For various reasons, the AS-NULL as-set scratched a rather pressing itch earlier today. In the interests of neutrality, would the RIPE NCC be interested in maintaining ownership of this object as an empty set? Nick
On Mon, Dec 15, 2014 at 04:05:14PM +0000, Nick Hilliard wrote:
For various reasons, the AS-NULL as-set scratched a rather pressing itch earlier today.
In the interests of neutrality, would the RIPE NCC be interested in maintaining ownership of this object as an empty set?
Can you elaborate on the use of AS-NULL? Isn't it common place to use something like "NOT ANY"? For what it's worth: I have no objections to RIPE maintaining an empty AS-SET named AS-NULL. Kind regards, Job
On 15/12/2014 16:13, Job Snijders wrote:
Can you elaborate on the use of AS-NULL? Isn't it common place to use something like "NOT ANY"?
"NOT ANY" is not the same as AS-NULL - they're not interchangeable. AS-NULL is a convenience placeholder token for where might be a need to have an empty as-set. My particular itch today involved needing to use an AS-SET of some form, but where the contents didn't particularly matter. It was more convenient to use an empty set rather than a set with some members in it because that gave an element of semantic consistency. There are other similar sets in the various irrdbs, e.g. AS-EMPTY occurs in RADB. This suggests that there is at least one other person in the world who needs this sort of construct. Nick
On 15/12/14 16:26, Nick Hilliard wrote:
On 15/12/2014 16:13, Job Snijders wrote:
Can you elaborate on the use of AS-NULL? Isn't it common place to use something like "NOT ANY"? "NOT ANY" is not the same as AS-NULL - they're not interchangeable.
AS-NULL is a convenience placeholder token for where might be a need to have an empty as-set. My particular itch today involved needing to use an AS-SET of some form, but where the contents didn't particularly matter. It was more convenient to use an empty set rather than a set with some members in it because that gave an element of semantic consistency.
There are other similar sets in the various irrdbs, e.g. AS-EMPTY occurs in RADB. This suggests that there is at least one other person in the world who needs this sort of construct.
I don't see any particular problem with the NCC creating and owning such an empty set Nigel
Nigel Titley wrote:
On 15/12/14 16:26, Nick Hilliard wrote:
On 15/12/2014 16:13, Job Snijders wrote:
Can you elaborate on the use of AS-NULL? Isn't it common place to use something like "NOT ANY"?
"NOT ANY" is not the same as AS-NULL - they're not interchangeable.
AS-NULL is a convenience placeholder token for where might be a need to have an empty as-set. My particular itch today involved needing to use an AS-SET of some form, but where the contents didn't particularly matter. It was more convenient to use an empty set rather than a set with some members in it because that gave an element of semantic consistency.
There are other similar sets in the various irrdbs, e.g. AS-EMPTY occurs in RADB. This suggests that there is at least one other person in the world who needs this sort of construct.
I don't see any particular problem with the NCC creating and owning such an empty set
Same here, in principle, but... After having a look at some RFCs I'd suggest to rather go for an AS-EMPTY, as the use of AS0 (AS zero or null) seems to have some semantics attached. I may easily be wrong. :-) Wilfried
Nigel
On 16/12/2014 08:51, Wilfried Woeber wrote:
After having a look at some RFCs I'd suggest to rather go for an AS-EMPTY, as the use of AS0 (AS zero or null) seems to have some semantics attached.
AS0 is autonomous system number 0 which is not what we're talking about here. "Null" has a very specific meaning: it refers to a set with no members. Given that RPSL is designed around set theory, AS-NULL is a more appropriate name than AS-EMPTY. Nick
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
AS0 is autonomous system number 0 which is not what we're talking about here.
"Null" has a very specific meaning: it refers to a set with no members. Given that RPSL is designed around set theory, AS-NULL is a more appropriate name than AS-EMPTY.
FWIW, AS-NULL makes more sense to me. Cheers, Rob -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.9 (Darwin) Comment: GPGTools - http://gpgtools.org iEYEARECAAYFAlSQXSMACgkQF83fs8P8zamP+wCfVptYE2meqtKuo4ERy9HK57MT ctEAoIpdE+rbb5qWfr4+vJl35ydeEJZD =HUmX -----END PGP SIGNATURE-----
Hi Group, This has been taken care off: Vurt:~ job$ whois AS-NULL -r | grep -v % as-set: AS-NULL descr: The NULL AS-SET. This is an empty AS-SET. admin-c: OPS4-RIPE tech-c: OPS4-RIPE mnt-by: RIPE-NCC-MNT source: RIPE # Filtered Case closed! Kind regards, Job
Dear working group, We were following the discussion with interest and waiting for consensus on the naming of this object AS-NULL vs AS-EMPTY. The object has now been created. Kind regards, Tim Bruijnzeels Assistant Manager Database Group RIPE NCC On 16 Dec 2014, at 18:02, Job Snijders <job@instituut.net> wrote:
Hi Group,
This has been taken care off:
Vurt:~ job$ whois AS-NULL -r | grep -v % as-set: AS-NULL descr: The NULL AS-SET. This is an empty AS-SET. admin-c: OPS4-RIPE tech-c: OPS4-RIPE mnt-by: RIPE-NCC-MNT source: RIPE # Filtered
Case closed!
Kind regards,
Job
participants (7)
-
Job Snijders -
Job Snijders -
Nick Hilliard -
Nigel Titley -
Rob Evans -
Tim Bruijnzeels -
Wilfried Woeber