
mention
most of those frameworks have a more consistent api interface which are
already
production proven (think Django / Plone / ...)

So, let us know if someone want to maintain the GTK client so that we
> can transfer the rights on the trunk branch. But if you want to do it,
> be ready to assign 5 full time developers on it. Otherwise it will not
> be possible for you to follow the future evolutions and you will be
> trapped on old versions, just because you mis-evaluated the web client
> potential.
>

A lot of time could be saved by drafting a client API specification as I
have already
requested earlier (see
http://nightwalkers.be/openerp-open-letter-api-specification)

If we have one solid specification it would be save a lot of time and
prevent a lot of bugs
and we could actually start writing tests that would emulate client
behaviour and thus
provide more code coverage.

-- 
Niels Huylebroeck
Lead Architect   --   Agaplan
Tel. : +32 (0) 93 95 98 90
Web : http://www.agaplan.eu

Follow ups

References