Re: [radext] #192 (bigger-packets): unclear meaning of "configured categories" of dynamically discovered servers

[Sam Hartman <[e-mail]>]

>>>>> "Stefan" == Stefan Winter <[e-mail]> writes:


    Stefan> Okay... We both agree that this parameter needs to be
    Stefan> configurable. None of us really knows how "categories"
    Stefan> should be defined exactly (nor do we know what will be
    Stefan> useful classification criteria in the field).

    Stefan> I think this is something actual implementations - and their
    Stefan> customer demand - can and will sort out themselves.

I agree with you so far.

    Stefan> The important bit is that it must be configurable in
    Stefan> principle IMHO; and that more than one global default is
    Stefan> something to consider for an implementation. So how about
    Stefan> this as a text proposal:

I also think it's important that this SHOULD  be configurable  along
with everything else that the implementation makes configurable for
dynamic discovery.
That is, I suspect that  the market  will help us decide what the
categories are, but I think we know that for each category an
implementation has, the implementation SHOULD make this parameter
configurable.

    Stefan> OLD:

    Stefan> If dynamic discovery mechanisms are supported, configuration
    Stefan> SHOULD be provided for the maximum size of clients and
    Stefan> servers in each dynamic discovery category.

    Stefan> NEW: If dynamic discovery mechanisms are supported, the
    Stefan> default maximum packet sizes for sending packets SHOULD be
    Stefan> configurable. Implementations MAY choose to allow
    Stefan> configuration of different defaults in different parts of
    Stefan> their dynamic discovery configuration.

Your text loses the should.  I don't find the ambiguity problematic in
the old text, so I prefer the old text.  However, provided that the
process gives my concerns a fair hearing, the new text is not something
I'd object to.  That is, I actively support the old text, but do not
raise an objection if the new text is the result of a reasonable
process.

I do not yet understand your concern with the old text.  It sounds at
some level like you're concerned the old text is ambiguous, but neither
of us seems to believe resolving that ambiguity is important.

--sam