Re: Extensibility strategies, was: Deciding in public (Was: SVGWG SVG-in-HTML proposal)

Boris Zbarsky wrote:
> Julian Reschke wrote:
>> Because I believe that URIs are the safest and most widely used way 
>> for disambiguation
> 
> I can buy this.
>
>> and mixing URIs with free-form identifiers just is asking for clashes.
> 
> Clashes between the URIs and the free-form identifiers?  I'm not sure I 
> buy that.  You really think that someone will create a free-form 
> identifier that looks like a URI and not _think_ of it as a URI?

Well, people have been known to use a colon to separate prefixes from 
local names, so I wouldn't be too surprised.

>> If you think it is so unlikely, why not make a statement that 
>> something that parses as a URI must be a URI under the control of the 
>> party minting the identifier?
> 
> One could, but people will just ignore such a statement, pretty much 
> like every other authoring requirement.  Especially because this one is 
> not machine-verifiable.

Agreed. But that's also true for many other things the spec contains.

>>> None of these problems exist if you don't have prefixes.
>>
>> Yes, that can happen.
>>
>> That being said, I've been working with XML + namespaces since these 
>> specs came out, and that problem never *ever* occurred to me in practice.
> 
> There've been a number of Mozilla bugs on just such issues in the XML 
> serializer.  In practice, some people create DOMs that look just like 
> that, then move the DOM nodes around, then try to serialize as XML.  We 
> have a whole bunch of code to fix up prefixes as we go when serializing 
> to handle the mess.

Understood. When I needed an XML serializer years ago (in Java), I 
finally wrote my own, because the ones shipping back then were just too bad.

> I have no idea how common this is, but common enough for people to have 
> cared enough to file a dozen or so bugs over the last few years.
> 
>> - the spec to give rules for the formats of these names, so clashes 
>> are avoided (if you use a URI, use one you have authority over)
> 
> Seriously, if someone is using a URI they don't have authority over, 
> they are either copy/pasting (in which case they won't even know about 
> the proposed requirement) or doing it on purpose (in which case they 
> will ignore the requirement even if they know about it).

I agree these are likely reasons, and that the spec can't prevent that.

One other thing I see frequently is people using something like 
"http://example.org" (reasoning; it's just a name, so who cares). Maybe 
we could educate at least these people.

BR, Julian

Received on Tuesday, 5 August 2008 16:53:41 UTC