[Standards] Stamping on one's head Re: Fwd: Minutes 20121011

Ralph Meijer ralphm at ik.nu
Wed Oct 17 11:47:02 UTC 2012


On 2012-10-13 09:37, Andreas Kuckartz wrote:
>> 4) Any other business
>> Brief discussion of XEP 71, to continue on standards list.
>
> This made me curious and I belong to those who reads conference minutes.
> And to my astonishment I found this:
>
> [15:18:46] <stpeter> I have one AOB.... XEP-0071 and how to come to
> agreement on structured elements vs. stylistic presentation, but I
> suppose I'll just post to the list about that
> [15:18:50] <Kev> I started and then got distracted, as so often happens.
> [15:18:57] <ralphm> I wish to stamp that on the head of everyone
> involved the slightest in anything related to publishing web pages
> [15:19:23] <Kev> stpeter: I think at this stage the answer is "We don't
> substantially change the XEP".
> [15:19:34] <Kev> But yes, that discussion can happen on-list.

I'm sorry you got caught in a misinterpretation of the flow of that 
discussion, of which this snippet is incomplete. My statement was 
related to what I wrote directly above that, concerning the unneeded 
port number added to the URL of our wiki and that cool URLs don't 
change. It may be worded strongly, but in that context I stand by it.

I'd like to note that chat logs like don't convey the relation between 
messages and messages may be out of order due to multiple people typing 
at the same time. Also it may involve latency issues due to network 
connectivity or painful on-screen keyboards on phones while in transit. 
Not to mention attempts at humor or sarcasm.

Full logs can be found here at our archive, with the start of the thing 
I responded to pointed to by the time-based anchor: 
<http://logs.xmpp.org/council/121011/#15:16:35>

> ...
> [15:19:47] <ralphm> To be more blunt. I'm a solid -1 on anything adding
> stuff
> [15:19:55] <ralphm> like rdfa

This was a direct response to Kevin's remark:
 > [15:19:23] <Kev> stpeter: I think at this stage the answer is "We 
don't substantially change the XEP".

> According to http://xmpp.org/about-xmpp/xsf/xmpp-council/ ralphm is
> currently a member of the XMPP Council.
>
> Can someone please let me know if he has a veto right there?

Like Kevin and Peter already mentioned I don't agree with changing this 
XEP this late in the game. In practice we see that implementations 
already have security issues because they pass the XHTML directly to a 
generic HTML rendering engine, without properly filtering elements and 
(style) attributes that are not included in this profile.

We have found that people tend to implement specifications by example, 
and then maybe read the surrounding prose. I know I have sometimes. Even 
though existing implementations should ignore the RFDa if they implement 
the current version of the spec, I am afraid they don't, and I believe 
that allowing more at this stage will promote implementations to not 
filter at all.

 > If that is the case there would be no reason for me to spend a minute
 > more on finding a reasonable way to add RDFa to XEP-0071. I generally
 > do not participate in discussions with a predetermined outcome.

I encourage any proposals to add or amend the specifications published 
by the XSF and generally tend to accept any proposed XEP as experimental 
because I believe it encourages discussing its merits. However, as 
council member I am tasked to take this discussion as input and then 
make a judgment call on advancing specifications in our process.

In this specific example, you engaged in discussing the addition of RDFa 
to XEP-0071 and I responded at length on this list [1], as did others. 
To conclude, I believe we should not add anything to XEP-0071 in 
general, and think that conveying RFD triples in messages can done 
without embedding it in this XHTML profile specifically.

That said, I do appreciate your input, even though I may disagree on 
this point. I'd sincerely regret it if you are discouraged by this and 
would still like to support any efforts to convey RDF triplets in 
another way.

Thank you for bringing this up.

[1] <http://mail.jabber.org/pipermail/standards/2012-October/026850.html>

-- 
ralphm



More information about the Standards mailing list