Re: ietf-nntp Three proposals

[Brian Hernacki <[e-mail]>]

David Johnson (Exchange) wrote:
> I think we should be careful about which commands are made mandatory. In
> my opinion XPAT is only marginally useful, so why should it be required?
> I would like to see a list of the newsreaders that won't function
> without XPAT support. If this list is some number greater than one then
> maybe it should be required. On the other hand, I have seen many clients
> that won't function without XOVER, so it is reasonable to make this a
> required command.

We need to define a criteria for taking a de-facto extension and making
it a required part of the base protocol.

> I made the assertion at the BOF that "X" and "X-" naming conventions
> don't work and I still stand by that. We should document the commands as
> they exist. The only reliable way to prevent namespace problems is to
> establish a registry for extensions.

It does seem a bit silly to rename and bust clients...or even encourgae
folks to develop implementations that would bust existing clients (what
I think will happen if we rename). But if we allow X-commands to become
"real"..then we might as well scrap the idea of X-commands being "local"
or "experimental". It shoudl be one way or ther other.

I think the plan for the extensions mechanism was to have some
semi-formal way of registering names for new extensions, so it seems
like only existing extensions are in question.

--brian