RE: The Perpetual Myth of iSeries Obsolescence

>>  Jim-
>> If you're not going to re-engineer the application to use modern UI
>> principles and you're not going to exploit the rich opportunities for
>> application improvements, you're just wasting time and effort.
You're
>> then binding your applications and businesses to complicated
technology
>> and introducing considerable back-end costs and drag without good
>> reason.

>  Tom-
>I am not sure if I agree with this or not.  Clearly, re-engineering
>applications solely to convert to GUI interfaces is a waste of time
>and money, unless the UI improvements improve the application. 
>However, if it is being done to position the application for future
>improvements, it may be a good investment.  

There may be scenarios where upgrading green screen to GUI positions the
application for future improvements, but I think it's a marginal step on
a long journey.  Re-engineering an app to truly take advantage of modern
UI principles is often pretty much like starting over.

>Also, I don't think that
>there is a whole lot of additional cost for the vehicles for a modern
>UI.  Now Websphere is expensive, but there are alternatives.  CGIDEV2
>is free, the incremental cost of serving .NET is pretty small, and
>Tomcat is free.

Linux is free too, unless you want to run with some degree of support in
a production environment.  You're either going to subscribe to something
like Red Hat or you're going to have to staff up, or both.  I do think
that there are a hell of a lot of additional costs for the vehicles for
a modern UI, if you're starting out as a traditional green screen shop.
Many of the costs for serving .NET are small if you're already support a
large Windows back end for file and print serving, and for mail.

I was referring to back end costs -- managing and administering various
technologies, DBA's, sys admins, web admins, etc.  You can run a pretty
large business on a green screen platform without much more than a
part-time or full-time admin and some operators.  Bring in Windows
servers, Oracle or SQL Server, Unix, and web server components and the
skill sets start to rise.  The costs for developers also rise, and the
development staff has to encompass more roles as well.

>> I personally don't believe that there is a majority of folks out
there
>> developing to good, modern UI principles.  For every decent,
>> productivity-improving GUI app (in my opinion) there has got to be a
>> dozen or so mediocre collections of forms and programs developed by
>> "programmers" without much true application design education or
>> experience, and developed without the benefit of business analysis.

>Agreed, but irrelevant.  First, remember that developers have to walk
>before they can run.  Second, there are as many crappy green screen
>apps as there are crappy GUI apps. Writing off GUI development because
>many of the apps suck is muddied thinking. 

Irrelevant is a strong word for an open discussion.  Also I wasn't
writing off GUI applications.  I think you're looking at this as purely
two-sided discussion.  I'm not advocating staying with green screen and
banishing the thought of GUI applications.  I'm saying that if you're
going to go to the effort you should achieve some real return to offset
the extra ongoing costs.  Actually I'm defining it from the negative --
folks who are going to GUI without comprehension, planning, or skills
are merely adding ongoing infrastructure expenses, and are migrating to
platforms which are more likely to require complex upgrades or rewrites
in coming years.