On Thu Oct 15 02:34:42 2009, Peter Saint-Andre wrote:
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
On 10/14/09 3:07 AM, Dave Cridland wrote:
> On Wed Oct 7 22:38:05 2009, Joe Hildebrand wrote:
>> What about XEP-156? For web-only clients that don't have DNS
access,
>> they can usually have the polling URL configured in, and for
desktop
>> clients, they can look up a TXT record to figure out where to
poll.
>>
>> Why do we have to hard-code a path?
>
> I think the intention here is that web-only clients need the URL
> hard-coded in, and the argument is that because this is often
less than
> intuitive, it'd be useful if the vast majority of BOSH
implementations
> listened to, and expected, the same path.
Aha, so that an end user needs to provide only a domain name (e.g.,
bosh.jabber.org) and not also a port and a path?
Well, the domain name should be whatever domain the web client is
hosted on. But that may be a biased opinion, because I don't think
there's any really sensible use-case for BOSH outside of web clients.
> The thing is, my understanding is that for typical "real"
deployments,
> the path is reverse proxied by a "real" webserver anyway, so I
wonder
> how much work we're saving here.
It would be good to gather operational data on this point so that
we're
not making unjustified assumptions.
Right. I know some people firmly believe that BOSH is a sensible way
to circumvent firewalls. I don't care about this use-case.
> Finally, I'd just like to say U-NAPTR. Just to see the looks of
horror.
Abject terror here. Close enough to horror for you? :P
I think abject terror is only justifiable by use of full NAPTR,
particularly as part of DDDS.
U-NAPTR is the simple variant, right?
Dave.
--
Dave Cridland - mailto:[email protected] - xmpp:[email protected]
- acap://acap.dave.cridland.net/byowner/user/dwd/bookmarks/
- http://dave.cridland.net/
Infotrope Polymer - ACAP, IMAP, ESMTP, and Lemonade
