1 |
On 03/28/2013 03:11 PM, Stroller wrote: |
2 |
|
3 |
> The search I made before posting led me the wikipedia article which |
4 |
> mentioned, for example, using thick and thin client models. |
5 |
> |
6 |
> http://en.wikipedia.org/wiki/Whois#Thin_and_thick_lookups |
7 |
> |
8 |
> One might assume, for example, that a thin client might tend to give |
9 |
> more accurate and up-to-date information, but of course there's also |
10 |
> the issue that the whois server for the domain might move. Thus the |
11 |
> client might need to be updated in a timely manner, too. |
12 |
> |
13 |
|
14 |
The thin model sort of works like DNS, except everything is unstructured |
15 |
and totally made-up on the server side and guessed-at on the client |
16 |
side. The clients are trying to parse the unstructured output, like you |
17 |
would if you were trying to screen scrape a webpage. As of ten seconds |
18 |
ago, this is what I get for a lookup of orlitzky.com: |
19 |
|
20 |
Domain Name: ORLITZKY.COM |
21 |
Registrar: GANDI SAS |
22 |
Whois Server: whois.gandi.net |
23 |
Referral URL: http://www.gandi.net |
24 |
... |
25 |
|
26 |
The "Whois Server:" for the domain is something like an NS record, where |
27 |
the guy higher up points you at the next level down. If the whois server |
28 |
for the domain changed, you wouldn't need to update the client -- you |
29 |
could just ask the top-level server for it again. What *would* make you |
30 |
update the client is if, say, that top-level server started outputting a |
31 |
space between "Server" and ":". |