[uf-discuss] URI schemes vs. visible data (was Re: communications log, "tel" microformat?)

Tantek Ç elik tantek at cs.stanford.edu
Tue Nov 29 18:15:48 PST 2005

On 11/26/05 1:07 AM, "Benjamin Carlyle" <benjamincarlyle at optusnet.com.au>

> Tantek,
> On Thu, 2005-11-24 at 10:25 -0800, Tantek Çelik wrote:
>> In the mean time, as Ryan King as pointed out, there is nothing stopping you
>> from simply marking up your phone calls using semantic class names, which
>> you could re-use from hCard, to note the phone number in particular, e.g.:
>> <span class="tel">+1.415.555.1212</span>
> I know there are various groups on the Internet, all wanting to encode
> the information of the world into their scheme. This is a neat
> microformat approach, but an alternative might be to use a telephony
> URI[1]. The number above might be represented instead as something like:
> <a href="tel:+1.415.555.1212">+1.415.555.1212</a>

As Karl pointed out, it has been documented as one of the techniques for
enhancing your hCard.

However, the larger issue is that of using URI schemes vs. human visible

> There are URI schemes for a variety of identifiable things.

Most of which have seen infrequent if any use at all.

Part of what microformats consider is whether or not a previous
scheme/standard has any adoption, or more to the point, interoperable

> It may save
> effort on the microformat side of the fence to make use of these where
> available. 

It's actually more work to do so, on the part of the content authors.

The point is, people are already publishing this information *visibly* on
their web pages, e.g.


or *maybe*


The point of microformats is to provide the minimum necessary to allow that
information to be automatically recognized and parsed, and in the case of
hCard, you can do this like this:

<span class="tel">+1.415.555.1212</span>

That's less work/change than

<a href="tel:+1.415.555.1212">+1.415.555.1212</a>

Of course there is nothing saying that you can't do both:

<a class="tel" href="tel:+1.415.555.1212">+1.415.555.1212</a>

And in fact, as Karl has documented on the hcard-brainstorming page.

> Of course, the function of the scheme and the class may
> differ. 


> While it is clear that I'm identify a telephone number in my
> href, it isn't clear what the telephone number means. Having a
> microformat that indicates whether the telephone is for home or business
> use and who it belongs to may be a more important focus.

That too.

> To possibly rub you guys up a little bit more of the wrong way,

Not at all.

> perhaps
> it is even worth reexamining the geo and addr microformats in favour of
> developing URI-based alternatives.

It's an interesting question to consider.

Here is the key. 

Authors aren't publishing links to geo and address information.

They're publishing *visible text* of geo and address information.

So the easiest thing to do, for the author, is to leave it as visible text.

Thus, it makes the most sense to do the simple thing of just wrapping that
visible text with a little bit of markup, rather than asking the author to
move (or copy) it into an attribute, which may or may not require a
reformatting of the data as well.

I can imagine, that for geo or addr information, maybe they'll want to wrap
it in a hyperlink that goes to a maps page or something, so that clicking it
actually does something.  If you forced them to use a hypothetical "geo:"
protocol instead, then that would interfere, since you can only hyperlink
something to one destination.



P.S. Note to self/community, add "URI schemes vs. visible data" this to the
microformats FAQ.

More information about the microformats-discuss mailing list