Thomas Corte wrote: > I just realized that the agreement seems to be that it is OK to specify a > larger fee than actually charged by the server.
Yes. And i think it's good. > I don't think this is a good idea, I'd prefer requiring a perfect match > of all fees. Sure, allowing the specification of larger fees still guards > the registrar from losing money, but it will also potentially lead to the > registrar unintendedly overcharging a customer if e.g. fees are > statically configured in a registrar's system, and a price change > notification is missed. We can never prevent registrars from "overcharging" a customer, and i do consider it out of scope for the IETF. What i don't consider out of scope of the IETF, however, is the robustness principle of "be conservative what you send, be liberal in what you accept". Especially in situations where there's a rush for names, a failed transaction just because someone "overbid" the registry could create problems. Further, requiring a "perfect match" would prevent models like dutch auctions, where prices slowly decrease over time. A check could never reflect the actual, current price, so "overbidding" is required in such situations. More hypothetically, but, possible. > It also raises the question what to do when multiple fees are involved. > If the server e.g. charges 50 for creating an initial application > (immediate) and 50 later upon a domain's allocation (delayed), should the > server accept it if the registrar specifies 60 (immediate) and 40 > (delayed), i.e. if the total sum of the fees in the create request is > sufficient, but the individual amounts don't match? No. Each fee involved would need to be equal or over the fee required by the server. > At the very least, I'd leave it up to a server's policy to accept fees > which are higher than the actually assessed fees. I do suggest that the text says something like the following only (only make clear that fees must be sufficient, but don't specify what happens if they are above the required value): "A server MUST reject a transform command if client supplied fee values for the fees involved in the transaction are lower than the server requires" (in better english ;) Alex _______________________________________________ regext mailing list regext@ietf.org https://www.ietf.org/mailman/listinfo/regext
