Modelling the modelling process (Re: Methods)

[Harald Tveit Alvestrand <Harald@Alvestrand.no>]

At 01:55 15.05.2000 -0700, Andrea Westerinen wrote:
>That is a very valid question - "What does the DMTF's MappingStrings
>qualifier mean?" This qualifier indicates where data could be mapped to/from
>the CIM property.  It does not specify the algorithm for the mapping - only
>that logically similar data is defined.

hmmm.
it seems to me that we've got a model of the modelling process in our minds,
which looks something like this:

   +---------------------+
   | System model        |-------+
   | (protocol independent)      |
   +---------------------+       |
                                 +--->
   +---------------------+                              Protocol description
   | Mapping mechanisms  |----------->  MAGIC  ------>  (eg MIB, repository or
   | (model independent) |                     --+      "other" for managing
   +---------------------+       +--->           |      system
                                 |               |
   +---------------------+       |               +---> Description of what
   | Fudges              |-------+                     can't be represented
   +---------------------+                             in this protocol

and that the requirements for the modelling include:

- keeping the amount of fudge small
- keeping the "magic" standard
- making the "can't be represented" heap small

noting that you lose fractional megabits(?) in mapping to CIM is a property 
of the mapping mechanism. What to do about it is probably a "fudge", since 
it depends on both the system and the protocol.

but probably we have to live with a reasonable amount of fudge for a good
long while still.

My guess at the metasystem we seem to be working on....

                  Harald

--
Harald Tveit Alvestrand, EDB Maxware, Norway
Harald.Alvestrand@edb.maxware.no