Re: power line (AS/400) article in computerworld

Steve:

You got much of what you have been asking for with the recent new hardware announcements and pricing. (POWER = i + p)

But your posts on this list over the past few months indicate a profound lack of understanding of just what Single Level Storage and MI architecture is all about.

You keep harping on the 16MB segment size as if it is a huge problem. It is not. If you truly need to address more than 16MB of contiguous storage, you can use teraspace (since V4R4).

MI pointers DO remain valid between IPLs. Only temporary pointers (such as space pointers into heap space, teraspace pointers, etc.) are not valid across an IPL.

ILE has been improved over the years, but many of the improvements were "under the covers" and might not have been obvious to you. In V6R1, i5/OS can now "load on demand" any *SRVPGMs used by an application, rather than loading them all at once. This can reduce the overhead to start up large ILE applications, because *SRVPGMs are not resolved and activated until they are actually used by the application.

Once again, you are harping on 10 character object names. It has been suggested (by others more knowledgeable than me) that eventually, the operating system will be enhanced to support longer object names in the QSYS.LIB file system, probably at the same time that this file system is enhanced to allow Unicode names or other CCSIDs. This is not really that big of a deal to most people, but you continue to rant and rave about it.

Give us a break!

Sincerely,

Mark S. Waterbury

> Steve Richter wrote:

the spirit of the article and comments reads as an embrace of the operating
system of the 1990s. The hardware people at IBM are doing world class work.
The software team is what is holding us back now. SLS is aged with its 16meg
segment limit and the inablility of pointers to stay valid between IPLs. ILE
has not been improved in 10 yrs. 10 char limit on object names is bad, Fix
it, ok.


On Thu, Apr 10, 2008 at 9:29 PM, Mike Cunningham <mcunning@xxxxxxx> wrote:

http://blogs.computerworld.com/frankly_speaking_not_dead_yet
--