(In declining order of information - I also couldn't find the info for
AFRINIC queries) I also couldn't find information on what fields they
have (and obviously how they map to one another). There are also a few
other whois servers around that I have no idea about: https://www.opensource.apple.com/source/adv_cmds/adv_cmds-149/whois/whois.c
Heh, heh, heh. There are just about as many whois output formats as there are back-end data-stores. Note that I say “data-stores” rather than databases. Some of them aren’t. So when you say “title” I assume you’re referring to half of a key-value pair. A concept some large whois sources don’t have.
So, you’re not running into a poorly-documented mystery, you’ve run afoul of one of the rotten armpits of the shub-Internet.
Is there a list of NIC (and other popular whois server) features (what
can be searched on) and what data they provide (and what title they
give it)?
Heh, heh, heh. There are just about as many whois output formats as there are back-end data-stores. Note that I say “data-stores” rather than databases. Some of them aren’t. So when you say “title” I assume you’re referring to half of a key-value pair. A concept some large whois sources don’t have.
Yes, I'm referring to mapping between key names.
So, you’re not running into a poorly-documented mystery, you’ve run afoul of one of the rotten armpits of the shub-Internet.
So there's no consensus between NICs for the information they should
have in whois and what search mechanisms they should provide? I guess
what you're saying is that whois is just a protocol definition and
nothing else?
> So, you’re not running into a poorly-documented mystery, you’ve run
afoul of one of the rotten armpits of the shub-Internet.
>
So there's no consensus between NICs for the information they should
have in whois and what search mechanisms they should provide? I guess
what you're saying is that whois is just a protocol definition and
nothing else?
Might not even qualify for "protocol definition", and RFC 3912 ACKs this:
So, you’re not running into a poorly-documented mystery, you’ve run afoul of one of the rotten armpits of the shub-Internet.
So there's no consensus between NICs for the information they should
have in whois and what search mechanisms they should provide? I guess
what you're saying is that whois is just a protocol definition and
nothing else?
Correct. It gets you a blob of text. Sometimes, a blob is just a blob. Other times, it contains what _appear_ to be key-value pairs, but are instead loosely-formatted text. Other times, it contains textually-represented key-value pairs that are programmatically generated from an actual database, and can thus be re-imported into another database. Depends what’s on the back end.
Scripting languages have modules that can parse many registrar whois
formats. However, most are incomplete due to the plurality of output
formats as stated above. I, and i suspect many others, wouls *love* to see
a more concrete key value format drafted and enforced by ICANN.
Yes, that's what I was looking at. And that REST API looks nice...
Though from what I read (admittedly not the whole doc yet) I didn't
see it defined what type of data is returned, nor what data should be
expected, which would leave me in the same place. If I'm only getting
a blob back (that would be, I guess, internationalized at this point)
I've still got to loop through with a regex expecting some type of
key/value thing and concatenate data after a line break to the last
value (probably removing spaces since they try to format it pretty).
Correct. It gets you a blob of text. Sometimes, a blob is just a blob.
Other times, it contains what _appear_ to be key-value pairs, but are
instead loosely-formatted text. Other times, it contains
textually-represented key-value pairs that are programmatically
generated from an actual database, and can thus be re-imported into
another database. Depends what’s on the back end.
If only we could create a committee to fix whois...
Scripting languages have modules that can parse many registrar whois
formats. However, most are incomplete due to the plurality of output
formats as stated above. I, and i suspect many others, wouls *love* to see
a more concrete key value format drafted and enforced by ICANN.
CRISP is dead. RDAP is real. If people need to script, then RDAP is
workable JSON and for once, has converged on sensible stuff in both names
and numbers.
the whois "problem" is a formalism owned by ICANN, but as DRC pointed out
the WHOIS solution is dispersed.
RPSL lies to one side btw. I wish it was a defined RDAP type. It may be
work-in-progress but thats not clear.
If only we could create a committee to fix whois...
Quite astonishingly, the IETF WEIRDS working group finished
successfully, and its documents will be published as RFCs when they
get through the editing queue in a month or two. The protocol is
called RDAP, the queries are http, the results are json.
ARIN, APNIC, and RIPE have prototypes already that are a lot easier to
script than the text WHOIS.
RDAP is also supposed to work for domain name WHOIS, but the timeline
there is less clear. ICANN has hired cnnic to do a freeware version
which we hope a lot of smaller registries will use.
Awesome thanks. That answers half of my original question (though this
route was much more insightful than I thought). I can run with that (php
isn't my language but the etl is pretty clear).