In <CC3A855E-069E-11D9-B753-000393A56BB6(_at_)glyphic(_dot_)com> Mark Lentczner
<markl(_at_)glyphic(_dot_)com> writes:
I am against them being included as part of the standard.
[....]
I am for implementations supporting them.
*UGH*
What next? Implementations should (*wink* *wink*) support ipv4: as an
alias for ip4:? And prt: is ok for ptr:? And putting an IP address
on a: is something implementations should support because there are
alread a lot of records out there that have it?
Of course "v=spf1" records can be almost certainly interpreted at
"spf2.0/mailfrom". Surely, any implementation can do this trivially.
They should. But when things go wrong with "v=spf1" records, for any
reason, better for us to be able to all say: "Use the standard"
This is circular reasoning. Things will go wrong with both the SPFv1
records and SPFv2 records. By excluding them from the standard you
are by definition making them non-standard. By making them
non-standard, people will not want to support them, or if they support
them, they won't have anything to guide them.
How fricken hard it is it put a sentence in the standard that says:
"Records that begin with v=spf1 MUST be interpreted as spf2.0/mailfrom,helo"
-wayne