Re: Laymans explaination for single level store?

I agree with most of Mr. Richter's points as being limitations that seem
out of place today. However, I have to quibble with a couple of them...

The limitations described are a function of the "implementation and
exploitation" of SLS and not of the SLS architecture itself. I suspect the
reason that most of these changes have not been made is that they end up
being extremely pervasive and would require a huge level of investment in
time and money. These sorts of changes tend to end up low on the priority
list. Now one could argue that IBMs support of the POSIX file system
provides much of the function that Mr. Richter says is missing (e.g. long
file names, symbolic links, directories within directories, etc.) mentioned
in his post...and most of this is done on top of SLS -- proving it's
flexibility. However I understand that the limits are still applied to the
QSYS file system.

>The ability of the system to run code from 25 years ago is due in part
> to the fact that the system has not been changed much in the last 25
> years!

I strongly disagree with this particular statement. The system has changed
drastically in the last 25 years. At least three different hardware changes
for one. In security alone we have made drastic improvements/changes  and
yet those 25 year old programs still work. This is only partly because of
the implementation of SLS. The real reason is because of the visionaries
that did the original architecture design (one of whom was Dr. Frank, but
there were a whole slew of other folks involved in developing the original
architecture...none of which were me :-)) who had the incredible foresight
to understand how one would architect a system in order to allow programs
to run on different hardware without re-compiling.

Patrick Botz
The opinions expressed here are mine and not necessarily those of my
employer...whomever that might be...