Re: AS/400 STILL wearing the undeserved "closed system" moniker (was PC connection via twinax ?)

Wow...   you know, there is not a shadow of a doubt in my mind that the
AS/400 is a "closed system".   The REAL question is, "is that a bad
thing?"

Lets say you wanted to write your own operating system to replace 
OS/400.  What kind of hoops would you have to jump through to get THAT
kind of info from IBM?   Or couldn't get get them at all?

Or how about writing a device driver for some piece of hardware you'd
like to sell, independent of IBM...   at *BEST* you'd have to be a
registered business partner, at *WORST* you wouldn't be allowed to at all.

When I'm writing a program on my FreeBSD system (which *is* open) I can
ask myself "hmmm... how much work does the system do when I call the
read() function?  Is it better to optimize the code so that I don't
have to call it as frequently?"  And then I can look at the source code
for the read() function and see EXACTLY what it has to go through, and
how efficient it is...   I can even recompile the whole operating 
system with debugging symbols included, and use a debugger to go step
by step through the OS code, just as I can do with STRDBG on my RPG
code...

Are these good things or bad things?   This closed-ness to the iSeries
architecture has ensured the stability of the system.  It would be very
easy to argue that a "closed system" is a "stable system"

But, there should be no doubt in anyone's mind that the AS/400 is a 
closed system.   It's VERY MUCH a closed system.



On Fri, 17 Aug 2001, Buck Calabro wrote:

> Comments inline.
> 
> >>They have given more "access" to this machine than any 
> >>other box they have ever sold. I don't know what the current 
> >>API count is, but what would be some examples of things 
> >>developers "need to know" that isn't currently available ???
> 
> >database access:
> >Can you write a 3rd party query tool that has the same performance as
> RunQry?
> I have a command line interactive SQL.  Does that count?
> 
> >Could a 3rd party odbc driver be written for the 400?
> Already exist.
> 
> >Is oracle able to port its database product to the 400?
> There's no market for Oracle, but SAP managed a port.
> 
> >Programming:
> >If you write a 3rd party programming language, how do get 
> >the objects referenced in a pgm written in that language 
> >to be included in dsppgmref?
> Read Leif's MI book.
> 
> >Much of ile is shrouded in secrecy.
> A bit vague.  Have you seen the Redbook?  The ILE manuals?  Taken any IBM
> education?
> 
> >Twinax devices:
> >Is it known how a pgm can write direct to a twinax terminal or printer
> device?
> Yep.  I have a 5250 datastream reference book in my hand, published in March
> 1987.
> 
> >Back in the day when converting from the s36 to 400, there was 
> >an ibm product that allowed data transfer thru the twinax port.  
> >Is it known how to do that?
> Yep.  They simply wrapped the data in a simple 5250 display and extracted it
> at the far end.  Albert York has a utility that would serve admirably as the
> basis for such a project if ftp, SNADS or Netserver somehow don't meet your
> needs.
> 
> I believe that the system still carries the "closed" stigma is because there
> are too few 3rd party software houses.  Why are there too few?  The midrange
> market as a whole is very budget conscious, and rarely buys software, or
> even hardware.  Want proof?  Look at the places still on V3R2 (CISC).  Look
> at the companies that still want *M36.  This isn't a rip on small companies
> - it's a simple acknowledgement that small companies carefully watch
> expenditures.  It's a reality of the midrange market space.
> 
> At any rate, about the only big thing that's missing from the platform is
> universal support for operational descriptors.  Give us those, and we'll be
> just about home free.
> 
> My opinion only,
>   Buck