Re: NZ whois server
On the 'liberal in what you accept' precept, I'd say that the server should either respond with the info, followed by somthing like % NOTE: Your whois client does not conform to the standard. % It failed to terminate the query string properly or even just provide the carp message and decline to provide the information, which is less helpful but more of an incentive to fix the broken client. In other words, if we want to throw rocks at the server, the most we can really complain about is that it failed silently when could have been a little more voluble. I'm no fan of 'liberal in what you accept'='I am a doormat, abuse me as you will, I will add code to work around your mistakes' for three reasons. 1. If the client writer doesn't know there is a problem, they can't fix it. 2. The server suffers code bloat. 3. When second-guessing what the client actually meant, it may well get it wrong (doesn't apply this time). However, if they are going to open the code to make the change you suggest though, the formating of the output could use some work. IMHO it's the ugliest whois server output I've ever seen, not to mention the most verbose. But as Don says, there are other things that need fixing more. -- Michael Newbery Technology Manager Saturn Communications Tel: +64-4-939 5102 Mobile:021-642 957 Fax:+64-4-939 5100
Don Stokes
- 30/5/00 6:38 PM >>> "Robert Gray" wrote: It's not BUFFOONZ but RIPE who's out of step with the rest of the world. I should have known ;)
Well, in this case, the RFCs are on Domainz's side, not RIPE's. Not to
mention every other whois client on the planet which sends CRLF at the
ends of their queries, as required by reputable network protocols the
world over.
Andy Linton
"Be liberal in what you accept, and conservative in what you send"
Indeed, hence the suggestion that it be made to terminate on any control character. But I suspect that there are higher priorities at the moment than making the whois server talk to non-compliant clients. (It's not my code.) -- don --------- To unsubscribe from nznog, send email to majordomo(a)list.waikato.ac.nz where the body of your message reads: unsubscribe nznog --------- To unsubscribe from nznog, send email to majordomo(a)list.waikato.ac.nz where the body of your message reads: unsubscribe nznog
"Michael Newbery"
However, if they are going to open the code to make the change you suggest though, the formating of the output could use some work. IMHO it's the ugliest whois server output I've ever seen, not to mention the most verbose.
The format is intended to be usable as an email template (with the addition of the necessary authentication fields of course), a la the RIPE/APNIC et al whois. Now, one could say that the email template is ugly, and I'd agree, but given the flamage the template format has produced in the past, I'd really rather someone else proposed changing it... 8-) -- don --------- To unsubscribe from nznog, send email to majordomo(a)list.waikato.ac.nz where the body of your message reads: unsubscribe nznog
On Wed, 31 May 2000, Don Stokes wrote:
Now, one could say that the email template is ugly, and I'd agree, but given the flamage the template format has produced in the past, I'd really rather someone else proposed changing it... 8-)
Never mind changing it; what about the fact that ISOCNZ published a policy before the old "test" domainz whois server was ever available mandating the output to be consistent with RIPE-181? The ISOCNZ policy document was actually drafted following some consultation with the community, of course, which is perhaps what scared Domainz away from it. To be of optimal usefulness, why not code a whois server which can output in a variety of formats, including RIPE-181 and Domainz's own template format? whois -h whois.domainz.net.nz patho.gen.nz [RIPE-181 format data] whois -h whois.domainz.net.nz -t patho.gen.nz [Domainz template format data] Joe --------- To unsubscribe from nznog, send email to majordomo(a)list.waikato.ac.nz where the body of your message reads: unsubscribe nznog
Joe Abley
On Wed, 31 May 2000, Don Stokes wrote: Never mind changing it; what about the fact that ISOCNZ published a policy before the old "test" domainz whois server was ever available mandating the output to be consistent with RIPE-181?
Uh, how exactly is "Representation of IP Routing Policies in a Routing Registry" (http://www.ripe.net/ripe/docs/ripe-181.html) relevant? I think you're thinking of RIPE-049. The short answer is that the RIPE-049 format simply doesn't match the data held by Domainz. Your own whois demonstrates this by burying a bunch of info in the remarks: field; frankly, I wouldn't want to have anything but free-form comments in a remarks: field; as such I'd consider the RIPE format essentially unparseable for the fields that appear in the DRS format that aren't specified by RIPE-049. Note that the offending terminology in the WHOIS policy (http://www.isocnz.org.nz/whoispolicy.html) is "... with output complying with the standard RIPE format", and doesn't actually specify a document defining exactly what is meant. The whole point of the RIPE whois format is that it can be used as the basis of a template to be submitted to the database. Thus, when I wrote the spec for the whois (rather late in the development, with some time pressure), I, uh, creatively interpreted that policy to mean something that met the goals of the RIPE format first, met the syntactical requirements second and the precise set of fields last. Just as RIPE-181 isn't appropriate for a RIPE format DNS registry, RIPE-049 isn't appropriate for DRS, since DRS is not a RIPE format database. (Whether it should be is of course a completely separate argument...)
To be of optimal usefulness, why not code a whois server which can output in a variety of formats, including RIPE-181 and Domainz's own template format?
whois -h whois.domainz.net.nz patho.gen.nz [RIPE-181 format data]
whois -h whois.domainz.net.nz -t patho.gen.nz [Domainz template format data]
That's a possibility of course. Note that the RIPE '-t' option returns a *blank* template, so doing exactly the above would be terribly non-compilant. But there could be other ways of specifiying which format to use. Personally though, I'd be (mildly) opposed to making RIPE-049 the default format for the reasons outlined above. -- don --------- To unsubscribe from nznog, send email to majordomo(a)list.waikato.ac.nz where the body of your message reads: unsubscribe nznog
On Wed, 31 May 2000, Don Stokes wrote:
Joe Abley
wrote: On Wed, 31 May 2000, Don Stokes wrote: Never mind changing it; what about the fact that ISOCNZ published a policy before the old "test" domainz whois server was ever available mandating the output to be consistent with RIPE-181?
Uh, how exactly is "Representation of IP Routing Policies in a Routing Registry" (http://www.ripe.net/ripe/docs/ripe-181.html) relevant?
4. The Routing Registry and the RIPE Database One of the activities of RIPE is to maintain a database of Euro- pean IP networks, DNS domains and their contact persons along with various other kinds of network management information. The database content is public and can be queried using the whois protocol as well as retrieved as a whole. This supports NICs/NOCs all over Europe and beyond to perform their respective tasks. The RIPE database combines both allocation registry and routing registry functions. The RIPE allocation registry contains data about address space allocated to specific enterprises and/or delegated to local registries as well as data about the domain name space. The allocation registry is described in separate documents [6,7] and outside the scope of this document. One of the points made at the time that John was asking for feedback on the draft policy was that the whois server should be capable of supporting a route registry for NZ operators as well as a respository for domain names, which perhaps helps to explain where 181 came from. Of course the idea of entrusting Domainz with routing policy now is rediculous, but at the time UoW was providing a good sanity layer between management idiocy and the network and this didn't seem like such a silly idea.
I think you're thinking of RIPE-049.
The short answer is that the RIPE-049 format simply doesn't match the data held by Domainz. Your own whois demonstrates this by burying a bunch of info in the remarks: field; frankly, I wouldn't want to have anything but free-form comments in a remarks: field; as such I'd consider the RIPE format essentially unparseable for the fields that appear in the DRS format that aren't specified by RIPE-049.
The long answer is that to be useful, an automated interface should return data which is easy to parse. Since there are many RIPE-181-format whois servers in the world, and a correspondingly large number of existing clients (user-driven and machine-driven), it makes sense to use the convenient format.
Note that the offending terminology in the WHOIS policy (http://www.isocnz.org.nz/whoispolicy.html) is "... with output complying with the standard RIPE format", and doesn't actually specify a document defining exactly what is meant.
But the conclusion that it refers to a whois server returning data in the same manner as whois.ripe.net is hardly difficult to reach.
The whole point of the RIPE whois format is that it can be used as the basis of a template to be submitted to the database. Thus, when I wrote the spec for the whois (rather late in the development, with some time pressure), I, uh, creatively interpreted that policy to mean something that met the goals of the RIPE format first, met the syntactical requirements second and the precise set of fields last.
Is that an apology, Don? :)
Just as RIPE-181 isn't appropriate for a RIPE format DNS registry, RIPE-049 isn't appropriate for DRS, since DRS is not a RIPE format database.
You seem to have missed the point. The vast majority of the hits I get on the RIPE-alike whois server running on maggie.automagic.org are from outside the country, and I would bet very few of them are concerned with registering a domain (they're more likely looking up e-mail addresses to complain about network abuse). Returning a template suitable for resubmission is pointless. Returning a result in a format that is different to every other whois server on the planet is just unhelpful.
To be of optimal usefulness, why not code a whois server which can output in a variety of formats, including RIPE-181 and Domainz's own template format?
whois -h whois.domainz.net.nz patho.gen.nz [RIPE-181 format data]
whois -h whois.domainz.net.nz -t patho.gen.nz [Domainz template format data]
That's a possibility of course. Note that the RIPE '-t' option returns a *blank* template, so doing exactly the above would be terribly non-compilant. But there could be other ways of specifiying which format to use.
That's right! Non-compliance is bad!
Personally though, I'd be (mildly) opposed to making RIPE-049 the default format for the reasons outlined above.
With respect, your personal views don't count for much. John polled the community at the time he wrote the ISOCNZ policy document, and the consensus was that a RIPE-format whois server was what was required. Joe --------- To unsubscribe from nznog, send email to majordomo(a)list.waikato.ac.nz where the body of your message reads: unsubscribe nznog
Joe Abley
But the conclusion that it refers to a whois server returning data in the same manner as whois.ripe.net is hardly difficult to reach.
Yes, that is the "obvious" interpretation. But the interpretation I prefered to go with, as being more useful to the direct community of interest was to look at what the RIPE format meant in actual practical terms, ie
that met the goals of the RIPE format first, met the syntactical requirements second and the precise set of fields last.
In my view, blithely using RIPE-049 without considering the intent of the RIPE formats would have been just as non-compliant and less useful than doing what I suggested.
Is that an apology, Don? :)
Absolutely not. I was asked to make a call and I made one. I haven't seen anything new on the subject that makes me want to change my mind. I'm sure that if there was convincing evidence that your interpretation is better than mine that can be accommodated.
resubmission is pointless. Returning a result in a format that is different to every other whois server on the planet is just unhelpful.
A quick survey of the whois servers listed in the whois-servers.net
domain gives 11 whois servers using the RIPE format (or at least
something that looked like it) and 26 that didn't. Three used Rwhois.
Note that about 48 domains (mostly smallish European domains) use RIPE
itself (ie whois.ripe.net). The count above is servers, not domains.
If we're going to count the likelihood of needing to use a RIPE server
to look up a given domain, note that the count above doesn't include the
COM/NET/ORG registrars, pretty much none of whom use the RIPE format.
In short, the claim that "every other whois server on the planet" uses
RIPE, or that the format used is "incompatible with everything else in
the world" is spurious at best. If you want to use "market share" as
the basis for choice, the NSI format would have to be the clear winner.
There's a strange kind of circular argument here. The RIPE format whois (and in some sense this was a mandate for a RIPE format database) was advocated/mandated by ISOCNZ as a matter of policy for Domainz long before the work commenced on the implementation of the current system. So much so that the bid that Netlink submitted for the replacement system advocated the RIPE format.
Andy, with all due respect, the RIPE specification in the document was put there at *your* urging. I hadn't really thought about the implications for the application already being built at the time (since my involvement with the application itself was and is fairly minimal). I can't speak for other Technical Committee members, but I suspect that they hadn't either. The (brief) discussion on the RIPE format was in the September - October timeframe (the draft went out for comment on 12 Oct), well after development was started. It may have been mentioned earlier, but that was the point the word RIPE went into the draft document. The database schema was (IIRC) drawn up as part of the tender response well before that. I'm not aware of any prior policy referring to RIPE. Your paragraph above implies that your insistence on the RIPE format at Technical Committee was a deliberate attempt by a losing bidder to influence the development via a supposedly non-partisan forum. I *hope* this isn't the correct interpretation... As to the database schema itself: it's not my baby. Questions, moving on: Ignoring what Domainz was or wasn't asked to do, do you think the whois server should be a part of the registry interface (which includes the email template) or just a casual query tool? What does it need to do to satisfy users' requirements, and why? -- don --------- To unsubscribe from nznog, send email to majordomo(a)list.waikato.ac.nz where the body of your message reads: unsubscribe nznog
On Wed, 31 May 2000, Don Stokes wrote:
Joe Abley
wrote: But the conclusion that it refers to a whois server returning data in the same manner as whois.ripe.net is hardly difficult to reach.
Yes, that is the "obvious" interpretation. But the interpretation I prefered to go with, as being more useful to the direct community of interest was to look at what the RIPE format meant in actual practical terms, ie
The direct community of interest (based on the hits I see on my whois servers) are not people who want to register domains under .NZ; they are people trying to look up domain contact addresses from elsewhere.
In my view, blithely using RIPE-049 without considering the intent of the RIPE formats would have been just as non-compliant and less useful than doing what I suggested.
<shrug>
Is that an apology, Don? :)
Absolutely not. I was asked to make a call and I made one. I haven't seen anything new on the subject that makes me want to change my mind.
<shrug>
I'm sure that if there was convincing evidence that your interpretation is better than mine that can be accommodated.
<shrug>
BTW: Joe, wasn't it you who once wrote to isocnz-l:
| RIPE and APNIC aren't actually name service organisations; they're | regional registries for ASN and IP number allocations. You might | use CRSNIC, NSI, Nominet, etc if you're looking for examples of | registries' and registrars' use of whois.
during the consultation process for the ISOCNZ draft? None of the above use anything even vaguely resembling RIPE-049.
It was. But I'm not sure what your point is here, Don. RIPE defined the format of the replies to their whois server; I was trying to correct a specific mis-wording in John's policy. Joe --------- To unsubscribe from nznog, send email to majordomo(a)list.waikato.ac.nz where the body of your message reads: unsubscribe nznog
At 12:46 PM 5/31/00 +1200, Don Stokes wrote:
The short answer is that the RIPE-049 format simply doesn't match the data held by Domainz. Your own whois demonstrates this by burying a bunch of info in the remarks: field; frankly, I wouldn't want to have anything but free-form comments in a remarks: field; as such I'd consider the RIPE format essentially unparseable for the fields that appear in the DRS format that aren't specified by RIPE-049.
There's a strange kind of circular argument here. The RIPE format whois (and in some sense this was a mandate for a RIPE format database) was advocated/mandated by ISOCNZ as a matter of policy for Domainz long before the work commenced on the implementation of the current system. So much so that the bid that Netlink submitted for the replacement system advocated the RIPE format. I think it's a pretty spurious argument to say the RIPE format doesn't match Domainz's requirements when the Domainz data structure was formulated after the event (and to the untrained eye it seems as though this was deliberate so that the monopoly held by Domainz would be hard to break). Certainly any request from ISOCNZ and the industry for this format has been studiously ignored on repeated occasions. My concern in all this is that we're very large numbers of dollars down the track - I've heard figures of $700,000 and more - and we have a format that's incompatible with everything else in the world --------- To unsubscribe from nznog, send email to majordomo(a)list.waikato.ac.nz where the body of your message reads: unsubscribe nznog
Domainz was instructed by ISOCNZ to provide a RIPE format "whois" in November 1999. This was after months of procrastination, and wallying around prior to this. (I estimate March/April 1999 that Domainz was really told, but never mind. Formal Instruction was issued in November 1999, in response to protracted procrastination.) Domainz were very severely slapped around at the last ISOCNZ council meeting for non-compliance. This was all done long before the DRS system was specified. (in fact the DRS tender/responses pretty much ignores most of the stuff network engineers care about, which is another matter entirely.) So it is extremely rich for Domainz, or it's representatives to plead that the required RIPE "whois" does not match the current Domainz database. I strongly suggest that several parties involved in this understand the chronological issue here. If Domainz would like to come back to ISOCNZ council and announce that it is unable to provide a RIPE format "whois", then perhaps it would like to consider how it might like to word that submission. Others may think this looks like wilful procrastination or technical incompetence, or a sublime mixture of the two. I couldn't possibly comment. Rgds Roger De Salis (is protracted procrastination good english???) Don Stokes wrote:
Joe Abley
wrote: On Wed, 31 May 2000, Don Stokes wrote: Never mind changing it; what about the fact that ISOCNZ published a policy before the old "test" domainz whois server was ever available mandating the output to be consistent with RIPE-181?
Uh, how exactly is "Representation of IP Routing Policies in a Routing Registry" (http://www.ripe.net/ripe/docs/ripe-181.html) relevant?
I think you're thinking of RIPE-049.
The short answer is that the RIPE-049 format simply doesn't match the data held by Domainz. Your own whois demonstrates this by burying a bunch of info in the remarks: field; frankly, I wouldn't want to have anything but free-form comments in a remarks: field; as such I'd consider the RIPE format essentially unparseable for the fields that appear in the DRS format that aren't specified by RIPE-049.
Note that the offending terminology in the WHOIS policy (http://www.isocnz.org.nz/whoispolicy.html) is "... with output complying with the standard RIPE format", and doesn't actually specify a document defining exactly what is meant.
The whole point of the RIPE whois format is that it can be used as the basis of a template to be submitted to the database. Thus, when I wrote the spec for the whois (rather late in the development, with some time pressure), I, uh, creatively interpreted that policy to mean something that met the goals of the RIPE format first, met the syntactical requirements second and the precise set of fields last.
Just as RIPE-181 isn't appropriate for a RIPE format DNS registry, RIPE-049 isn't appropriate for DRS, since DRS is not a RIPE format database.
(Whether it should be is of course a completely separate argument...)
To be of optimal usefulness, why not code a whois server which can output in a variety of formats, including RIPE-181 and Domainz's own template format?
whois -h whois.domainz.net.nz patho.gen.nz [RIPE-181 format data]
whois -h whois.domainz.net.nz -t patho.gen.nz [Domainz template format data]
That's a possibility of course. Note that the RIPE '-t' option returns a *blank* template, so doing exactly the above would be terribly non-compilant. But there could be other ways of specifiying which format to use.
Personally though, I'd be (mildly) opposed to making RIPE-049 the default format for the reasons outlined above.
-- don --------- To unsubscribe from nznog, send email to majordomo(a)list.waikato.ac.nz where the body of your message reads: unsubscribe nznog
-- \_ Roger De Salis Cisco Systems NZ Ltd ' +64 25 481 452 L8, ASB Tower, 2 Hunter St /) +64 4 496 9003 Wellington, New Zealand (/ roger(a)desalis.gen.nz rdesalis(a)cisco.com ` --------- To unsubscribe from nznog, send email to majordomo(a)list.waikato.ac.nz where the body of your message reads: unsubscribe nznog
Domainz was instructed by ISOCNZ to provide a RIPE format "whois" in November 1999. This was after months of procrastination, and wallying around prior to this. (I estimate March/April 1999 that Domainz was really told, but never mind. Formal Instruction was issued in November 1999, in response to protracted procrastination.)
Domainz were very severely slapped around at the last ISOCNZ council meeting for non-compliance.
None of this would be a problem if there was a shared registry in place. I see ISOCNZ council has thrown out the hine report because it strongly recommended the introduction of a shared registry model for .nz Shame on them Regards Peter Mott Chief Enthusiast 2day.com -/- --------- To unsubscribe from nznog, send email to majordomo(a)list.waikato.ac.nz where the body of your message reads: unsubscribe nznog
Peter Mott wrote:
I see ISOCNZ council has thrown out the hine report because it strongly recommended the introduction of a shared registry model for .nz
Not true. The Hine report has generated considerable debate within council, and I know of several other papers that have been generated. The issue currently is "should the paper be adopted as is" or what changes/re-write might be appropriate. I have considerable sympathy with the authors desire not to re-write the report en-mass, given time requirements. At the moment, I am not aware of a conclusion either way. R
Shame on them
Regards
Peter Mott Chief Enthusiast 2day.com -/-
--------- To unsubscribe from nznog, send email to majordomo(a)list.waikato.ac.nz where the body of your message reads: unsubscribe nznog
-- \_ Roger De Salis Cisco Systems NZ Ltd ' +64 25 481 452 L8, ASB Tower, 2 Hunter St /) +64 4 496 9003 Wellington, New Zealand (/ roger(a)desalis.gen.nz rdesalis(a)cisco.com ` --------- To unsubscribe from nznog, send email to majordomo(a)list.waikato.ac.nz where the body of your message reads: unsubscribe nznog
I see ISOCNZ council has thrown out the hine report because it strongly recommended the introduction of a shared registry model for .nz
Not true. The Hine report has generated considerable debate within council, and I know of several other papers that have been generated. The issue currently is "should the paper be adopted as is" or what changes/re-write might be appropriate.
Prof hine waltzes all over the country side, getting views from people who are interested, puts together a useful document and ISOCNZ council want it to be changed or re-written. Unbelievable! Regards Peter Mott Chief Enthusiast 2day.com -/- --------- To unsubscribe from nznog, send email to majordomo(a)list.waikato.ac.nz where the body of your message reads: unsubscribe nznog
Prof hine waltzes all over the country side, getting views from people who are interested, puts together a useful document and ISOCNZ council want it to be changed or re-written.
Unbelievable!
There you go again, Peter. Getting all confused over the meaning of the words "unbelievable" and "predictable". -- Andrew P. Gardner Never underestimate the power of stupid people in large groups. --------- To unsubscribe from nznog, send email to majordomo(a)list.waikato.ac.nz where the body of your message reads: unsubscribe nznog
On Wed, 31 May 2000 16:21:55 +1200, Peter Mott wrote:
I see ISOCNZ council has thrown out the hine report because it strongly recommended the introduction of a shared registry model for .nz
Not true. The Hine report has generated considerable debate within council, and I know of several other papers that have been generated. The issue currently is "should the paper be adopted as is" or what changes/re-write might be appropriate.
Prof hine waltzes all over the country side, getting views from people who are interested, puts together a useful document and ISOCNZ council want it to be changed or re-written.
It would obviously have been easier and saved a lot of time if the WG had ignored all the public meetings and consultations and just come up with the recommendations that those who were not on the WG decided the WG should have. DPF ________________________________________________________________________ <david at farrar dot com> NZ Usenet FAQs - http://www.dpf.ac.nz/usenet/nz ICQ 29964527 --------- To unsubscribe from nznog, send email to majordomo(a)list.waikato.ac.nz where the body of your message reads: unsubscribe nznog
participants (8)
-
Andy Gardner
-
Andy Linton
-
Don Stokes
-
DPF
-
Joe Abley
-
Michael Newbery
-
Peter Mott
-
Roger De Salis