Re: Creating a Web Service with COBOL...


"softWare design" <sabraham@xxxxxxxxxxxxx> wrote in message
news:1182910437.457329.69930@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
On Jun 26, 6:30 pm, "Pete Dashwood"
<dashw...@xxxxxxxxxxxxxxxxxxxxxxxxx> wrote:
"softWare design" <sabra...@xxxxxxxxxxxxx> wrote in message

news:1182888614.062525.103150@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx


Your links didn't work for me, but I have a fair idea what you're talking
about...

Pete.


Pete, thanks for your comment. See if you can access the links now.

http://216.239.51.104/search?q=cache:JFZTMPGZ5W0J:www.cobolportal.com/files/egilityedge.pdf+cobol+web+service&hl=en&ct=clnk&cd=10&gl=us

OK, I read this. Normally, I'm not too keen to waste time reading marketing
stuff but most of this was sound, and it was generally sensible. I went off
it when they quoted the usual Gartner bull*** from 7 years ago at the
start, but I persevered and was generally pleased to see they are advocating
component based solutions and recognise the importance of Web Services. I
certainly don't disagree that leveraging existing COBOL into these
environments is a useful thing to do and you can't blame MicroFocus for
seeking to extend the useful life of COBOL in this way.

I think they are still going to hit the paradigm shift but at least existing
stuff will be salvageable. This is of critical importance for many
companies.

http://216.239.51.104/search?q=cache:kyloOiakJUMJ:www.adtmag.com/article.asp%3Fid%3D9472+cobol+web+service&hl=en&ct=clnk&cd=1&gl=us

This one was a bit of a hodge-podge. Some pretty hyperbolic marketing claims
("we are the ONLY people..." etc. when, in fact, if you're prepared to do it
manually, ANYONE (with access to a network and (easily acquired) knowledge
of XML/WSDL can produce and consume web services.) It was lengthy, but
there were some interesting nuggets in amongst the dross. The main value I
see is in raising awareness in mainframe world that they better get with the
action or they will be left behind. (However, "getting with the action" is
not as difficult as some people who are trying to market products, would
have you believe...)

http://216.239.51.104/search?q=cache:tR7JwXw62jQJ:www.adtmag.com/article.asp%3Fid%3D7730+cobol+web+service&hl=en&ct=clnk&cd=2&gl=us


Not a lot of substance here and mainly a vehicle for MicroFocus to reprise
the catchphrase of being the last chance for COBOL. (Nothing against MF;
they very possibly are...) The company's commitment to the future of COBOL
was restated (at least they see a future and are prepared to state it; more
than the COBOL Guardians at J4 were able to do...) and they are doing what
they can to see that COBOL gets to play in the brave new world of SOA. I
hope it goes well for them. I believe their efforts could well extend the
useful life of COBOL, but I'm still not prepared to extend the 2015 date...

The problems they will run into are:

1. Unless they can actually persuade the current user base to move to OO
(and so far, that hasn't worked at all), their customer base will be eroded
as more and more shops realise that maintaining computer programs line by
line is just far too expensive and opt out. (As someone pointed out in one
of the above articles, it doesn't matter whether you are using Java, C++ or
COBOL, as long as you are tied to source code line by line, you're screwed.
The only reason more haven't already done so is because they are locked into
their current COBOL investment. (Or think they are... actually, they are
not, but as long as they continue developing with the status quo, there will
be no progress. If there is no real progress, eventually the cost of
in-house development will just get too expensive and the whole development
department will simply collapse.) At this point companies outsource IT, or
outsource a package solution, or simply buy the service.

2. Although they can provide bridges to CICS and IMS that will allow web
services to be accessed from the mainframe, unless there is a radical
rethink in the way things get done (development methodology, if you
like...the Waterfall is still "state of the art" in many COBOL shops), no
actual benefit will accrue. There's no point in having access to components
if you don't understand how to build component based systems, and reuse
existing objects. The whole concept of network infrastructure and object
reuse is beyond the ken of most existing COBOL shops. Basically, they don't
see anything broken, so there's nothing to fix.

3. As more sites DO start to cotton on to the real advantages of SOA they
will realise that, even if COBOL CAN be used for OO it still isn't as facile
as other languages that were designed to do so from the start.

(I loved OO COBOL and used it for nearly 10 years.Built loads of reusable
components and tools with it, and thought it was great. But the simple fact
is that in an OO environment, C# is simply light years ahead of it (so is
Java). I wouldn't go back to OO COBOL development now, and I use COBOL only
for existing stuff until such time as I can get around to rewriting it.) The
point is that as knowledge grows, and the scales fall from people's eyes,
COBOL is revealed as a less attractive option, in the 21st century
environment we are faced with.

Given all of 3 of the above (which can be summed up as the "paradigm shift")
I believe MicroFocus are on a hiding to nothing long term.

Hope I'm wrong, but current trends seem to be bearing out what I say.

The three articles are well worth reading, but remember there is sludge in
with the gold dust... :-)

Pete.


.

Quantcast