I sometimes wonder, since gtlds are not permanent fixtures (they are sort of subscription payments to ICANN), and a lot of people are seemingly removing their delegations now, presumably to save money, are we going to hit the point where URLify function is not gonna have any idea what is currently a fully qualified host name and what is currently not?
Will anything remember the fact that comcast was once a TLD?
@benjojo how many organisations make any real use of their .brand TLDs? The only one that I can recall seeing in the wild is .ntt.
ICANN made a load of money selling shovels in a gold rush.
@edward oh I think it's pretty abysmal. I know NTT uses their gTLD (But at least I know from anecdotal data that their employee's still use their older emails rather than the gTLD one)
(I wrote a bit about this in 2018 when Sony let go of a couple of their gTLD's)[https://blog.benjojo.co.uk/post/the-death-of-a-tld] and it seems that sometimes it's used for email, but not much else.
In the case of comcast, they had:
and that is basically it...
@benjojo @edward I think a lot about Barclays’ status page for retail banking- https://status.uk.barclays/
@unlobito @benjojo @edward Stockholm municipality has .stockholm that is exclusively for the municipality, not even Stockholm county stuff is under it. so I guess it's a brand TLD and it does get use as most municipal things are now under it.
@benjojo it’s relatively easy to get a copy of the root DNS zone.
When will things (other than DNS servers) use this as priming data?
@drscriptt But regexs are easy :), To be fair, the safer (and easier than zone file parsing) is to just test if the suspected DNS name actually resolves or not, before "URLifying" it.
That being said, that does carry risks such as described in this iTerm2 bug report https://gitlab.com/gnachman/iterm2/-/issues/6050
@benjojo I hear you but I raise you named-compilezone et al. That produces simple line oriented <record owner name> <record data> that is very safe to pass through awk et al. to get the first column that is the TLDs in the root zone.
And I think about people parsing (read: deserializing) XML, YAML, and JSON, all of which are more complex to parse than tabular data.