Matthew Elvey <matthew(_at_)elvey(_dot_)com> wrote:
On 10/10/2004 9:46 AM, Mark sent forth electrons to convey:
... what is the point of (being allowed to) point to an reputation
service of your own choice? Yea, of your own making even? Then people
will just do what you did: set up an extra record for themselves,
and give themselves a big A+.
Let me try to answer this in the form of a suggested addition to the FAQ:
Q: Why does CSV allow the sender to specify the reputation services that
will vouch for it?
Good question!
A: It's an efficiency thing. The sender does this, but the receiver
will only look up the reputation at reputation services it respects.
This is more efficient than the sender checking for a reputation with
all the reputable reputation services it knows about.
If the sender specifies a disreputable reputation service or a
reputation service that itself lacks a reputation, it will be ignored.
It's more efficient given the (IMO valid) assumption that there will
typically be more reputable reputation services than reputable
reputations of a given FQDN. I anticipate a dozen or more of the former
and just a few of the latter will be typical.
Let me try rephrasing a few things:
" A: When the sender "suggests" accreditation service(s), this aids both
" the receiver and other accreditation services to know where to check
" for up-to-date information about the sending domain. Otherwise, there
" can be months-long delays before a particular service checks the
" performance of that domain and updates its report.
"
" Senders should be careful to "suggest" accreditation services which
" themselves have a good reputation. Otherwise, folks are likely to
" ignore whatever they report. Often, it will be best to suggest more
" than one accreditation service, since opinions of them may vary.
(I realize that didn't come out looking much like your suggestion.
But I think it's critically important to make the point about quickness
of updates; and it's also worth recommending listing more than one.
Feel free to comment on my wording -- we can't update the FAQ anyway
until Dave Crocker and I agree, since there are separate copies
currently published.)
--
John Leslie <john(_at_)jlc(_dot_)net>