RE: Web enablement options (previously private)

["Joe Pluta" <joepluta@xxxxxxxxxxxxxxxxx>]

> From: jkrueger@andrewscg.com
>
> Using the Jacada Innovator technology, the RPG applications are
> converted to
> batch server programs with CL wrappers.  The workstation calls in the RPG
> applications are replaced with messaging calls that communicate
> with the new
> thin-client code through an intermediate layer called the Jacada
> server.

This is an excellent summary, Janet.  In fact, it's an excellent summary of
the architecture that was outlined in my eDeployment book, the architecture
that both Innovator and the JACI/400 product are based on.  The book was
published in October of 2000, and basically outlined this entire process,
but of course I'd given discussions about the idea at COMMON and OMNI in
1999.  Still do, as a matter of fact, even though I don't think Jacada even
showed up at the last COMMON; I don't recall seeing their booth, or
Seagull's for that matter, while I was manning the Linoma Logic booth.

I'm not sure how Innovator handles various things, but issues to look out
for and address are:

1. Hidden fields
2. Support for PAGEUP/PAGEDN and SFLFOLD/SFLDROP (without returning to the
program)
3. Support for INFDS
4. Softcoded (i.e. via CSS) support for attributes
5. Support for message subfiles
6. Support for upper case only and numeric shift
7. Support for autorecord advance
8. Support for modulus checking
9. Support for embedded blanks and attributes in fields
10. Ability to change appearance dynamically based on indicators

These are just a few of the issues that begin to come into play when you're
handling a graphic enablement of a 5250-based program.  A few other high-end
issues:

1. Full command key support (F-keys, roll keys, etc.)
2. Support for the back button
3. Clean timeouts, both browser and application
4. Connectivity to other web applications (and back)
5. Support for 5250 emulation mode (fixed font, traditional colors)


> >> If it's the later, how much is automated and how much is manual effort?
>
> The conversion from the interactive applications to batch is completely
> automated -- Jacada does recommend that all applications be
> completely retested
> after the conversion, with a statement that at least 95% of the
> conversion will
> have been done correctly. I don't think much fix up is needed on
> the RPG side,
> and in general, the converted programs end up being just as
> maintainable as the
> original programs were.

With the goal, of course, being 100% conversion, as PSC400 does.  PSC400
does complete conversion of programs, with no manual effort, and at the same
time automatically generates the 5250 emulation JSP and the original "gray
screen" JSP, ready to run.  Not only that, it's all done from a single
AS/400 command.  No workstation required, no workstation software, no
graphical interface.  A single command, CNVPGM, and you're done.


> Most of the manual effort will be involved in
> determining how the new user interfaces should look and feel,
> unless you're
> satisfied with just having web delivery of screens that look/feel
> just like they
> did before.

Which, unless you're using a 5250 emulation mode, is not pretty.  At least
the 5250 emulation mode, with fixed spacing and all the colors you've grown
to know and love, looks just like your original screens, and so users don't
get culture shock.  Converting existing screens to proportional fonts (or
worse, a combination of proportional and non-proportional fonts) without any
other work is really, really ugly.  But on the other hand, if the tool gets
you started (by putting things in tables and rows, like our gray screen JSP
does) then it's really not that difficult to add your own look and feel.
And once you've done one, applying it to others is very easy.


> The actual amount of effort will depend on how much
> variety there is
> in your DDS panels -- if they adhere to fairly clear and
> consistent standards,
> it will be relatively easy to build a new standardized look and
> feel that fits
> naturally into web delivery.

In the best of all worlds, the tool would take your customized look and
feel, build a template from it, and use that on subsequent conversions.
We're not there just yet, not sure how the Innovator folks plan to do it.