[Standards] File hosting XEP?

Sergey Dobrov binary at jrudevels.org
Fri Aug 17 08:36:48 UTC 2012

On 08/16/2012 10:17 PM, Simon Tennant wrote:
>     Yes, sorry, you're right. I just meant that buddycloud invents it's own
>     protocols instead of edit existent XEPs, write new, with an XMPP
>     community participation. It's a valid approach but differs from mine.
> Ok I'll bite.
> buddycloud is built on pub-sub and federates using XMPP.
> buddycloud designs social software to power millions of users using
> open-source, open protocols and provides great reference implementations.
> To do this we have two choices:
>  1. deploy a XEP-0060 pub-sub server and hope that uses suddenly find it
>     useful even through it was really designed for machine to machine
>     communication. Now you have to force users to start using it. And
>     they don't.
>  2. Or, look at what users really want, try and adapt standard so that
>     developers can use existing design fundamentals and existing
>     libraries and match these up with your users' real-world needs.
> So, I agree - buddycloud isn't pure XEP-0060 and that's totally fine by
> me. My mom cares about communicating, not XEP-0060 and there's a lot
> more "my moms" out there than developers than developers sitting aroudn
> the spec campfire. Developers care about being able to pick up a pub-sub
> library and build. To me that's a win for both parties.

But we have a lot of services which can provide us with their
experience, so the problems which will arise are obvious right now and I
am trying to solve them because I can't be sure that something will not
be changed in incompatible way earlier. Anyway, I just dislike
buddycloud architecture. But that's just mine subjective opinion and I
really glad that buddycloud exists in our universe.

Anyway, I think the topic is interesting but should be separated in an
another thread if we want to continue to flame about it.

> Following on from this (and we're getting closer to this point now) when
> the protocol kinks have been worked out, we loop back and plough that
> learning into a spec (we're tracking our pub-sub adaptions here:
> https://buddycloud.org/wiki/XMPP_XEP).
> tl;dr: writing a spec first without real users and a real product means
> your spec will sit on a shelf gathering dust.

I have a few real users and I can't provide all needed features even to
them, why have I think about someone else then? :)

> S.
> -- 
> Simon Tennant | buddycloud.com <http://buddycloud.com> | +49 17 8545 0880

With best regards,
Sergey Dobrov,
XMPP Developer and JRuDevels.org founder.

More information about the Standards mailing list