Multiple hCards - a problem?
Christian Scholz / Tao Takashi (SL)
I am wondering what happens if we put a link to a profile page with
hCard format into the service catalogue and this profile page has more
than one hCards.
Like a page where on top you have your profile information but in
another section a list of your friends, each with hCard markup.
I guess this will then be a problem to choose which profile
information to use or am I missing something?
(of course I can ask the user but that rather inconvenient).
-- Christian
--
Christian Scholz
Tao Takashi (Second Life name)
taota...@gmail.com
Blog/Podcast: http://mrtopf.de/blog
Planet: http://worldofsl.com
Company: http://comlounge.net
Tech Video Blog: http://comlounge.tv
IRC: MrTopf/Tao_T
Darren Bounds
hCard probably wasn't the best mechanism to discover information about
the owner of a URI/Identity Page. Rather it would make more sense to
establish a new service, YADIS discoverable, and had similar
characteristics to vCard but unlike hCard was designed purely for
machine consumption (e.g. straight XML, RDF, etc). We've been playing
around with it at Beemba but it's still in refinement and lacking a
name (e.g. http://darren.beemba.com/uoi/).
I'm curious what the group thinks about this.
Darren B.
Beemba.com
http://darren.beemba.com/
--
Thank you,
Darren Bounds
Julian Bond
>
>We asked that same question some time ago. Our conclusion was that
>hCard probably wasn't the best mechanism to discover information about
>the owner of a URI/Identity Page. Rather it would make more sense to
>establish a new service, YADIS discoverable, and had similar
>characteristics to vCard but unlike hCard was designed purely for
>machine consumption (e.g. straight XML, RDF, etc). We've been playing
>around with it at Beemba but it's still in refinement and lacking a
>name (e.g. http://darren.beemba.com/uoi/).
>
>I'm curious what the group thinks about this.
This is FOAF with the vcard namespace, isn't it? Why create another one?
--
Julian Bond E&MSN: julian_bond at voidstar.com M: +44 (0)77 5907 2173
Webmaster: http://www.ecademy.com/ T: +44 (0)192 0412 433
Personal WebLog: http://www.voidstar.com/ skype:julian.bond?chat
Contains Flammable Gas Under Pressure
Christian Scholz / Tao Takashi (SL)
page then maybe
hCard is not the way to go. Or the URL in the service catalogue needs
to point to
a HTML fragment identifier as the spec says to only parse this then.
On the FOAF side of things this seems not to be a problem as you can
put the FOAF
URL to _your_ FOAF file directly into the service catalogue and in
FOAF it's with the
VCard namespace pretty much defined.
On Sun, May 18, 2008 at 11:41 AM, Julian Bond <julia...@voidstar.com> wrote:
>
> Darren Bounds <dbo...@gmail.com> Sat, 17 May 2008 20:03:24
>>
>>We asked that same question some time ago. Our conclusion was that
>>hCard probably wasn't the best mechanism to discover information about
>>the owner of a URI/Identity Page. Rather it would make more sense to
>>establish a new service, YADIS discoverable, and had similar
>>characteristics to vCard but unlike hCard was designed purely for
>>machine consumption (e.g. straight XML, RDF, etc). We've been playing
>>around with it at Beemba but it's still in refinement and lacking a
>>name (e.g. http://darren.beemba.com/uoi/).
>>
>>I'm curious what the group thinks about this.
Well, I wonder the same as Julian: Would it make sense to just use FOAF with
VCard namespace?
Darren Bounds
FOAF vocabulary has no mechanmism to define address information (e.g.
state/province, country, city, street address).
Darren B.
Beemba.com
Christian Scholz / Tao Takashi (SL)
On Sun, May 18, 2008 at 2:39 PM, Darren Bounds <dbo...@gmail.com> wrote:
>
> I suppose it depends on what you hope to glean from the document. The
> FOAF vocabulary has no mechanmism to define address information (e.g.
> state/province, country, city, street address).
Well FOAF not directly but you can add the VCard namespace to it:
http://www.w3.org/TR/vcard-rdf
(which reminds me that I need to update the technical recommendation
doc to list it)
Would that solve your problem? I think the advantage is that many
people use it already and there are RDF-Parsers out there which
hopefully understand it :-)
Darren Bounds
--
Thank you,
Darren Bounds
Julian Bond
>
>I suppose it depends on what you hope to glean from the document. The
>FOAF vocabulary has no mechanmism to define address information (e.g.
>state/province, country, city, street address).
Which is why a high proportion of the FOAF out there uses the vCard
namespace for that information.
--
Julian Bond E&MSN: julian_bond at voidstar.com M: +44 (0)77 5907 2173
Webmaster: http://www.ecademy.com/ T: +44 (0)192 0412 433
Personal WebLog: http://www.voidstar.com/ skype:julian.bond?chat
No Bits Of Pulp
Bob Ngu
On May 18, 2:33 pm, Julian Bond <julian_b...@voidstar.com> wrote:
> Darren Bounds <dbou...@gmail.com> Sun, 18 May 2008 08:39:07
Julian Bond
>
>Is it RDF+VCard or FOAF+VCard?
It's kind of like this. It's all RDF. The FOAF namespace provides the
macro structure and the most common fields. The vCard namespace and
others provide the detail. So you typically have the following
namespaces
rdf, rdfs //basic RDF in XML
dc //document structure and things like document title
foaf // foaf:Person, foaf:mbox_sha1sum, etc for identifying people
foaf:knows for the simplest relationship between people
vCard // address data
geo // Lat-Long coordinates of their address
bio // CV-Resume type fields
A typical social network FOAF structure (much simplified) looks like
this:-
A document description block
A person description block about the "owner" of this data.
A long list of foaf:knows about this person's contact list each
containing at least:-
- A human readable name
- A machine readable identifier
- A link to their FOAF file.
For a fairly complex example of this see:-
http://www.ecademy.com/module.php?mod=network&op=foafrdf&uid=1
In the context of using FOAF for data portability as a transport format,
I think this is what we mean by a foaf file. It's:-
- RDF in XML
- It contains at least the foaf namespace, almost certainly the vcard
namespace and probably several others.
- It has a lot of info about one person
- It has a list with a little info about everyone that person follows
with links to their file.
- There's at least one identifying property against every person
referenced.
Now ultimately it's all RDF and there's a lot of RDF out there, used in
all sorts of ways. Some of it uses the FOAF namespace but would look
quite unlike what I've described above. But I think the above is the
bare minimum and the most understandable to specifically use this for
data portability. vCard on its own isn't enough. And FOAF on its own
isn't enough either.
--
Julian Bond E&MSN: julian_bond at voidstar.com M: +44 (0)77 5907 2173
Webmaster: http://www.ecademy.com/ T: +44 (0)192 0412 433
Personal WebLog: http://www.voidstar.com/ skype:julian.bond?chat
Schrodinger's Cat For President
Darren Bounds
services. As a consumer, why should I be expected to download a 200k
FOAF document when all I'm ever interested in 5k in basic identity
information?
FOAF files can be enormous and if they gain mass adoption, they will
be. Take your ecademy file for example. You have appoximately 620 FOAF
definitions in there. What if I had 1000? or 5000? Would the consumer
be expected to download a 1MB document every time they want to consume
updates about your identity? That's expensive and a waste.
Sure it exists today, but is it the right thing to do?
--
Thank you,
Darren Bounds
Julian Bond
>
>I really think identity discovery and FOAF file belong in distinct
>services. As a consumer, why should I be expected to download a 200k
>FOAF document when all I'm ever interested in 5k in basic identity
>information?
>FOAF files can be enormous and if they gain mass adoption, they will
>be. Take your ecademy file for example. You have appoximately 620 FOAF
>definitions in there. What if I had 1000? or 5000? Would the consumer
>be expected to download a 1MB document every time they want to consume
>updates about your identity? That's expensive and a waste.
>
>Sure it exists today, but is it the right thing to do?
One way to deal with this is to split the profile data from the contact
list data. It doesn't have to be in the same file. It's just convenient.
And that file has 1000 people in the contact list. That's relatively
large.
When you say "Identity discovery" what are you referring to?
--
Julian Bond E&MSN: julian_bond at voidstar.com M: +44 (0)77 5907 2173
Webmaster: http://www.ecademy.com/ T: +44 (0)192 0412 433
Personal WebLog: http://www.voidstar.com/ skype:julian.bond?chat
Always Be A Responsible Rider
Darren Bounds
machine/application/consumer to discover basic identity information
(see below) about the owner of a URI (e.g. an OpenID).
Firstname
Lastname
Street address
City
State/Province
Country
Email
OpenID
Profile image
Home/Mobile/Office phone numbers
other relatively common, universal identity artifacts
Personally, while VCard RDF exists, I'm not so sure it's as relevant
as it once was. I feel similarly about RDF and I'm not sure it should
be the foundation of everything we're trying to build.
--
Thank you,
Darren Bounds
Julian Bond
>
>"Identity Discovery" being a service to allow a
>machine/application/consumer to discover basic identity information
>(see below) about the owner of a URI (e.g. an OpenID).
>
>Firstname
>Lastname
>Street address
>City
>State/Province
>Country
>OpenID
>Profile image
>Home/Mobile/Office phone numbers
>other relatively common, universal identity artifacts
>
>Personally, while VCard RDF exists, I'm not so sure it's as relevant
>as it once was. I feel similarly about RDF and I'm not sure it should
>be the foundation of everything we're trying to build.
Well it comes down to the same thing. The main use case of copying
profile data and using contact list data means that fundamentally we
need two things.
- Discovery of where those files are
- Transport of the data.
So if the main candidates are hCard/XFN in HTML and VCard/FOAF in
RDF/XML we then need 2 discovery standards for each. And there are two
main candidates for that
- <link> in the <head> section of the html
- XRDS-Simple with two Service entries.
So far so good. But now we come down to the real problem. We really need
two discovery standards.
- Here's the URL of my profile data (in format X)
- Here's the URL of my contact list data (in format y)
FOAF has one discovery standard "here's my FOAF". hCard and XFN don't
really have a discovery standard at all, at all.
Everything else is well understood and defined. Apart from actually
implementing some of this stuff, the discovery is the last missing
piece, just as it was back in Jan.
Bob Ngu
> as it once was. I feel similarly about RDF and I'm not sure it should
> be the foundation of everything we're trying to build.
>
what's more relevant?
Chris Messina
http://microformats.org/wiki/uid
We did solve this problem quite a while ago.
Chris
On May 17, 7:51 pm, "Christian Scholz / Tao Takashi (SL)"
> Hi!
>
> I am wondering what happens if we put a link to a profile page with
> hCard format into the service catalogue and this profile page has more
> than one hCards.
> Like a page where on top you have your profile information but in
> another section a list of your friends, each with hCard markup.
>
> I guess this will then be a problem to choose which profile
> information to use or am I missing something?
> (of course I can ask the user but that rather inconvenient).
>
> -- Christian
>
> --
> Christian Scholz
> Tao Takashi (Second Life name)
Julian Bond
On a brief reading of the pages around that I'm properly confused. :(
The need here for data portability during signup (or later) is to get a
single URL from the user and then to use that to discover the location
of:-
1) The URL of the page containing their profile description
2) The URL of the page containing their list of contacts
And then to extract the data from those pages unambiguously. And for 2)
to be able to see identifiers that can be matched with identifiers
already held locally. Preferably without having to spider outwards to
the full profile page of each contact.
I still don't feel I understand the pattern for structuring all this
using just microformats. I can see bits and pieces in the microformat
approach that all help but I can't see an overall pattern that can be
recommended as a Best Practice.
--
Julian Bond E&MSN: julian_bond at voidstar.com M: +44 (0)77 5907 2173
Webmaster: http://www.ecademy.com/ T: +44 (0)192 0412 433
Personal WebLog: http://www.voidstar.com/ skype:julian.bond?chat
If You Need Help, Press F1
Bob Ngu
http://microformats.org/wiki/representative-hcard-parsing
On Jun 21, 12:48 am, Julian Bond <julian_b...@voidstar.com> wrote:
> Chris Messina <chris.mess...@gmail.com> Fri, 20 Jun 2008 10:41:57
Darren Bounds
philosophy of a microformat leads me to shy away from them for this
particular use case.
--
Thank you,
Darren Bounds