The most fundamental characteristic

[prasadgc]

Hi all,

I've been thinking about the most fundamental characteristic that
would be common to all the architectural variants we've been
discussing. My thoughts keep coming back to the principle of "Low
Coupling, High Cohesion" as applied to Presentation Logic and Business
Logic. Application of this principle causes presentation logic to be
factored out of the server (low coupling) and to coalesce around the
client (high cohesion).

All other aspects, i.e, XML or JSON, thin or rich client, application
download strategies, etc., result in different variants such as SOFEA.

What do you think?

If this is acceptable, what we probably need to do next is find a good
name for the fundamental architecture:
LoCHicAPABLo (Low-Coupling High-Cohesion Architecture for Presentation
And Business Logic) is probably descriptive but unsuitable.

Ganesh

[Peter Svensson]

That's a very good way to summarize it. Maybe we could call it LC/HC ?
I've also been thinking about ways to criticize existing methods and frameworks more formally. Often when we bring these ideas up a souting match ensues with a lot of he said - she said.

If we can start with this as a basis, maybe we could also build a set of provable or reasonable by statement rules from which we then begin to show why certain ways of building applications is 'better' than others.

I'm thinking in terms such as complexity, LoC's produced, how many different places are affected by a new entity, and stuff like that. Does anyone know if there has been any formal studies on these, or related kinds of things?

Cheers,
PS

[Ganesh and Sashi Prasad]

Roy Fielding's PhD dissertation on REST builds up the model step-by-step, applying layers of constraints one by one until he arrives at the REST architectural style (See page 76 onwards).

I wonder if we can follow the same approach and "prove" the model by deriving it from first principles.

Regards,
Ganesh

2008/7/29 Peter Svensson <psve...@gmail.com>

[Peter Svensson]

That could be a very good process to copy, in spirit and format.
But I think also that it would be beneficial to refer to other, existing studies in efficiency, code complexity, maintainability, et.c.
I'll dig around and see what I can come up with.

Also, reread the dissertation :-P  Argh! :)

Cheers,
PS

[Mario Valente]

Glad that you mention it. I think that the constraints/requisites
that are applied by Roy to get at the REST architecture apply
not only to the browser/view <-> webserver/bizlogic layer but also
to the webserver <-> dataserver/data layer. Thats why I believe
that the REST pattern should be applied at all interface points
of the presentation/logic/data or MVC models.

-- MV

Ganesh and Sashi Prasad wrote:
> Roy Fielding's PhD dissertation on REST

> <http://groups.google.com/group/sofea/web/Roy%20Fielding%20PhD%20Dissertation%20%28REST%29.pdf>

> builds up the model step-by-step, applying layers of constraints one by
> one until he arrives at the REST architectural style (See page 76 onwards).
>
> I wonder if we can follow the same approach and "prove" the model by
> deriving it from first principles.
>
> Regards,
> Ganesh
>

> 2008/7/29 Peter Svensson <psve...@gmail.com <mailto:psve...@gmail.com>>

[Peter Svensson]

Hi Guys. I just re-read Kris Zyp's excellent article on client-server web tech with a good rundown of implementations. He also has a list of reasons for not using traditional methods (and great diagrams :).

http://www.sitepen.com/blog/2008/07/18/clientserver-model-on-the-web/

Cheers,
PS

[Ganesh and Sashi Prasad]

Thanks, Peter. I've invited Kris to join our Google Group.

Regards,
Ganesh

2008/8/14 Peter Svensson <psve...@gmail.com>

[Kris Zyp]

Thanks for the invite, I joined. Glad to hear about your group.
Kris

On Aug 14, 12:36 am, "Ganesh and Sashi Prasad" <g.c.pra...@gmail.com>
wrote:

> Thanks, Peter. I've invited Kris to join our Google Group.
>
> Regards,
> Ganesh
>

> 2008/8/14 Peter Svensson <psvens...@gmail.com>

>
> > Hi Guys. I just re-read Kris Zyp's excellent article on client-server web
> > tech with a good rundown of implementations. He also has a list of reasons
> > for not using traditional methods (and great diagrams :).
>
> >http://www.sitepen.com/blog/2008/07/18/clientserver-model-on-the-web/
>
> > Cheers,
> > PS
>

> > On Sat, Aug 2, 2008 at 12:46 AM, Mario Valente <mfvale...@gmail.com>wrote:
>
> >>   Glad that you mention it. I think that the constraints/requisites
> >>  that are applied by Roy to get at the REST architecture apply
> >>  not only to the browser/view <-> webserver/bizlogic layer but also
> >>  to the webserver <-> dataserver/data layer. Thats why I believe
> >>  that the REST pattern should be applied at all interface points
> >>  of the presentation/logic/data or MVC models.
>
> >>    -- MV
>
> >> Ganesh and Sashi Prasad wrote:
> >> > Roy Fielding's PhD dissertation on REST
> >> > <

> >>http://groups.google.com/group/sofea/web/Roy%20Fielding%20PhD%20Disse...

>
> >> > builds up the model step-by-step, applying layers of constraints one by
> >> > one until he arrives at the REST architectural style (See page 76
> >> onwards).
>
> >> > I wonder if we can follow the same approach and "prove" the model by
> >> > deriving it from first principles.
>
> >> > Regards,
> >> > Ganesh
>

> >> > 2008/7/29 Peter Svensson <psvens...@gmail.com <mailto:
> >> psvens...@gmail.com>>

>
> >> >     That's a very good way to summarize it. Maybe we could call it LC/HC
> >> ?
> >> >     I've also been thinking about ways to criticize existing methods and
> >> >     frameworks more formally. Often when we bring these ideas up a
> >> >     souting match ensues with a lot of he said - she said.
>
> >> >     If we can start with this as a basis, maybe we could also build a
> >> >     set of provable or reasonable by statement rules from which we then
> >> >     begin to show why certain ways of building applications is 'better'
> >> >     than others.
>
> >> >     I'm thinking in terms such as complexity, LoC's produced, how many
> >> >     different places are affected by a new entity, and stuff like that.
> >> >     Does anyone know if there has been any formal studies on these, or
> >> >     related kinds of things?
>
> >> >     Cheers,
> >> >     PS
>

> >> >     On Tue, Jul 29, 2008 at 3:22 PM, prasadgc <g.c.pra...@gmail.com