[uf-discuss] Addressing bits of information

[Chris Casciano]

On May 25, 2006, at 1:00 PM, Ben Ward wrote:

> Tantek Çelik wrote:
>> But discussions of extending URL/URI semantics for addressing bits of
>> information?  I'm not sure that actually has much bearing on parsing
>> microformats.  But like I said, perhaps I am missing something.
>
> I don't know if my interpretation brings this on topic or just 
> clarifies (or confuses…) but the problem as I understand it, is 
> exemplified like so:
>
> I have a page listing 12 events using hCalendar. None of the hevent 
> nodes have an ID and therefore cannot be referenced through fragments. 
> However, I want to pass one of these events to a web service (say a 
> Technorati pipe to generate an ICS file for the specific event). 
> There's no way to do that.
>
> Same thing with a page containing multiple hCards.
>
> However, I honestly think that this problem is best solved by 
> emphasising the need/benefit of putting IDs on vcard/vevent nodes, 
> rather than inventing/adopting something more complex. Personally, I 
> wouldn't think it unreasonable to have 'microformats must have an ID' 
> as part of each spec.
>
> I don't think we should overcomplicate the issue with talk of 
> technology like XPointer or Selectors which are unimplemented in the 
> areas needed to 'fix' this.
>
> Ben

Thanks for the clear post Ben.

This is very much the same issue that I've mentioned a few times WRT 
consuming hatom documents.... the added 'bonus' if you will with hatom 
is that you're typical usage will be to return to that document (or 
document fragment) repeatedly over time so its more important to handle 
some error conditions or other document changes over time.

I do think that 'root element must have id' goes a long way to solving 
the problem on the hatom side, but I don't know if that alone gets us 
all the way there.

And that says nothing about the questions of how a typical user might 
be able to identify the specific hcard/hevent/hatom/xoxo fragment to 
pass to another application if there was an ID requirement in place. In 
the above scenario you'd still have to have a way of learning that 
#event-002 is *the* event you want to spread around -- other then 
knowing because you're the author or you viewed source.

-- 
[ Chris Casciano ]
[ chris at placenamehere.com ] [ http://placenamehere.com ]