On Sat, Oct 14, 2017, at 07:06, Jonas Wielicki wrote:
> I am still not keen on obsoleting XHTML-IM before we have an actual 
> alternative ready. I don’t think that this will achieve anything good. 
> Instead, I think that one of two things will happen:

Someone suggested in the council meeting today that these specific
points have not been addressed, so I'd like to make a few observations:

> (a) Clients continue to implement XHTML-IM because it is the only actual
>     way to convey markup right now (this is what I’ll do until there’s a
>     replacement).

With regards to (a), I think that is exactly what will happen, and I am
okay with that. By not obsoleting it, clients will also continue to
implement it, so this doesn't really change things either way. All we
are doing by obsoleting it is saying that we, the XSF, do not recommend
new implementations of this protocol because it has a history of
security issues. Naturally people may still implement it for
compatibility. At best we stop a few new implementations from running
into the same security issues we've seen time and time again, at worst
there is no effect, so this doesn't seem like a compelling argument for
not obsoleting before a replacement is ready to me.

> (b) The ecosystem will fracture in islands of different, underspecified, 
>     plain-text markups put in <body/>.

With (b) I think that's only likely to happen if the council decides to
accept multiple different formatting specs as experimental and work on
all of them in parallel. With my council hat on, I don't think that's
likely to happen. I would also be interested in a protactive measure to
prevent this, say, starting a group to come up with requirements and
eventually a spec for the next generation of formatted messages (this is
something I was already planning on proposing, I'm sure many people who
have spoken up on this list would be interested in working together
towards a single spec).

> I also wonder what it would look  like to have the only markup protocol with 
> actual deployment being obsoleted 

I don't think it would look any different than today. Clients that
support it will continue to support it, clients that don't support it
probably won't. Maybe one or two people add an implementation, maybe one
or two people drop it, but it's not likely to drastically change
anything in the short term.

I'd also like to point out that not deprecating or obsoleting old XEPs
that we don't recommend is one of the reasons fragmentation happens. For
instance, I've had several people express confusion to me about which
spec they should use for archiving (this community knows that MAM is
recommended and what most implementations do, the layman however sees
that Message Archiving has a big green banner at the top and starts to
implement that). Another example is the fragmentation between Privacy
Lists (now deprecated, thankfully) and the Blocking Command. For quite a
while people insisted that we couldn't deprecate Privacy Lists until
there was a feature-parity alternative, and this led to lots of client
and server incompatibilities, these were mostly resolved before we
deprecated Privacy Lists, but it took a lot of work. Since XHTML-IM
isn't just a spec with a few issues that make it hard to implement, it
causes security issues, this seems like a good reason to stop
recommending it as the way forward. In fact, I'd go so far as to say
that it is irresponsible of the XSF to continue recommending a protocol
with a security history as bad as XHTML-IMs regardless of whether the
spec is "technically" secure or not (technically doesn't count for
anything for the reasons outlined by multiple other people numerous
times in this thread; in the real world, it's broken regardless of what
the spec says and we can certainly do better).


—Sam
_______________________________________________
Standards mailing list
Info: https://mail.jabber.org/mailman/listinfo/standards
Unsubscribe: [email protected]
_______________________________________________

Reply via email to