RE: NT vs AS/400 -- MIDRANGE-L


Hi Nathan,

I'm glad you decided to jump in. I'm going to snip out any part of your
comments that aren't directly relevent to my response, so I hope I don't
take anything out of context in the process.


> That's true.  We have more favorable to say for Rochester when we're
> considering Windows vs. OS/400, rather than iSeries  vs. xSeries.

Actually, it's us techies that make these types of comparisons. The ones
that are writing the cheques tend to be more concerned with comparing the
cost of different solutions. In that arena, cash-flow often rules. The
higher the initial cost, the greater the impact on *short term* cash flow.
Short term cash flow is viewed as more important than long term cash flow
because mismanaging the former renders the latter moot.


> For example, we could talk about the close coupling of ILE languages and
> record level database access.  That proven combination leads to the most
> functional, best performing, most fault tolerant applications.

Not to discount your experience, or my personal opinion; but this is
subjective. There are no qualitive measures that I'm aware of to prove this.


> Contrast
> that with an ODBC or similar connection to the database.  We could talk
> about the productivity of ILE languages and supporting tools.  I
> have a lot
> of experience with Windows based software development, and I can
> unequivocally say that I and my peers have been less productive
> with Windows
> based languages and tools.

Overall, I'd tend to agree with this. Simplicity facilitates productivity,
and programming on the AS/400 is nothing if not simple. However, it's also
rather rigid. Joe Pluta has posted some examples of the work he's done with
servlets using various decorators that provide a level of flexibility and
productivity which would be difficult to achieve using the traditional ILE
toolset.


> We could talk about application deployment
> across multiple hardware tiers vs. hosting applications on a
> single server.
> The former leads to extra requirements, a need for multiple skill sets,
> increased complexity, less reliability, extra systems management,
> and poorer
> performance.

This is an architectural issue. There is nothing inherent in an non-as400
solution that forces you to design by crossing hardware tiers. Scalability
*may* require the use of multiple hosts, but that also provides certain
benefits, such as redundancy and load balancing. The performance issue has
already been decided; Cray's are being dumped in favour of multi-processor
server farms that deliver better price/performance numbers.

> These are but a few points.  So, while it's unfortunate that Rochester has
> priced itself out of some significant markets, I think there's a
> future for
> iSeries in the Web Application Server market.

Indeed, you're banking on it!

One thing that these discussions often overlook is the "good enough"
principle. There is ample historical evidence demonstrating that the best
solution is not usually the one to prevail in the marketplace. Many
companies have decided that *nix/Apache, or W2K/IIS is good enough. The very
fact that we're even having this discussion is a testament to that fact; how
many AS/400's is this message likely to cross before it reaches you?

As a final thought, I'd like to ask you to consider *why* IBM is moving away
from the low end of the server market with the new iSeries. Perhaps OS/400
isn't nearly as strategic to them as it is to you. Here's an interesting
article with Dr. Frank that I think you'll find quite illuminating:

http://search400.techtarget.com/originalContent/0,289142,sid3_gci537131,00.h
tml


John Taylor
Canada

+---
| This is the Midrange System Mailing List!
| To submit a new message, send your mail to MIDRANGE-L@midrange.com.
| To subscribe to this list send email to MIDRANGE-L-SUB@midrange.com.
| To unsubscribe from this list send email to MIDRANGE-L-UNSUB@midrange.com.
| Questions should be directed to the list owner/operator: david@midrange.com
+---