omniORBpy / IDL mapping question...

[Martin von Loewis]

cbbrowne at news.hex.net (Christopher Browne) writes:

> As an arguably odd-ball choice, you could use a DBM file, so that the
> values actually reside in a database outside the Python world.  It
> doesn't have to be a plain old variable...

Indeed, this is one rationale for mapping attributes that way (Paul
Prescod calls it "computed attributes", and complains every now and
then that Python does not really support them).

The other rationale for mapping attributes to operation pairs is that
the CORBA Core specification says that attributes really *are*
operation pairs, so the IDL notation is just a short-hand notation.

Regards,
Martin