Le mercredi 18 octobre 2017, 19:19:53 CEST Florian Schmaus a écrit :
> On 18.10.2017 18:47, Goffi wrote:

> > CommonMark is specified, but how can you garantee that the dev will use
> > libraries compliant with CommonMark?
> 
> No one can, but that applies to all approaches, even to <body/> text.

It would not apply to a syntax with only one specified way to write things. 
The thing is there are dozen of libraries (which probably don't even all 
support CommonMark), and devs not aware of Markdown issues will take the first 
one they find, resulting in troubles.

> I think the security situation is, while not perfect, far better
> compared to XHTML-IM.

If we base the debate on devs not really taking care of security (which was 
the initial issue with XHTML-IM) or path of less resistance, they will most 
probably just send the raw Markdown to the list, were HTML can be executed.

At least with XHTML(-IM or not) we KNOW that it must be cleaned.
  

> > It would be used a a rich syntax, so it is definitely relevant
> 
> I really don't think so.

Why would people put Markdown if there is not intent to use rich markup then?



> > The ProtoXEP says "This list is not comprehensive and not meant to be.",
> > so it clearly means that any flavour can be used.
> 
> Good point. I didn't notice that this could be misinterpreted. What I
> tried to express is that the list does not include every existing markup
> language. But only the "Markup Language Identifier" defined in the XEP
> can be used with BMH. And I plan to specify some requirements for markup
> languages to be included in that list. Like the already mentioned
> requirement that the language must be reasonably readable when treated
> as plain text.

OK, that's a bit better, but still doesn't invalidate other points.


> > Again allowing a protoXEP like this would mean using it as a rich syntax
> > vectore, so the issue is relevant.
> 
> No, it is totally irrelevant. Please recall the situation BMH tries to
> improve: You already have the data formatted using e.g. CommonMark.
> Whether or not CommonMark supports feature X does not matter.

Where and which clients do already put CommonMark in regular <body>? I've 
never seen that, I've always had XHTML-IM when somebody wanted to put rich 
text.


> > And what do you do when you have HTML tags? You put them raw? What if
> > Markdown syntaxes are inside? Tou remove? how can I send HTML then?
> 
> Similar as above: You usually wont have custom HTML in those situations.
> And if you want to send CommonMark+HTML formated data, then you should
> not use BMH.

My point was not to put HTML to be interpreted, but to put HTML to be 
displayed like text, when you want to paste some code for instance.


As an additional remark, don't forgot that you can have many <body>, and also 
<subject>, there is nothing talking about that in the protoXEP.



I think the best way to fix the situation is to have a (single!) syntax rich 
and extensible enough to satisfy everybody.


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

Reply via email to