> I agree that speed entry is a major problem in the current web client v5. 
> Having editable grids (see my previous posts) and avoiding superfluous clicks 
> when creating new one2many items are the best way to fix this, in my opinion. 
> In the mean time, users who want faster entry are stuck with the gtk client 
> (which only works efficiently on LAN).


I agree with all those points and consider it a hot issue, way hotter than 
relooking indeed...

Moreover, here are some other extra ways to require less user clicks:

- make dynamic domain work on selection widget: in a lot of place where 
many2one widget popups are used, the user will never want to create a new 
resource and there are only a few choices (the product configurator is a good 
example of it). So it's important to be able to swicth the many2one widget to a 
dropdown (selection), avoiding one click and one slow popup with a server round 
trip.
https://blueprints.launchpad.net/openobject-server/+spec/dynamic-domain-on-selection-widget

- make it easier to create a new resource in just one click on many2one 
widgets. For instance for creating a new partner in a quotation form, since the 
edit partner button is disabled when no partner is selected why the hell can't 
we have instead a new button to create a new one in one click instead of 2? I 
believe it was like that in early 4.x Tiny versions, why did that change?

- currently forms can pass parent form params to on_change methods that can 
themselves alter the form content, fine. However, it makes no sense the 
on_change can not also change parent forms. Again the product configurator is a 
great illustration of that limitation: once you selected all dimensions values, 
the matching product variant can't be selected automatically because you were 
triggering on_change in an inner form.
https://blueprints.launchpad.net/openobject-server/+spec/form-parent-field-update-on-change

- new osv_memory wizard lack a few miles to replace old non OOOP wizard 
completely (can't be on the right side or apply to a group of record). This is 
sad because new osv_memory wizards support on_change which can save some clicks 
to the user.

The sad part is that the little Tryton fork already fixed most of these things 
and this hasn't been spotted by Tiny as something important (no blueprint 
feedback, no feedback yet on expert teams lists).

Hope Tiny considers that important. Let fix that first instead of making tons 
of new modules with on a limited UI. Once we have a solid platform, module will 
come on a stronger basis with a more sexy UI. Also please don't let tons of new 
modules develop on sand one more year before fixing those things.

We all know that once bad things are done it costs more to remove/change them 
than re-do them properly from scratch. Because osv_memory are limited or 
because of the on_change limitations, lot's of crappy code will be made in the 
module to work around the limitations: old wizards styles will be copied over 
and over in non compatible manner... Then it might stay here for ever because 
customer X is in prod with it. So in order avoiding making X unhappy with his 
bad yet working code, it costs the whole community/future market, this is not 
how open source is supposed to advance...
So, aside form short term marketing strategy, I see no point in letting that 
happen.

Basically, I'm not saying Tiny should invest more by no mean, I know they are 
doing a lot already. I just call for considering that sort of things a priority 
and at least let the community merges enter before 5.2 regarding those points.

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

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




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

--
http://www.openobject.com/forum/viewtopic.php?p=49512#49512

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


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

Reply via email to