RE: Platform Independence (was George is Farr from where we want tobe)

You make a lot of good points Pete, and from what I can see we are both on
the same page.  You, as a software reseller, want to write software in a
language that wont cause you to lose sales because you had to sell a piece
of expensive hardware to go with it.  I don't disagree with your approach
and it is great that the iSeries supports things like Java and PHP for your
purposes.

Here's what I struggle with: There are die hard RPG programmers writing home
grown applications for companies that have already chosen the iSeries as
their platform of choice.  They understand, for the most part, why they dish
out the many thousands of dollars a year.  They understand that they gain
ROI by using the same technologies repeatedly (RPG, DSPF, PF's, CL, DB2,etc)
year in and year out.  They have become very productive in meeting their
business needs.  Now IBM comes and says: "We have a roadmap for you to get
to GUI, and here it is.  You use these wizards and walla! You have GUI."
Well, in the middle of that they introduced a whole bunch of things that I
would consider far from being native.  They essentially took their tools
normally used for cross platform purposes and tried to make them fit with
intimate native practices.  True blue RPG shops don't want this (speaking
mostly for myself but believe others share my sentiment).  They don't want
to introduce another language.  They are not looking to migrate or run on
Windows.  These are they people that aren't getting as much development time
from IBM.

Thanks for your comments Pete,
Aaron Bartell

-----Original Message-----
From: midrange-l-bounces@xxxxxxxxxxxx
[mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of Pete Helgren
Sent: Wednesday, March 28, 2007 1:20 PM
To: Midrange Systems Technical Discussion
Subject: Re: Platform Independence (was George is Farr from where we want
tobe)

Aaron,

I am having mixed feeling about this whole topic because there are so many
nuances to investigate that finding a simple solution seems impossible to
me.  Backing up a second and looking at the System i in relation to all
other platforms, I think we all agree that it is software ("solutions") that
sells the system.  That is, it is very hard to just lead with the platform
and then sell a software application to go with it.  However, when someone
is looking for a software package they have folks in the IT department
recommending that they stick with something that runs on their current
platform(s).  I work primarily with school districts and outside of the
System i's I deal with it is probably about 90% Wintel and perhaps 10%
Linux/Mac (mostly Mac).  So if I am developing an application I have a
choice of whether to develop an application that will run ONLY on the System
i (read: RPG) or develop it to run on any of the platforms at the school
district.  In this environment I could chose any approach and still be OK.
In fact, for my System i shops I probably have 40% of my apps written with
or using System i centric languages or platform specific middleware, 40%
where it could be running on any platform and 20% that is Microsoft centric.
I am OK with that and, in fact, I love that environment because I can use a
broad range of skills and technologies to solve problems.

But, from a business viability point of view, I am in a bit of a quandary.
There are approximately 25,000 school districts out there and I am going to
WAG that only 500 or so have System i's.  So when I write an application, do
I write an application that has a potential market of 25k or of 500?
Instinct tells me to go for the broader market.  Don't get me wrong, there
is nothing wrong with "specializing" in System i.  
In fact, I became an System i reseller so I could package and sell a System
i with my solution. However, selling a System i into a Wintel shop is a
challenge.  Even if you can sell them on System i stability and reliability,
you have a "value" issue when they compare to Wintel.  
In short, the battle is all uphill. Further, IBM has made the whole process
of becoming a "small" System i reseller so difficult that in the end the
energy expended just doesn't seem worth it.  It is much easier IMHO to write
in Java or PHP and lead with a System i for hardware but if you lose that
battle you can still get the sale with something that is platform
independent.

Nobody has to sell me on the value of a System i, in some ways I am blinded
by my love of the platform, but that perspective comes from being on the
platform for 20 years.  Folks who are new to the System i are a much harder
sale.  Not impossible, just harder to sell.  I guess if I was a crack
salesman myself or could find a sales person that could easily sell a System
i, I wouldn't sweat the platform independence.  
However, my experience has been it isn't all that easy to get a System i
into an existing Wintel, Linux or Mac shop for all of the reasons that have
been floated on this list.

So, my vote for a platform independent approach isn't based on technology,
just business. Just generating a GUI app on a System i isn't going to
guarantee success.  The challenge will be to displace Wintel servers with
System i's and it will take more than just a GUI to do that.  If anything
*could* easily displace a Wintel server it would be a killer app that is
ONLY available on a System i. Finding that app will be about as easy as
finding the Holy Grail.

Pete Helgren


albartell wrote:

... but I have heard of way to many System i systems or applications 
being
    

replaced because they are not modern/gui/whatever.  Given that, 
doesn't something have to change?  And how should it be accomplished?

That is EXACTLY what I am talking about in my rant to George.  IBM is 
trying to introduce us to COMPLETELY new languages (Java/EGL) and servers

(WAS).

Changing to those vs. modifying RPG to meet the same need is where I 
think they have gone wrong.  The change needs to happen to the RPG 
language/platform and not through new introduction.  Sure they are 
making consistent changes to RPG (small changes here and there), but 
they have ignored the big elephant in the room of us needing RPG 
native ways to do web programming and RPG native thick client interfaces.

Thanks Mike,
Aaron Bartell


-----Original Message-----
From: midrange-l-bounces@xxxxxxxxxxxx
[mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of Crump, Mike
Sent: Wednesday, March 28, 2007 11:13 AM
To: Midrange Systems Technical Discussion
Subject: RE: Platform Independence (was George is Farr from where we 
want
tobe)

Aaron,

I would definitely say that with any of the modern language 
environments that the risk of 'performance' problems is much higher 
than it is with our working languages.  "your mileage may vary"  
"buyer beware" "it depends" are definite terms to be used when going down

this road.

Don't forget that within the RPG world we have decades of experience 
and the learning was typically done at a nice pace (read linear).  
Today the people using JAVA (even if they are knowledgeable) probably 
won't get to the same level of comfort for quite a while.  And the 
expectations are that they pick this up in a exponential fashion.

I know this goes to your point but the best JAVA programmers are the 
guys who have been doing C or C++ programming for decades.  Same with 
.NET weenies.  The complexity transition to any new technology is not 
to be trivialized. The ripple effect is a nice way to put it.  Seems 
to be it some where between a ripple and a rogue wave.  The 
administration of this is even more complex.  And problem determination?

Generally, I agree with most points but I'm not necessarily willing to 
damn the new technologies.  There is a need and a fit out there.  
Change for the sake of change is never good but I have heard of way to 
many System i systems or applications being replaced because they are 
not modern/gui/whatever.  Given that, doesn't something have to 
change?  And how should it be accomplished?

Michael Crump
 
Manager, Computing Services
Saint-Gobain Containers, Inc.
1509 S. Macedonia Ave.
Muncie, IN  47302
765.741.7696
765.741.7012 f

Don't ever be the first, don't ever be the last and don't ever 
volunteer to do anything.
This email and its attachments may be confidential and are intended 
solely for the use of the individual to whom it is addressed. Any 
views or opinions expressed are solely those of the author and do not 
necessarily represent those of Saint-Gobain.  If it did, it would be 
folded, mutilated, watered down, politically corrected, and would show 
up a week later if at all.  If you are not the intended recipient of 
this email and its attachments, you must take no action based upon 
them, nor must you copy or show them to anyone.
Please contact the sender if you believe you have received this email 
in error.
 
-----Original Message-----
From: midrange-l-bounces@xxxxxxxxxxxx
[mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of albartell
Sent: Wednesday, March 28, 2007 9:56 AM
To: 'Midrange Systems Technical Discussion'
Subject: RE: Platform Independence (was George is Farr from where we 
want
tobe)

  

in essence implemented on a system that is already running a decent
    

workload, does not have enough memory, is utilizing an older 
processor, is not written very well, etc.

Interestingly enough, these are all things that we "can get away with"
when
coding RPG.  Yes an RPG program can still kill a system, but I think 
that happens much less often than in the Java environment.  I didn't 
put the word "scholarly" in there simply because I was lacking word 
choice.  You have to know A LOT about the ins and outs of Java components

to 'get it right'

when
creating a large application with the potential for a lot of 
concurrent users.  In RPG you are there much quicker.
 
Good to hear about your other failures to successes with Java. I don't 
hear much of the successes believe it or not.  But like you commented 
in #1, a lot of these things could/should simply be implemented in RPG 
as that is the language most shops support.  Introducing a new 
language into the mix in any shop has MANY ripple effects (think QA, 
Support, Training, Debugging, Split knowledge on multiple platforms, 
etc)

Thanks for your comments Mike,
Aaron Bartell

-----Original Message-----
From: midrange-l-bounces@xxxxxxxxxxxx
[mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of Crump, Mike
Sent: Wednesday, March 28, 2007 8:33 AM
To: Midrange Systems Technical Discussion
Subject: RE: Platform Independence (was George is Farr from where we 
want
tobe)

Platform independence?  Not very.  We use JAVA more so to allow for 
using the ubiquitous browser and for independence from MS.

Although, I would also have to respectfully disagree with some of 
Aaron's points.  While I would not argue that plenty of people have 
horror stories regarding native Java performance I think that you will 
find a lot of them are of there own making - in essence implemented on 
a system that is already running a decent workload, does not have 
enough memory, is utilizing an older processor, is not written very well,

etc.

1.)  We have a native JAVA program that processes our finite 
scheduling system.  This is program that does a lot of IO and I argued 
up front that it should have never been written in Java because it was a

classic fit for RPG.

I lost but I will have to say that application runs blazingly fast and 
processes many functions and millions of records.  I doubt that we 
would see any improvement if it was converted to RPG.  In addition, by 
most testing it is as fast as running it on an independent WINTEL 
platform.  I have compared it run wise on our i5 to an IXS system (to 
eliminate remote database
issues)
and there is no comparison.  Right now, I am not in the position to 
test on a faster Bladecenter type of unit but perhaps in the future.  
The database is remote to the native application as well.

2.)  We originally implemented WAS (v3) on a 720 years ago.  80% of 
the application runs perfectly fine but certain components were a dog.
Processor was not being taxed, we were not short of memory, we just 
did not have the processor crank and L2 cache that the application needed.
Shortly thereafter we upgraded to an 825 and 100% of the application 
ran fine.  A classic example of implementing Java in a not so perfect

platform.

I would argue that as we have climbed the server ladder 7xx, 8xx, 5xx 
the potential for this problem has decreased.

3.)  We run WAS and WPS on our 570's and generally speaking these guys 
perform very well.  Our major complaint on WPS comes from portlets 
that we rely on our parent company for and their responsiveness is not

very good.

4.)  The native JVM can be a resource hog.  That is why the J9 was

invented.

Certain machines and customers should not be running the native JVM.
They
do not have the horse for it.  Is it a problem with the native JVM?
Perhaps.  I would just argue that it is as much environmental as it is 
platform.

I cannot argue with the complexity.  However, I think that this is not 
a JAVA problem only.  Every existing 'modern' application is much more 
complex than the traditional RPG model.  Whether it is Java, .NET, Domino,

etc.

I
think that they are much more complicated either in architecture or in 
resulting code length.

Michael Crump
 
Manager, Computing Services
Saint-Gobain Containers, Inc.
1509 S. Macedonia Ave.
Muncie, IN  47302
765.741.7696
765.741.7012 f

Cluelessness
There are no stupid questions, but there are a LOT of inquisitive idiots. 
This email and its attachments may be confidential and are intended 
solely for the use of the individual to whom it is addressed. Any 
views or opinions expressed are solely those of the author and do not 
necessarily represent those of Saint-Gobain.  If it did, it would be 
folded, mutilated, watered down, politically corrected, and would show up

a week later if at all.

If
you are not the intended recipient of this email and its attachments, 
you must take no action based upon them, nor must you copy or show 
them to anyone.
Please contact the sender if you believe you have received this email 
in error.
 
-----Original Message-----
From: midrange-l-bounces@xxxxxxxxxxxx
[mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of Michael Ryan
Sent: Wednesday, March 28, 2007 8:10 AM
To: Midrange Systems Technical Discussion
Subject: Platform Independence (was George is Farr from where we want 
to
be)

How important is platform independence to the community? It has little 
importance to me. I believe RPG is the best language on iSeries, and I 
don't need to write the same code to run on multiple platforms. I *do* 
write code on multiple platforms (VB and C on Wintel), but I don't 
expect or want to run the *same* code on other platforms.

On 3/28/07, albartell <albartell@xxxxxxxxx> wrote:

  

While in that session I was very close to asking the group whether
    

they gave
  

a rip about platform independence.
    

--
This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing 
list To post a message email: MIDRANGE-L@xxxxxxxxxxxx To subscribe, 
unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx Before posting, please take 
a moment to review the archives at http://archive.midrange.com/midrange-l.


--
This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing 
list To post a message email: MIDRANGE-L@xxxxxxxxxxxx To subscribe, 
unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx Before posting, please take 
a moment to review the archives at http://archive.midrange.com/midrange-l.

--
This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing 
list To post a message email: MIDRANGE-L@xxxxxxxxxxxx To subscribe, 
unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx Before posting, please take 
a moment to review the archives at http://archive.midrange.com/midrange-l.


  

--
This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing list
To post a message email: MIDRANGE-L@xxxxxxxxxxxx To subscribe, unsubscribe,
or change list options,
visit: http://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx Before posting, please take a
moment to review the archives at http://archive.midrange.com/midrange-l.