Re: Adobe's RIA Technologies

You know what would be great? (I am guessing this idea isn't original and
there is probably already something out there) It would be great to have an
application spec that each framework could implement (i.e. simple order
entry application with shopping basket for end customers, internal screens
for "wizard-less" shopping baskets, item looking, pricing module, customer
lookup, PDF creation, email notification, etc). That way each framework
could go through the motions of implementing all of the same features and it
would be much easier to see the complexities of each framework.

Anybody ever heard of such a thing?

Most of the frameworks out there only go to a certain extent showing you
patterned implementation strategy (i.e. how do I write application xyz with
JSF/PHP/RoR/.NET/RPG) by showing you how to do things like submit a single
form, occupy a dropdown, display a result set, etc. But they don't give you
a solid idea of how to make all those work in unison on the same page in a
multi-page process. I think this is partly because once an application is
written in xyz framework the creator can't distribute it because of
liabilities/copyright and that is where a public spec for a simple
application addressing complex capabilities would be great.

Aaron Bartell
http://mowyourlawn.com

On Fri, Oct 31, 2008 at 11:37 AM, Niels Liisberg <liisberg@xxxxxxxxxxxxx>wrote:

Aaron;

You are absolutely right. I think we are on the same page here. I was
actually addressing the UI more that the server side code - even though I
was referring to it as one .. my fault not being to clear ..

One interesting thing though in that discussion: When we are making
statistic on how much time we are using in our company divided up in the
MVC
model it is rather scary

Model : 10%
View: 80%
Control: 10%

(This is for normal customer web 2.0 application using IceBreak and ExtJS.)

My guess is - if you go back to a good old 5250 application you will find
that that the same 80% of code is used to manage the interaction with the
displayfiles. If you separate the business logic and data access you only
find around 10% to 20% of efficient code - give and take …

The idea of just picking the UI of the day for fun can be rather time
consuming = pricy. So - Yes - of cause you need some kind of long term
strategy.

So before diving into Adobe's Flex or decide to give ExtJS a try - you
better make some prototypes and ask if they are funky enough for years to
come - or say - how soon will they be outdated?



Best regards


Niels Liisberg
IceBreak Chief SW Architect