Paul Thorpe wrote: > > On Tue, 25 Jun 2002, Egon Andersen, Talura wrote: >
> > > > To solve this problem and some other difficulties with parameterization, > > where the parameter is a DefinedObjectClass, I would suggest a change in > > the definition of Parameter, so that DefinedObjectClass should be > > preceded by "CLASS" ":". > > (Allow ParamGovernor to be "CLASS" in this case and change X.683 clause > > 8.3 a) to reflect this.) > > This would also remove the need for the NOTE in X.683 clause 8.5. > > This would make it easier for a human reader to understand and it would > > also make it easier to make compiler support. > > The only drawback I see is that protocols using DefinedObjectClass in > > parameterization, should be updated, but that is no big problem. > > Hi Egon, > > I have seen specifications that use DefinedObjectClass as a parameter. > The ASN.1 committee is very sensitive to changes that might cause > backwards compatibility problems. In this case, the note on X.683 clause > 8.5 indicates that you might not be able to determine what that dummy > parameter represents until the parameterized definition is used. > Hi Paul, You are right, there could be other ambiguities than just the objectclassreference / typereference problem. I also know that the ASN.1 committee is quite conservative in relation to changes. To some extend I agree to this, but I think that it is even more crutial to have a good standard that is simple to understand and hopefully also to implement. One of the major problems with ASN.1 is that it is difficult to understand for an average designer and it doesn't help that there are ambiguities / problems of the kind Steven pointed out here. So if we in any way can make the specifications more clear, I'm definitely in favour of that also if it cost a little on backwards compatibility. Best regards Egon Andersen -- * Talura ApS * Phone: +45 43 52 50 00 * * Baldersh�j 24 B * mailto:[EMAIL PROTECTED] * * DK-2635 Ish�j * http://www.talura.dk *
