Re: dev tool help... -- MIDRANGE-L

Jon 

I think that maybe we are both missing each others point!  I said, in a
very short note, that it was possible with our product to maintain live
applications, but, for brevity's sake, I did not say when I thought this
was desirable.  By live, I meant in use.  This includes testing, where it
very convenient.  Whether it is desirable on a production appplication will
depend on all sorts of factors but certainly it will not always be nor did
I mean to imply that it was.  Nevertheless you can add new attributes to
ERROS entity types which are in use in live production applications without
any impact on those apps unless you wish to process any of the new
attributes in an existing app.  Even then, you can make additions with high
security, so that ordinary users are unaware of them, test the changes, and
when you are happy, remove the additional security and so put the changes
into production without stopping the application.  But I am only too aware
that there are many situations where this is not desirable.

> 
> On a more serious note, I don't see that the implementation technology
(code
> generator or "interpreter" or whatever) has anything at all to do with
how
> intelligent the software is.  There is no connection between the two.
>

In my original post, I said that I believed that a truly Intelligent tool
would not require any coding or program generation.  Our ERROS Neural
Database achieves this a lot of the time, but not all.  It is not an
interpretive compiler. I repeat that I did not claim in my post that it was
intelligent.  
 
In my view,  software itself (i.e. hard coded programs) can never be
intelligent.  The von Neumann model splits rules and data by encapsulating
rules in programs (and therefore in concrete) and putting the information
to which those rules apply in a database.  This artificial split is the
cause of all the problems in the computer industry (complexity and frailty
of apps., difficulty of maintenance, etc.).  As I am sure you are aware,
John von Neumann designed his model for complex calculations on relatively
small data sets and it was appropriate for that.  He was not trying to
store and retrieve vast quantities of data nor was he trying to do MRP or
word processing or run a business.  I imagine that, if he were alive today,
he would be horrified that we (including me) are using his model for a
process for which it was not intended.

Only by treating the rules and data as one and the same and storing all in
the context provided by the connections, as I believe the brain does, can
we hope in due course to achieve some form of intelligence (whatever that
might be).  I believe that our Neural Database which stores most of the
rules and all the data together in a context provided by the connections
stored with the data is a step in that direction (but clearly there are
more required).  If and when it becomes appropriate to call any system
intelligent, it will, in my view, be the database not the software. 

Best wishes,

Rob Dixon
Erros plc
----------
> From: Jon Paris <paris@ca.ibm.com>
> To: MIDRANGE-L@midrange.com
> Subject: Re: dev tool help...
> Date: 14 July 1998 05:34
> 
> >> I did not say our Neural database was intelligent but ....
> 
> I think you missed my point (which was supposed to be humorous - rats
failed
> again!) .  I was merely pointing out that I don't think performing
maintenance
> on a live application is a terribly sensible thing to do.
> 
> On a more serious note, I don't see that the implementation technology
(code
> generator or "interpreter" or whatever) has anything at all to do with
how
> intelligent the software is.  There is no connection between the two.
> 
> 
> Jon Paris - AS/400 AD Market Support - paris@ca.ibm.com
> Phone: (416) 448-4019   -   Fax: (416) 448-4414
> +---
> | 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
> +---
+---
| 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
+---