RE: How to move forward -- MIDRANGE-L

> From: Pat Barber
> 
> I ain't laughing.... I been wondering the same thing.
> 
> Based on the responses, I would say this is a real concern.
> 
> I noted that almost every single reply gave a "different"
> way to "move forward", which is exactly what this "no standard"
> method I see in this whole thing.
>
> If you hear of a solid, well documented, proven solution, let me
> know.

I've avoided this conversation up until this point because it tends to
bog down into that semi-religious "mine is better than yours" type of
discussion, and I've been involved in too many of those over the years.
However, there is a reasonably simple answer depending on one specific
business issue.

That issue is whether or not you plan to stay with the iSeries, and if
so, if you plan to stay with IBM's strategic direction.  If the answer
to both questions is yes, then it's absolutely clear what your choice of
tools should be: WDSCi to develop JSP-based applications to run under
WebSphere.

Ask any of the development groups at IBM, and the entire focus of the
tools teams is on developing for JSPs and WebSphere, and the great
majority of the budget for tools development is being dumped into WDSCi
(and the rest of the WebSphere development suite, such as WSAD and
WSSD).

This is not an opinion, not a "nice to have", not a marketing gimmick,
not a pet language.  All IBM development is geared towards this brave
new world, and whether you agree with it or not, any money you spend on
other solutions is going to be wasted.  This includes screen scrapers,
CGI, Net.Data, Python, Perl, and .NET.  Just like I'm no big fan of SQL,
but I know I have to embrace it as part of the strategic direction, so
too is JSP, especially JSP Model II, IBM's architecture of choice.

There are still a ton of issues here.  Hans' points on XHTML and CSS
barely scratched the surface of the web design question; issues such as
Section 508 accessibility, JavaScript, and document layout are all still
being hammered out at this time.  But it is clear that IBM has chosen
WebSphere and JSP as the strategic solution to deliver that information
to the browser.

The only real competition at this point is the .NET architecture.  But
because of the Microsoft lock-in and the obvious lack of working
relationship between the two behemoths, it comes down to a matter of
which side of The Force you choose to join - Redmond or Armonk.  I leave
that choice to you.

All of the other approaches, from RPG-CGI to Net.Data to the many
Unix-based specialty languages, are little more than stopgaps.  None
have reached any level of widespread acceptance.  This statement might
be challenged by some, but rather than rely on zealous advocacy, you
should instead do some research to see how many businesses in your
particular market niche use any given technology.  Also, all of these
lesser known solutions tend to be at least somewhat proprietary.  Sure,
you can run Python on an iSeries, but who's going to support you when
your website crashes with an unknown error?  Not IBM.

On the other hand, all of those niche technologies are very good for
learning, for pilot projects, and for non-strategic applications.  They
allow you quick results with a limited investment, and you can get
started immediately.

Of course, the cost issue is no longer a big one these days with
WebSphere Express and WDSCi; $2000 gets you a working copy of Express
and the WDSCi tools come with the compiler.  The learning curve is
longer, but the strategic positioning is unassailable.

And that's my .02.  Remember, I sell a JSP based product, so I'm biased.
However, I've been selling a JSP-based product for years now, and I'm
still selling it, so it must be an architecturally sound choice.

Joe Pluta 
www.plutabrothers.com