Re: Development ideas -- MIDRANGE-L

inline also

Adam Lang
Systems Engineer
Rutgers Casualty Insurance Company
http://www.rutgersinsurance.com
----- Original Message -----
From: "Pat Barber" <mboceanside@worldnet.att.net>
To: <midrange-l@midrange.com>
Sent: Friday, October 18, 2002 2:48 PM
Subject: Re: Development ideas


> Comments inline:
>
> Adam Lang wrote:
> >
> > You are overlooking the major issue with developing a critical app on
> > Windows technology.  Microsoft breaks their tech every 2 years.
>
>      This is either "new" or "old" news depending on your perspective.

I said this in regards to put in a recommendation against MS technology for
the backend.

> I thought you supported writting a "new" application in Java ???

Yes.  I like Java.  But i was trying to be unbiased and say yes, some of the
same issues with developing in VB are there for Java, ie a private company
has ultimate control.

> Yeppp, you want to use a language everybody gets to change when
> they don't like a feature.

But atleast you have an ANSI standard to back you up.

> I don't know that number and I suspect, neither do you. There are huge
> numbers of COBOL programmers world wide. There are also huge numbers of
> C++
> programmers, I guess. Computer related degrees in schools are now not
> in fashion since that little "dot.gone" thing of the last year or two.
>
> Nobody wants to be a programmer anymore, since it doesn't allow you to
> become a zillionaire in two years or less.

Well even so.  If comp sci degrees are nto fashionable, how much even less
are COBOL majors?

>       They said the same thing about PL/1 a few years back. That
>       didn't work out quite like everybody wanted. This was a first
>       cousin of C++ and few other language thrown in for good luck.

Well, C does have a 20 year track record of being around as well as an open
standards (ANSI) body supporting it.  It is still wildly supported by all
somputing platforms (hell, you can write programs for Lego Mindstorms in a
C).

COBOL has a long track record also.  Just not as wildly supported (not
saying it isn't)

> I recall Fortran having a similar calling when I was in college. This
> was back in the "neanderthal" period of computers. We punched our cards
> by candle light.

And it lasted a long time, you could say.  Granted, 20 years from now C
might be in the state Cobol and fortran are. No new share, but lots of
developers supproting.  But at this current point in time, I think you could
argue COBOL developers and support is waning a lot faster than C/C++
resources.

It is also arguable Java could be a flash in the pan compared to these other
two languages, in terms of longevity.  Again, a lot of that is hinged on Sun
being able to let it grow unlike C/COBOL where standards bodies control it.

Java probably only has gotten as far as it has because of the Internet.

> I think you just described a 5250 data stream using pc's...

In essence yes.  But in genral, people like GUIs over green screen.  It is
also a lot easier to sell.  Hell, when I write apps, I would love to be able
to get away with making the interface jsut labels and text boxes and a
couple command buttons.

multi-tier development is not a new strategy.  From what I see, the
AS/400-mainframe was doing it since the 80s.  Only difference is, it could
handle all multi tier functions on one box, whereas in the PC revolution,
you need a desktop for the client, another server to host the business logic
and another to run database.

What someone mentioned on another post is very valid.

You break up your project into logical and modular sections.  You see what
you want to do.  THEN you find the technology that best solves the problem.
Not find a technology and make your application to it.

So in essence say this project is broken into, for simplicity sake, data,
business logic, client.

After lookign at the client piece they see the itneface is pretty simple and
straightforward, they figure they can make a nice java app that will run in
a browser.  They see it as easily supported and will address and local user
they woudl posibly market too.  As well as not scare off any possible people
by seeign a "green screen app".

They analyze the logic end of it.  They see it will be crunching large
chunks of data and othe functions that an AS/400 is the best solution for.
They now have to decide on language.  They have to weight against RPG being
supported for the forseeable future against ever possibly movign to another
platform (so using C or Java). As well as weigh in the current trainign
ability of programmers and getting future help to meet demand and turnover.

Then they analyze the data end of it.  They always liek DB2 and have
experienced peopel for it and have always liekd it, so they go with that.

It all comes down to defining what you want to accomplish and THEN seeing
what tech is out there to support it.