RE: InetAddressType and InetAddress

["C. M. Heard" <heard@pobox.com>]

On Tue, 19 Nov 2002, Wijnen, Bert (Bert) wrote:
> OK, if 20 is the upper limit (which I now see, cause that Address is also 
> used in another table as INDEX together with the long name), then I
> would still expect it to be a renage o (0..20) and not (4.. 20).
> Do we at least agree on that? If I looked at our earlier discussion
> (thanks for including that, I had lost it when my laptop crashed a few
> months ago), then I think that statement was made as well.

That's not unreasonable, although I don't think it is terribly important
(the point escaped my notice in the previous discussion).

> I would also want to see the restrictions (that only a subset of
> InetAddressTypes is required to be supported) listed in the MODULE
> COMPLIANCE. In a description clause if that is all we can do in the
> current SMI.

That's quite reasonable.  As things presently stand, a compliant
implementation is expected to support row creation for all address
types that fit within the size restriction, and a revision to the
InetAddressType and InetAddress TCs could cause a silent change in
meaning.  Spelling out what's expected in the various MODULE-COMPLIANCE
DESCRIPTION clauses would alleviate this.

Mike