SEDS reports for ISO 10303-11

Markus Maier maier at pdtec.de
Tue Feb 19 05:32:12 EST 2002


Hi,

the use of TYPEOF can be at least partially avoided,
by using 
EXISTS(x\typename) instead of 
'SCHEMA_NAME.TYPENAME' IN TYPEOF(x).
Of course this works only for entity instances.

Concerning the proposed "IS" extension, I agree that it's
a missing feature in EXPRESS to get the "most special
subtypes" of an instance. However, Jochen's proposal does not
take complex entity instances into account, we'd need
something like a IS type1&type2.  
However, "IS" is a dangerous keyword, 
it might break existing EXPRESS code if it is used
as attribute or variable identifier.

I also agree that TYPEOF is hard to read an that 
the long form generation affects the type names 
in the strings. But this also applies to ROLESOF and
USEDIN. So if we find something better for TYPEOF then we should
do the same for ROLESOF and USEDIN.
I'd recommend to have a look at the EXPRESS-X (CD or DIS proposal) 
path operators: 
- they are a good replacement fo USEDIN
 (and partially for QUERY/TYPEOF too) 
- readability and code size of WHERE rules would be 
  100% better compared to current EXPRESS.
- no new keyword is introduced, i.e. there is 100% upward compatibility.
- with some minor extensions these operators
  could be a full replacement for USEDIN/TYPEOF (and probably
  ROLESOF too).

Regards,
	Markus

_____________________________________________________
Markus Maier, PDTec GmbH      tel:   +49 721 61844 21
Haid-und-Neu-Str. 7           fax:   +49 721 61844 19
D-76131 Karlsruhe, Germany    email: maier at pdtec.de



More information about the wg11 mailing list