Havard Eidnes <Havard.Eidnes@runit.sintef.no> writes: * > Actually, what the document says (maybe not too clear) is that when * > you send in an object that has guarded attributes, and you have * > defined them different than what the files say, the object WILL be * > updated, BUT all guarded attributes will be reset to their "guarded" * > value. * * Ok, sounds reasonable. Let me suggest a change (addition to the last part * of the paragraph) to the text along the lines of: * * If an update is sent to the database software using another mechan- * ism (i.e. mail to auto-dbm@ripe.net) that contains a guarded attri- * bute, this will not be allowed to change the guarded attribute. If * the value of the attribute is the same as what is currently * registered in the database then no warnings will be given. However, * if the update contains a value for a guarded attribute that is dif- * ferent to that registered in the database, a warning will be sent to * the originator and the guarded value will remain unchanged. Any * changes of other (unguarded) fields in the update will be checked * for syntactic correctness and if they pass will go through to the * database irrespective of any conflicts for the guarded fields. * * * - Havard Havard, sounds great - I'll the changed text in before the meeting. --Tony.