On Mon, May 31, 2004 at 01:47:01AM -0700, Greg Connor wrote:
1. The poll is pretty much useless for our purposes. I would really like
to see more choices, so we can get to the heart of the issues and possibly
solve them.
* Opposed to any XML, due to bloat, patent/IP issues, hate MS, other (list
any that apply)
* Let the receivers implement both, so that the publishers may publish
either one.
* Use XML only
Well, ok, that's more fine grained than just 'yes' or 'no', but it still
misses options that would be relevant to the debate. I think on the one
end we have a full blown discussion, on the other end a 'yes' 'no' vote.
Both extremes have their merrits, the former is useful for gaining
technical insights and come up with new ideas. The second gives you a
strong indication of how the market thinks about it. I think most admins
will look at spf, quickly decide whether this is too complicated or not
to implement, and then label it 'yes' or 'no'. If spf includes xml, the
chance of it being labelled 'no' rises significantly, as the poll
showed. So if we want to predict adoption rates, the poll is certainly
not a waste of time. But I think we share this opinion already.
Here is the boiled-down version.
* If extensibility is a requirement, XML wins. It at least has syntax
extensibility. So far it has hand-waving in place of feature
extensibility, but that can be sorted out easily.
It's extensibility is at the same time it's weak point (and I'm
repeating earlier arguments made by others here). The fact that it is
infinetelly extensible, with xml comments for one, will result in
infinite extensibility. You will have to put some brake on the extending
forces, before every single company has it's own extensions, and most
importantly, before one company comes up with patended extensions that
only that company may use. Can this be prevented? If yes, how?
* Yes, extensibility IS a requirement, and version numbers don't cut it.
If we want to keep it simple, yes version numbers do cut it imho. If we
want extensions, let's discuss those extensions, come up with the proper
way to implement these extensions and then if a certain treshold of new
extensions has been reached, increase the version number.
* This is the tricky part. NO I AM NOT -- REPEAT NOT -- SAYING THAT XML
IS BETTER. I am saying that we have an opportunity RIGHT NOW to FIX the
SPF spec so that it has as much or even more extensibility than XML and
still remain lean.
Agreed, and you mentioned "Should SPF be Frozen or Extensible? (XML
insights)". I really like his idea of keeping with the linear parsing,
prepending thinks like 'unsupported=+' before the extensions. Maybe I
missed something, but is there a good reason we need the nesting xml orA
other solutions in said post offer?
* If market forces are to decide, and I think they will, I think SPF plain
syntax will win. I sure want it to!
Judging the poll so far, it will win. I really think most admins will
not spend too much time looking into each and every protocol like spf
and get into the technical benefits and contra-indications (in an ideal
world, they will, but this world is far far from ideal). They will
simply stumble across something, see if it fits their particular needs
and if it can be deployed without too much hassle (such as signing
agreements, adding xml parser libraries to their mail servers, etc..).
I really want to stress this: adding bulk to email servers is not
something i really like. Ideally, i'd like just the one package i can
place in any (virtual) server anytime i want, without digging into
library dependencies etc.. Some of the mail servers I administer run on
low-capacity hardware, and my customers really like to keep it that way.
Big companies can just buy themselves another 10000 euro (=12000 dollar
i believe) server and add whatever bulky stuff they want, smaller
companies have to use what they have got.
Regardless of how you feel about XML, IF you think extensibility is
important, will you PLEASE take the time to read and respond on that
thread? Michael Brumm wrote an excellent summary about two hours ago under
the thread "Should SPF be Frozen or Extensible? (XML insights)"
The reason i did not participate in this discussion earlier is that
really, I don't know whether and if so to what extend we really need
extensibility. I can imagine we need some, but i think infinite nesting
is not needed. I'd like to see more arguments on this though.
Koen Martens
--
http://www.sonologic.nl/