Meng, on his site, laid down SPF for citizens of two worlds: an explanation
for system administrators, and the executive summary. I myself toil, and
have my existence, primarily in the former world. I cringe at having XML in
DNS; I loathe MS business ethics, or rather the lack thereof, as the next
guy. When I visit a site like Meng's, I invariably click on the
administrative link. And when I think of "cost", I think in terms of memory
foot-print, CPU usage, bandwidth, and rarely ever consider human resources.
Ever so often, however, I peek over the rim of the first floor and try and
imagine what the world looks like, wearing an executive hat. At those
moments I ask myself questions like, "Are there other than technical
considerations to go one way or the other?" And sometimes I answer yes. If
you speak to the Mark from the first floor, you see me agree,
wholeheartedly, with those opposed to XML for all the right reasons. But if
you catch me wearing that imaginary executive cap, you will also hear me
say, "There is more to introducing a standard than having clever specs; like
having big players onboard, for instance."
Sometimes I try and look beyond my Calimero-shell, and try not to think of
every compromise as "selling out", but wonder what genuine added value their
clout can bring. The system administrator in me belches out, "XML is DNS is
ridiculous!" But wearing that executive cap, I wonder, "Hmm, with both MS
and the sendmail people onboard, we suddenly have a real shot at
accomplishing two other important things: getting our own RR record, and
have the world adopt the new RFROM attribute."
See, here is where black and white start to blur. Because it is not just
about the question whether XML belongs in DNS, but also what good an
"alliance" with *really* big players can do for SPF. RFROM is brilliant,
really; and it is also the result of a "merger" between the pre- and post
DATA phase camps. But having a clever spec, as said before, remains but a
wet dream without real big players. Personally, I do not even think MS, in
and by itself, has much to contribute to SPF. My eye, really, is more keen
on watching who else joins the table when MS does. And if that player is
IETF, as has been alluded to, then I am inclined to extend MS a credit line
at this table, so to speak.
No one here, really, made their personal bias against MS their trump card.
Which speaks for our maturity, of course. But even if you really loathe MS,
then, still, eventually it comes down to this one question: do you hate MS
more than you love SPF? RFROM is good for SPF; XML in SPF sucks. If getting
MS onboard means having the IETF opening doors for us, heretofore firmly
closed, like getting our own RR type, and a bonafide RFROM extension for
SMTP, then I cordially invite people to consider the issue, for a brief
moment at least, in a broader sense than just the question whether or not
XML belongs in DNS. Especially if keeping our existing ASCII records is not
even off the table.
- Mark
System Administrator Asarian-host.org
---
"If you were supposed to understand it,
we wouldn't call it code." - FedEx