DataPortability throws down the Gauntlet

11 views
Skip to first unread message

Chris Saad

unread,
Dec 19, 2007, 12:18:38 AM12/19/07
to Social Network Portability
Check out the GraphSync page to see the first challenge from the
DataPortability.org workgroup ;)

http://www.dataportability.org/graphsync/

Here's a snippet:

Your challenge, should you choose to accept it...

1. Pick a silo of proprietary social graph data
2. Write some open source code to extract the data
3. Place that data into the open formats listed below.
4. Link to the code repository on the DataPortability Wiki.
5. Win the love and admiration of a grateful community

http://www.dataportability.org/graphsync/

Ben Werdmuller

unread,
Dec 19, 2007, 4:58:03 AM12/19/07
to social-networ...@googlegroups.com
On Dec 19, 2007 5:18 AM, Chris Saad <chris...@gmail.com> wrote:

> Your challenge, should you choose to accept it...
>
> 1. Pick a silo of proprietary social graph data

Wait, you're basically saying, "harvest the data hosted on a site you
don't own; publish it"? Isn't this likely to attract the attention of
lawyers and such, not to mention the potential ire of the users whose
data it rightfully is? And shouldn't there be some user choice in all
of this?

--
Ben Werdmuller
Company http://curverider.co.uk/ | Personal http://ben.elgg.com/

Benjamin Nowack

unread,
Dec 19, 2007, 7:09:33 AM12/19/07
to social-networ...@googlegroups.com, Chris Saad

As I understand the challenge, it's more a call for exporters, not
for actually exported data. Not sure why dataportability.org doesn't
mention RDF, though. It's a nice model that allows you to easily
merge and consolidate extracted Web information. Hmm.., ok, the
focus so far seems to be on "formats", not the underlying data
model. Would be nice to extend the DP homepage with a "model" column.
Chris, is that in scope of the effort already?

Benji

--
Benjamin Nowack
http://bnode.org/

Craig Wood

unread,
Dec 19, 2007, 9:18:48 AM12/19/07
to social-networ...@googlegroups.com
On Dec 19, 2007 4:58 AM, Ben Werdmuller <ben...@gmail.com> wrote:

On Dec 19, 2007 5:18 AM, Chris Saad <chris...@gmail.com> wrote:

>  Your challenge, should you choose to accept it...
>
>    1. Pick a silo of proprietary social graph data

Wait, you're basically saying, "harvest the data hosted on a site you
don't own; publish it"? Isn't this likely to attract the attention of
lawyers and such, not to mention the potential ire of the users whose
data it rightfully is? And shouldn't there be some user choice in all
of this?

This is exactly the point. Break down the barriers old institutional
ideals of 'ownership'.  Digital Rights Rebels have begun their 'Long March'
that will ultimately topple over all current notions of intellectual
property and privacy.  Unfortunately the current mass media is
no different than our current political regime - they are unwilling
to change for fear of losing some power, therefore ultimately they
will have to lose it all.

-Craig


--
Craig Wood
====================
Trend Spotter
http://www.wikirage.com
------------------------------------
Avid Search Enthusiast
http://www.craigsblog.com
------------------------------------
516-670-7889
YIM: w3ace

kevin curry

unread,
Dec 19, 2007, 9:27:41 AM12/19/07
to social-networ...@googlegroups.com
If the developer runs it on his own user id, or on someone who has granted permission, then no.  But this ain't the gauntlet.  I need this for all of my silos , not just one.  And there are many working on this already, no?  That's point of this group.  To figure out things like "what happens when two or more silos don't agree on the resource description for a person?"  Such are the ups and downs of ETL 2.0

On Dec 19, 2007 4:58 AM, Ben Werdmuller <ben...@gmail.com> wrote:

Dale Newfield

unread,
Dec 19, 2007, 12:40:50 PM12/19/07
to social-networ...@googlegroups.com
Craig Wood wrote:
> This is exactly the point. Break down the barriers old institutional
> ideals of 'ownership'.

You must have forgotten the details of the "terms of use" agreements
you've accepted to create those accounts. I suggest you go read them again.

-Dale

Craig Wood

unread,
Dec 19, 2007, 1:02:12 PM12/19/07
to social-networ...@googlegroups.com
I'm making reference to the Chinese Cultural Revolution and you think
I care about some 'contract' that I you say I agreed to because I
clicked a button on a website ?  I like the oligarchy, I've done well
in with the current greed is good mentality, but I've seen the way the
next generation uses bit torrents instead of Cable TV and my prediction
is soon there will be upheaval.

Dale Newfield

unread,
Dec 19, 2007, 1:16:32 PM12/19/07
to social-networ...@googlegroups.com
Craig Wood wrote:
> I'm making reference to the Chinese Cultural Revolution and you think
> I care about some 'contract' that I you say I agreed to because I
> clicked a button on a website ?

As long as I don't have to stand next to you in court, you're welcome to
do whatever you'd like.

-Dale

Michael and Diana Finney

unread,
Dec 19, 2007, 9:07:44 PM12/19/07
to social-networ...@googlegroups.com
Prove to me that taking "... proprietary ... data ..." is legal. ;)

I wish to understand.

Thanks.
Michael

--
Michael Finney - "Always Striving To Serve You Better Every Day"
fin...@ACM.org
http://www.SmilingSoftwareSolutions.com

http://www.dataportability.org/graphsync/

Here's a snippet:

http://www.dataportability.org/graphsync/
No virus found in this incoming message.
Checked by AVG Free Edition.
Version: 7.5.503 / Virus Database: 269.17.4/1189 - Release Date: 12/18/2007
9:40 PM

No virus found in this outgoing message.
Checked by AVG Free Edition.
Version: 7.5.503 / Virus Database: 269.17.4/1189 - Release Date: 12/18/2007
9:40 PM

Nathaniel Ford

unread,
Dec 19, 2007, 11:56:31 PM12/19/07
to social-networ...@googlegroups.com
I believe they may mean it in the sense that the definition provided in
the jargon file does:

http://www.catb.org/~esr/jargon/html/P/proprietary.html

2. In the language of hackers and users, inferior; implies a product not
conforming to open-systems standards, and thus one that puts the
customer at the mercy of a vendor able to gouge freely on service and
upgrade charges after the initial sale has locked the customer in.

This does not necessarily preclude it from being legal to take the data.
That would, presumably, be dependent on the silo you're writing for.

Finally, it is worth noting that though they ask you to place the data
into an open format they do not actually ask you to distribute that
data. Presumably, if the data is available to you, they don't care what
you do with it so long as you're not distributing it. It does, however,
mean that you're writing software with which this could be done. On the
one hand, this may cause some silos to decide "Hey, this isn't going to
cost a lot to do.", and on the other hand, may make other silos worried
that their data will be stolen.

So, in truth, they're not asking for anything illegal. They're asking
you to reduce the energy it would take to sync these graphs by writing
the software that would do it. That said, it may cross a line for you;
in which case I don't see why you'd 'take up the gauntlet'.

Anyway, looking forward to more traffic that is less to do with the
legal issues (as fascinating as they are...) and more to do with the
technology...

-Nathaniel

victorc

unread,
Dec 20, 2007, 4:47:47 AM12/20/07
to Social Network Portability
Hi,

Following the DataPortability challenge, is my project OSocial
http://www.osocial.net/network can be a challenger ?

And so how to post something in the Wiki ? ;)

Victor

Danny Ayers

unread,
Dec 20, 2007, 4:57:29 AM12/20/07
to social-networ...@googlegroups.com, paolo...@gmail.com, Story Henry, bno...@semsol.com

With the provisos mentioned already (even if the silo maintainers
aren't the true owners of the data inside, they'll still get miffed if
we grab it in bulk), this sounds a grand challenge. But is the
challenge in itself enough motivation to get the work done?

A very similar challenge has been around the semweb community for a
good few years now, scoped at making arbitrary data interoperable
through exposing/extracting it in the RDF model. The historic problem
here has been classic chicken and egg - for such data to be useful you
also need applications. (This problem is finally evaporating now
there's a significant amount of RDF available, and a handful of usable
apps).

Ok, in the social graph context we already have one class of
application in place - the social network site. But to motivate the
portability side, we need something that fits into another space.

Paolo Valdemarin recently proposed an application 'Ring' [1] which
would act as a desktop contact book, but with knowledge of one's
personal subgraphs and the ability to interface with the various
networking (and other) sites.

The nearest I know of to an implementation of this for the desktop is
Henry Story's Beatnik [2]. Benji's Knowee [3] is a similar Web-based
project. I get the impression Henry's been more focussed on putting
together something that would be immediately useful within the
enterprise (he's been RDFizing Sun Microsystems for a while now ;-)
and Benji's closer to the social graph in the wild.

One thing I'd say they both get right is using the RDF model under the
hood. The specific formats in which the data is gathered and exposed
becomes a comparatively trivial issue.

For example, go to LinkedIn, export your data as vCard (.vcf file) go
to FOAFGen [4] and upload the data as "local vCard". Voila, portable
FOAF. Using standard kit this can, for example, be exposed as hCards
(see [5]).

Architecturally, there are at least two big advantages in taking this
approach. When we have an environment which contains N different
formats/APIs which can express the social information, either we need
N^2 different converters to enable interop *or* we can map to a common
model, and just need N different parsers/serialisers - still
non-trivial, but a lot easier. The second advantage, which is the
reason an RDF-backed is appropriate, is having the flexibility to
compatibly include whatever (other) information you like in the common
model, beyond the limitations of the typical individual formats/APIs.
Check Tim Berners-Lee's FOAF [6].

Anyhow, I'll cc Paolo, Henry and Benji - with a tiny bit of convention
on the protocol side (HTTP...er, anything else?), Knowee and Beatnik
could (maybe can already) transparently interoperate, and work in
concert as an implementation of Ring...

Cheers,
Danny.

[1] http://paolo.evectors.it/2007/07/02.html
[2] http://blogs.sun.com/bblfish/entry/beatnik_change_your_mind
[3] http://knowee.org/
[4] http://toxi.co.uk/foafgen/
[5] http://dannyayers.com/2007/11/04/foaf-sample-data-and
[6] http://www.w3.org/People/Berners-Lee/card.n3

--

http://dannyayers.com

Danny Ayers

unread,
Dec 20, 2007, 5:02:18 AM12/20/07
to social-networ...@googlegroups.com
On 20/12/2007, victorc <victor.co...@gmail.com> wrote:
>
> Hi,
>
> Following the DataPortability challenge, is my project OSocial
> http://www.osocial.net/network can be a challenger ?

Very nice.

Quick question - is there any way I can pass the system my own copy of
my personal social, without going through another OpenSocial
container? (FOAF? hCard/XFN?)

Cheers,
Danny.

--

http://dannyayers.com

Henry Story

unread,
Dec 20, 2007, 5:27:16 AM12/20/07
to social-networ...@googlegroups.com
Thanks Danny,

This is what Beatnik looks like after I did the 3 following actions:

1. dragged and dropped the foaf file from my home page http://bblfish.net/
onto the first column
Beatnik fetched my foaf file, and adds my name to the first column
and the list of all the people that I know in the second column.
2. clicked the right arrow which moves the cursor to the list of the
people I know,
and clicked the down arrow until I reached Tim Berners Lee's info
located on the w3c site
3. clicked space. This downloads Tim Berner's Lees foaf file and
displays the people he foaf:knows in the right column
4. clicked the right arrow, then the down arrow until I reach Kingsley
Idehen.
5. clicked the space bar. Beatnik then fetches Kingsley's foaf file,
which is on yet another server somewhere else
... so it is really easy to explore a distributed social network

Picture 1.png

Michael Wechner

unread,
Dec 20, 2007, 6:12:52 AM12/20/07
to social-networ...@googlegroups.com
Henry Story wrote:

>
> you see someone you know in some column you can just drag and drop
> them into the column of people you know. That's how easy it is to add
> a friend to one's social network.
>


very cool :-)

>
> Ok the right hand side of the user interface really sucks currently.
> I put it together really quickly to show some of the info available
> in the foaf file.
>
> Things that I am doing now:
>
> - I am in the process of adding inferencing support. When that works
> we will see a lot more information come together.
>
>
> Things to do:
> - add a publish button to publish foaf files using scp, ftp, atom
> publication protocol, or web dav
>

We are currently working on an Open Social Server and will have the
prototype finished very soon now. I thought it would be great to use
Beatnik as a client, whereas I think this should work just as email does
;-) Within Beatnik one should be able to specify the open Social server
domain, e.g. http://foaf.foo.bar and that's it.

Everything else would be done through a FOAF protocol resp.
introspection via HTTP e.g. the communication between Beatnik and the
Server could be as follows:

- Beatnik: GET http://foaf.foo.bar

- Server Response: 200 <html><head><link rel="foaf-introspection"
type="application/foaf-server+xml" href="foaf-introspection.xml"/>...

- Beatnik: GET http://foaf.foo.bar/foaf-introspection.xml

- Server Response: 200 <?xml version="1.0"?> <introspection><add-user
href="..." method="POST"/><search-person href=""/>

- Beatnik: POST http://foaf.foo.bar/add-user

- .... (authentication ...)


This way end-users do not have to worry about what is running on the
server side and I think this would really free them.

> - improove the right hand user interface (it used to be a lot better
> when it was written in JavaFX )
> + make it more beautiful
> + make it editable
> - add search tool
>

I think this should also be done through a protocol resp. accessing an
Open Social Provider Server, which would then again access a disovery
server ....

>
>
>
> Any help would be welcome. This is an open source project at
> https://sommer.dev.java.net/ . Just
> join the users mailing list, or contact me. I'd love to add some
> comitters to the project. Swing wizards, html
> artists, documenters, testers, deployers are all welcome. :-) There
> is more than I can do by myself here. And this is a great way to get
> your hands on a real semantic web application and learn the ropes,
> or invent them, as there is a lot to discover here.
>

very much agreed. Btw, you might want to read some notes we made how we
think the archtitecture and communication should look like between
client and server, etc.

http://yanel.wyona.org/specification/foaf/index.html

resp.

http://yanel.wyona.org/en/documentation/deployment/deployment_within-foaf-environment.html

Cheers

Michael

>
>
> Henry


>
>
> On 20 Dec 2007, at 07:57, Danny Ayers wrote:
>
>> The nearest I know of to an implementation of this for the desktop is
>> Henry Story's Beatnik [2]. Benji's Knowee [3] is a similar Web-based
>> project. I get the impression Henry's been more focussed on putting
>> together something that would be immediately useful within the
>> enterprise (he's been RDFizing Sun Microsystems for a while now ;-)
>> and Benji's closer to the social graph in the wild.
>
>


--
Michael Wechner
Wyona - Open Source Content Management - Yanel, Yulup
http://www.wyona.com
michael...@wyona.com, mi...@apache.org
+41 44 272 91 61

Henry Story

unread,
Dec 20, 2007, 6:59:41 AM12/20/07
to social-networ...@googlegroups.com, us...@sommer.dev.java.net
(cc ing the beatnik mailing list)

On 20 Dec 2007, at 09:12, Michael Wechner wrote:
> Henry Story wrote:
>> [snip]


>> Things to do:
>> - add a publish button to publish foaf files using scp, ftp, atom
>> publication protocol, or web dav
>
> We are currently working on an Open Social Server and will have the
> prototype finished very soon now. I thought it would be great to use
> Beatnik as a client, whereas I think this should work just as email
> does ;-) Within Beatnik one should be able to specify the open
> Social server domain, e.g. http://foaf.foo.bar and that's it.

Yes, definitively. I was looking for a server I could have Beatnik
publish foaf to. (other than simple Apache servers)

> Everything else would be done through a FOAF protocol resp.
> introspection via HTTP e.g. the communication between Beatnik and
> the Server could be as follows:
>
> - Beatnik: GET http://foaf.foo.bar
>
> - Server Response: 200 <html><head><link rel="foaf-introspection"
> type="application/foaf-server+xml" href="foaf-introspection.xml"/>...
>
> - Beatnik: GET http://foaf.foo.bar/foaf-introspection.xml
>
> - Server Response: 200 <?xml version="1.0"?> <introspection><add-
> user href="..." method="POST"/><search-person href=""/>
>
> - Beatnik: POST http://foaf.foo.bar/add-user
>
> - .... (authentication ...)
>
>
> This way end-users do not have to worry about what is running on the
> server side and I think this would really free them.

I had been thinking of using the Atom Publication Protocol as one of
the ways to publish foaf. Have you thought of using that? What you
have written above looks quite similar. (I have not looked at Atom for
a few months now) Do you need it to look any different from the atom
publication format? I think Atom is designed to accept any type of
content, including rdf/xml. So it would be a way to be standard
compliant and not have to re-invent anything.

Anyway I think supporting something like that would be excellent for
Beatnik. If you have a server that accepts atom, then whoever will
implement it in Beatnik (me presumably when I get round to it) will
have something to test it with.

>> - improove the right hand user interface (it used to be a lot
>> better when it was written in JavaFX )
>> + make it more beautiful
>> + make it editable
>> - add search tool
>>
>
> I think this should also be done through a protocol resp. accessing
> an Open Social Provider Server, which would then again access a
> disovery server ....

Yes. A SPARQL endpoint on a server would be the most powerful way of
doing that. But it should be easy to map other existing mechanisms to
SPARQL queries.
( a fun post I wrote on this subject: http://blogs.sun.com/bblfish/entry/sparqling_altavista_the_meaning_of
)

>> Any help would be welcome. This is an open source project at https://sommer.dev.java.net/
>> . Just
>> join the users mailing list, or contact me. I'd love to add some
>> comitters to the project. Swing wizards, html
>> artists, documenters, testers, deployers are all welcome. :-) There
>> is more than I can do by myself here. And this is a great way to
>> get your hands on a real semantic web application and learn the
>> ropes, or invent them, as there is a lot to discover here.
>>
>
> very much agreed. Btw, you might want to read some notes we made how
> we think the archtitecture and communication should look like
> between client and server, etc.
>
> http://yanel.wyona.org/specification/foaf/index.html
>
> resp.
>
> http://yanel.wyona.org/en/documentation/deployment/deployment_within-foaf-environment.html

yes, you should add "(ftp), scp, atom, webdav" between the client and
foo bar.
Let's keep in contact on this!

Henry

Michael Wechner

unread,
Dec 20, 2007, 7:40:25 AM12/20/07
to social-networ...@googlegroups.com, us...@sommer.dev.java.net
Henry Story wrote:

>>
>
>
> I had been thinking of using the Atom Publication Protocol as one of
> the ways to publish foaf. Have you thought of using that? What you
> have written above looks quite similar. (I have not looked at Atom
> for a few months now) Do you need it to look any different from the
> atom publication format? I think Atom is designed to accept any type
> of content, including rdf/xml. So it would be a way to be standard
> compliant and not have to re-invent anything.


very much agreed

>
> Anyway I think supporting something like that would be excellent for
> Beatnik. If you have a server that accepts atom, then whoever will
> implement it in Beatnik (me presumably when I get round to it) will
> have something to test it with.


I will download and install Beatnik now and will start digging ;-)

>
> Yes. A SPARQL endpoint on a server would be the most powerful way of
> doing that. But it should be easy to map other existing mechanisms
> to SPARQL queries.
> ( a fun post I wrote on this subject:
> http://blogs.sun.com/bblfish/entry/sparqling_altavista_the_meaning_of )


ok, will check it out

>
> Let's keep in contact on this!

I have subscribed to the beatnik (users and dev mailing lists). Maybe we
can try getting a prototype running soon and then go from there ;-)

Cheers

Michael

Evaldas Taroza

unread,
Dec 20, 2007, 9:42:34 AM12/20/07
to Social Network Portability
Hi Henry,

> Things that I am doing now:
>
> - I am in the process of adding inferencing support. When that works
> we will see a lot more information come together.
What kind of inferencing are you adding? By the way, can anyone point
out something about FOAF from the logic perspective: how expensive it
is to reason about FOAFs for a machine, what can you derive, etc.
(just to get some motivation to look at all the RDFs :D)

I saw many people are suggesting to keep it simple and leave out FOAF,
but when I think that FOAF would enable reasoning about the social
network, I really think it is worth suffering a bit with all the RDFs
to get such exciting intelligent applications.

Evaldas

Henry Story

unread,
Dec 20, 2007, 10:29:58 AM12/20/07
to social-networ...@googlegroups.com, us...@sommer.dev.java.net

On 20 Dec 2007, at 12:42, Evaldas Taroza wrote:
> Hi Henry,
>
>> Things that I am doing now:
>>
>> - I am in the process of adding inferencing support. When that works
>> we will see a lot more information come together.
> What kind of inferencing are you adding?


Well at present Beatnik works by default without inferencing. And you
can get quite a lot done that way. The problem is when someone
describes themselves indirectly via an inverse functional property
such as foaf:mbox. I have diagrams illustrating this here:

http://blogs.sun.com/bblfish/entry/beatnik_change_your_mind

In N3 you get somthing like this

@prefix foaf: <http://xmlns.com/foaf/0.1/> .

:me foaf:knows [ a foaf:Person;
foaf:mbox <mailto:j...@smith.com>;
foaf:name "Joe Smith" ;
rdfs:seeAlso <http://smith.com/joe.rdf> ].


Then in joe's rdf file we have the following

<> a foaf:PersonalProfileDocument;
foaf:primaryTopic [ foaf:name "Joe Smith";
foaf:mbox <mailto:j...@smith.com>;
foaf:knows <http://bblfish.net/people/henry/card#me
] .

Ok so in neither of these files is Joe identified by a URI, but just
by what is called a blank node . Those two blank nodes are indirectly
identified via the foaf:mbox relation. That relation is inverse
functional (you can find that out by looking at the foaf ontology at http://xmlns.com/foaf/0.1/
, and so a reasoning engine will be able to merge the two nodes and so
know that I know Joe Smith and that he knows me. That is very simple
reasoning, but it is very important to help bring data together.

It should be very easy to program. I have just implemented reasoning
using the still alpha level Networked Graphs sail for Sesame http://isweb.uni-koblenz.de/Research/NetworkedGraphs
(can't connect to it from here for some reason )

I can then add the following rules to the graph:

#foaf:box is inverse funtional
grph: ng:definedBy """
CONSTRUCT { ?a <http://www.w3.org/2002/07/owl#sameAs> ?b . }
WHERE {
?a <http://xmlns.com/foaf/0.1/mbox> ?mbx .
?b <http://xmlns.com/foaf/0.1/mbox> ?mbx .
FILTER ( ! SAMETERM (?a ,?b) )
} """^^ng:Query .

# indiscernability of identicals
grph: ng:definedBy """
CONSTRUCT { ?b ?rel ?c . }
WHERE { ?a <http://www.w3.org/2002/07/owl#sameAs> ?b .
?a ?rel ?c .
FILTER ( ! SAMETERM(?rel , <http://www.w3.org/2002/07/
owl#sameAs>) )
} """^^ng:Query .


And it will calculate in a backtracking way when two things are
identical. I like the idea of using SPARQL for the rules. It is very
powerful, and should allow me to express some very interesting
things... We'll see how it works out.

Currently this does not work well because my java objects do no
caching, and so any call to the fields of the method sets off the
reasoning engine. So I have to implement caching.

But there are many other tools out there to help you do reasoning.
Some of them can reason of billions of relations. And exactly what
tool to use is going to be very dependant on what you are using, as
usual. You could also do this type of reasoning by hand, btw. But it
is a lot more tedious to write.


> By the way, can anyone point
> out something about FOAF from the logic perspective: how expensive it
> is to reason about FOAFs for a machine, what can you derive, etc.
> (just to get some motivation to look at all the RDFs :D)

The Description Logics people will be able to say some interesting
things there. But remember you don't have to work out all the
consequences of something for information to be useful. Just as in
human life some people don't reason very well, but we can still
communicate with them. It is often nicer to be able to communicate
with people who do reason though.

Danny Ayers

unread,
Dec 20, 2007, 12:09:36 PM12/20/07
to social-networ...@googlegroups.com
On 20/12/2007, Evaldas Taroza <tar...@gmail.com> wrote:

> I saw many people are suggesting to keep it simple and leave out FOAF,

Like it or not, social networks are *not simple* (some of my best
friends are cats - yeah I should get a life :-)

Jack Park

unread,
Dec 20, 2007, 12:14:36 PM12/20/07
to social-networ...@googlegroups.com
It is said that some people's best friends are little pills...

André Luís

unread,
Dec 20, 2007, 1:12:33 PM12/20/07
to social-networ...@googlegroups.com
Ok guys, I have question relating to FOAF...

Why has it been out there for such a long time (i remember it from my
early days in livejournal, over 3 years ago) and from where I'm
standing, it hasn't matured into the defacto standard to describe
relationships. XFN seems to be more widely accepted. Am I wrong about
this? Can you point me towards some literature that claims FOAF as
widely accepted?

Thanks, I'm honestly curious, not throwing a jab at foaf. ;)

Cheers,
André Luís

Julian Bond

unread,
Dec 20, 2007, 1:48:37 PM12/20/07
to social-networ...@googlegroups.com
Danny Ayers <danny...@gmail.com> Thu, 20 Dec 2007 10:57:29

>Check Tim Berners-Lee's FOAF [6].

>[6] http://www.w3.org/People/Berners-Lee/card.n3

Just mildly amused that when attempting to open this in a browser
(firefox), it didn't know what to do with it, and offered to save it or
for me to browse for a suitable application.

--
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
Medicated Treatment Formula

Danny Ayers

unread,
Dec 20, 2007, 1:54:25 PM12/20/07
to social-networ...@googlegroups.com
On 20/12/2007, Julian Bond <julia...@voidstar.com> wrote:
>
> Danny Ayers <danny...@gmail.com> Thu, 20 Dec 2007 10:57:29
> >Check Tim Berners-Lee's FOAF [6].
>
> >[6] http://www.w3.org/People/Berners-Lee/card.n3
>
> Just mildly amused that when attempting to open this in a browser
> (firefox), it didn't know what to do with it, and offered to save it or
> for me to browse for a suitable application.

The Tabulator plugin helps:

http://www.w3.org/2005/ajar/tab

Danny Ayers

unread,
Dec 20, 2007, 2:20:52 PM12/20/07
to social-networ...@googlegroups.com
On 20/12/2007, André Luís <andrel...@gmail.com> wrote:
>
> Ok guys, I have question relating to FOAF...
>
> Why has it been out there for such a long time (i remember it from my
> early days in livejournal, over 3 years ago) and from where I'm
> standing, it hasn't matured into the defacto standard to describe
> relationships. XFN seems to be more widely accepted. Am I wrong about
> this? Can you point me towards some literature that claims FOAF as
> widely accepted?
>
> Thanks, I'm honestly curious, not throwing a jab at foaf. ;)

I don't know of anything that'd provide useful figures, but the
majority of Semantic Web-oriented tools/applications use FOAF to
represent people, and there are a lot of those, e.g. see
http://blogs.talis.com/nodalities/this_weeks_semantic_web/
I believe quite a few blogging tools provide XFN by default
(WordPress?) so there's a lot of that about.

But the comparison with XFN is a little bit apples vs. oranges. XFN is
great for publishing relationship descriptions, FOAF is great for
doing stuff with those descriptions.

XFN provides a domain-specific syntax (for embedded data in the HTML
format), FOAF provides a domain-specific vocabulary (for describing a
particular kind of thing in the RDF model).

XFN can be interpreted as FOAF, e.g. see
http://www.w3.org/TR/grddl-primer/#hotel

(Some of) FOAF can be expressed as XFN - I've not seen any public code
specifically for XFN but it'd be very straightforward, I did some
hCards from FOAF not long ago:
http://dannyayers.com/2007/11/04/foaf-sample-data-and

André Luís

unread,
Dec 20, 2007, 3:31:54 PM12/20/07
to social-networ...@googlegroups.com
Thanks Danny. That cleared up some doubts I had. I'll look into it
more carefully when I get the chance.

--
André

Chris Saad

unread,
Dec 20, 2007, 6:30:13 PM12/20/07
to Social Network Portability
Hey guys - this is getting a little complicated again, boil the ocean
sorta stuff...

Check out my post here: http://groups.google.com/group/graphsync/browse_thread/thread/a66828fe24b591b8

Chris

On Dec 21, 5:20 am, "Danny Ayers" <danny.ay...@gmail.com> wrote:
> On 20/12/2007, André Luís <andreluis...@gmail.com> wrote:
>
>
>
> > Ok guys, I have question relating to FOAF...
>
> > Why has it been out there for such a long time (i remember it from my
> > early days in livejournal, over 3 years ago) and from where I'm
> > standing, it hasn't matured into the defacto standard to describe
> > relationships. XFN seems to be more widely accepted. Am I wrong about
> > this? Can you point me towards some literature that claims FOAF as
> > widely accepted?
>
> > Thanks, I'm honestly curious, not throwing a jab at foaf. ;)
>
> I don't know of anything that'd provide useful figures, but the
> majority of Semantic Web-oriented tools/applications use FOAF to
> represent people, and there are a lot of those, e.g. seehttp://blogs.talis.com/nodalities/this_weeks_semantic_web/
> I believe quite a few blogging tools provide XFN by default
> (WordPress?) so there's a lot of that about.
>
> But the comparison with XFN is a little bit apples vs. oranges. XFN is
> great for publishing relationship descriptions, FOAF is great for
> doing stuff with those descriptions.
>
> XFN provides a domain-specific syntax (for embedded data in the HTML
> format), FOAF provides a domain-specific vocabulary (for describing a
> particular kind of thing in the RDF model).
>
> XFN can be interpreted as FOAF, e.g. seehttp://www.w3.org/TR/grddl-primer/#hotel

Chris Saad

unread,
Dec 20, 2007, 7:18:13 PM12/20/07
to Social Network Portability
Hey guys - this is getting a little complicated again, boil the ocean
sorta stuff...

Check out my post here: http://groups.google.com/group/graphsync/browse_thread/thread/a66828fe24b591b8

Chris


On Dec 21, 5:20 am, "Danny Ayers" <danny.ay...@gmail.com> wrote:
> On 20/12/2007, André Luís <andreluis...@gmail.com> wrote:
>
>
>
> > Ok guys, I have question relating to FOAF...
>
> > Why has it been out there for such a long time (i remember it from my
> > early days in livejournal, over 3 years ago) and from where I'm
> > standing, it hasn't matured into the defacto standard to describe
> > relationships. XFN seems to be more widely accepted. Am I wrong about
> > this? Can you point me towards some literature that claims FOAF as
> > widely accepted?
>
> > Thanks, I'm honestly curious, not throwing a jab at foaf. ;)
>
> I don't know of anything that'd provide useful figures, but the
> majority of Semantic Web-oriented tools/applications use FOAF to
> represent people, and there are a lot of those, e.g. seehttp://blogs.talis.com/nodalities/this_weeks_semantic_web/
> I believe quite a few blogging tools provide XFN by default
> (WordPress?) so there's a lot of that about.
>
> But the comparison with XFN is a little bit apples vs. oranges. XFN is
> great for publishing relationship descriptions, FOAF is great for
> doing stuff with those descriptions.
>
> XFN provides a domain-specific syntax (for embedded data in the HTML
> format), FOAF provides a domain-specific vocabulary (for describing a
> particular kind of thing in the RDF model).
>
> XFN can be interpreted as FOAF, e.g. seehttp://www.w3.org/TR/grddl-primer/#hotel

Tim Berners-Lee

unread,
Dec 20, 2007, 8:15:50 PM12/20/07
to social-networ...@googlegroups.com
On 2007-12 -19, at 09:18, Craig Wood wrote:



On Dec 19, 2007 4:58 AM, Ben Werdmuller <ben...@gmail.com> wrote:

On Dec 19, 2007 5:18 AM, Chris Saad <chris...@gmail.com> wrote:

>  Your challenge, should you choose to accept it...
>
>    1. Pick a silo of proprietary social graph data

Wait, you're basically saying, "harvest the data hosted on a site you
don't own; publish it"? Isn't this likely to attract the attention of
lawyers and such, not to mention the potential ire of the users whose
data it rightfully is? And shouldn't there be some user choice in all
of this?

This is exactly the point. Break down the barriers old institutional
ideals of 'ownership'.  Digital Rights Rebels have begun their 'Long March'
that will ultimately topple over all current notions of intellectual
property and privacy.  Unfortunately the current mass media is
no different than our current political regime - they are unwilling
to change for fear of losing some power, therefore ultimately they
will have to lose it all.


Hmmm .. it was Chris' point, not yours -- are you sure that's what he meant?

This list is about interoperability, I thought. I have leaned that it is unwise to mix messages.
If you also think all friend data should be public (I don't, only professional network data)
then that is a different argument.
I had to be careful both with WWW and now with the sematic web that the drive is
that you shouldn't be frustrated by lack of interop. If people want to keep data to
temselves, or within a firewall, that is often omportant.
There is some data, like goverment data, which should be public, and that message
of making data publci is a separate message.

Eg. Linked Data = interop, RDF, etc,
compare with Linked Open Data = Public  linked data drive, dbpedia, etc.

They are obviously connected -- you can't have LOD without LD standards.
I think "Linked Public Data" might have been a better name.

Tim

Tim Berners-Lee

unread,
Dec 20, 2007, 8:27:36 PM12/20/07
to social-networ...@googlegroups.com
Brilliant.

I hope you get the save back working soon.
Actually I think it should save back your FOAF file whenever it makes
a change, without asking, with webdav.

Can you make it so that if you don't have a FOAF file, it creates one,
asks you where you have write access?

Tim

> <Picture 1.png>
> If you see someone you know in some column you can just drag and

> drop them into the column of people you know. That's how easy it is
> to add a friend to one's social network.
>

> Ok the right hand side of the user interface really sucks currently.
> I put it together really quickly to show some of the info available
> in the foaf file.
>

> Things that I am doing now:
>
> - I am in the process of adding inferencing support. When that works
> we will see a lot more information come together.
>
>

> Things to do:
> - add a publish button to publish foaf files using scp, ftp, atom
> publication protocol, or web dav

> - improove the right hand user interface (it used to be a lot better
> when it was written in JavaFX )
> + make it more beautiful
> + make it editable
> - add search tool

> - build a cache
> - build a cache viewer so that it is easy to see where all the
> information came from
>
> Interesting things to do later
> - edit information on other people
> - add a way to evaluate and set trust values on information found on
> the web
> - find better ways to show what information was downloaded, and what
> it affected


>
>
> Any help would be welcome. This is an open source project at https://sommer.dev.java.net/
> . Just
> join the users mailing list, or contact me. I'd love to add some
> comitters to the project. Swing wizards, html
> artists, documenters, testers, deployers are all welcome. :-) There
> is more than I can do by myself here. And this is a great way to get
> your hands on a real semantic web application and learn the ropes,
> or invent them, as there is a lot to discover here.
>
>

> Henry
>
>
> On 20 Dec 2007, at 07:57, Danny Ayers wrote:

Chris Saad

unread,
Dec 20, 2007, 9:52:56 PM12/20/07
to Social Network Portability
Yes Tim, you are right - that was not quite my point. The goal of
DataPortability.org, is not to fight the man - only to allow the user
to take control of their own personal data for their own personal
needs.

The goal if the GraphSync project is not to extract all data from
Social Networks - only to provide tools for users to export their own
personal data just like they do when importing their friends

Chris

On Dec 21, 11:15 am, Tim Berners-Lee <ti...@w3.org> wrote:
> On 2007-12 -19, at 09:18, Craig Wood wrote:
>
>
>
>
>
>
>
> > On Dec 19, 2007 4:58 AM, Ben Werdmuller <benw...@gmail.com> wrote:
> > YIM: w3ace- Hide quoted text -
>
> - Show quoted text -- Hide quoted text -
>
> - Show quoted text -

Julian Bond

unread,
Dec 21, 2007, 4:21:47 AM12/21/07
to social-networ...@googlegroups.com
Danny Ayers <danny...@gmail.com> Thu, 20 Dec 2007 19:54:25

Well that killed my browser. Oh, well, Alpha software. :(

I don't mean to be rude about the FOAF and RDF community but I just
don't see how it's going to become mainstream. The most promising
efforts I can see at the moment for Social Network Portability are in
these areas:-
- Google's OpenSocial Data APIs
- OpenID Attribute exchange
- Open APIs in the style of those from websites like Twitter, Plaxo,
Spoke.
- Building real world aggregators and tools for the Social data that's
already out there.

Now FOAF, RDF, hCard and all the rest can provide a lot of help to these
to avoid re-inventing wheels and to benefit from avoiding pitfalls.
There's a whole lot of prior work that can be taken advantage of by
things like:-
- Encouraging the use of FOAF and VCard tags in OpenSocial People Data
- Helping the OpenId AX people to get the schemas right [1]
- Encouraging the use of mbox_sha1sum as a common ID across systems
- Promoting auto-discovery techniques
and so on.

[1]http://www.axschema.org

Danny Ayers

unread,
Dec 21, 2007, 6:14:22 AM12/21/07
to social-networ...@googlegroups.com
On 21/12/2007, Julian Bond <julia...@voidstar.com> wrote:

> >The Tabulator plugin helps:
> >
> >http://www.w3.org/2005/ajar/tab
>
> Well that killed my browser. Oh, well, Alpha software. :(

Yeah...get on #swig and give the man a bug report ;-)

> I don't mean to be rude about the FOAF and RDF community but I just
> don't see how it's going to become mainstream.

Sure, I know you've been around this stuff long enough to know what
you're talking about, got a lot of respect for your position.

The most promising
> efforts I can see at the moment for Social Network Portability are in
> these areas:-
> - Google's OpenSocial Data APIs
> - OpenID Attribute exchange
> - Open APIs in the style of those from websites like Twitter, Plaxo,
> Spoke.

I think you're probably right that any real progress in the near
future will include some or all of these.

> - Building real world aggregators and tools for the Social data that's
> already out there.

There's the rub. Each of the the above has it's own quasi-proprietary
view of the data. They may be open, but there's only minimal reuse of
standards and commonality in the way the information is represented.
There's no real harm as such in having lots of different, largely
incompatible APIs or even in the kind of reinvention found in
Attribute Exchange. But to work across this stuff, a *lot* of glue
will be needed.

[If you get really bored over Xmas, a while back I did an article "The
Shortest Path" arguing that things like the Web 2.0 APIs are taking
us in very much the same general direction as the Semantic Web stuff,
only via a longer route than broader adoption of RDF etc would allow -
http://www.talis.com/platform/resources/publications.shtml ]

> Now FOAF, RDF, hCard and all the rest can provide a lot of help to these
> to avoid re-inventing wheels and to benefit from avoiding pitfalls.

Absolutely.

> There's a whole lot of prior work that can be taken advantage of by
> things like:-
> - Encouraging the use of FOAF and VCard tags in OpenSocial People Data
> - Helping the OpenId AX people to get the schemas right [1]
> - Encouraging the use of mbox_sha1sum as a common ID across systems
> - Promoting auto-discovery techniques
> and so on.

I agree.

While I have every intention to continue singing the praises of
FOAF/RDF at every opportunity (*grin*), the real key is simply reusing
the approaches of Web architecture which are proven to work on a
global scale. Basically URIs as identifiers (which RDF extends from
just documents by using URIs as identifiers for entities and
relationships, i.e. typed links), and using HTTP as it was designed
(cue RESTful cliches).

IMHO the semweb community has generally only relatively recently
started taking this on board, after a long time treating RDF stores as
just a nifty kind of unstructured database and overlooking the
significance of the URI on_the_Web. While apps like e.g. Joost and
Seeismic may take advantage of the flexible modelling offered by
Semantic Web technologies internally, the real progress (still in
IMHO) has been around Linked Data
(http://en.wikipedia.org/wiki/Linked_Data).

Things like OpenID and OpenSocial (in front of the numerous social
network sites) are new, and when used in concert with WebArch offer
huge potential.

Though there was quite a big change in practice between the home pages
of Web 1.0 and the RSS-enabled blogs etc of Web 2.0, the syndication
side is still human-readable document oriented. Even mashups tend to
be client-side hacks on such data. It's still Web in the sense of
timbl's "Giant Global Graph".

But social networks are *not* about document content, they're about
people. A conceptual shift is needed to thinking in terms of Web data
(a graph of related things) rather than stretching the doc-oriented
stuff further.

Just because FOAF isn't ubiquitous doesn't mean the approach is
flawed, or even that there's a more compelling alternative (I can't
see one). BradFitz's Social Graph piece was a bit of a wake-up call
that "We're not in Kansas any more". FOAF's been sorting out the
Munchkins in this space a long time now.

Henry Story

unread,
Dec 21, 2007, 6:51:09 AM12/21/07
to social-networ...@googlegroups.com, us...@sommer.dev.java.net
On 20 Dec 2007, at 23:27, Tim Berners-Lee wrote:
> Brilliant.
>
> I hope you get the save back working soon.
> Actually I think it should save back your FOAF file whenever it
> makes a change, without asking, with webdav.

Yes. This is the most important part of Beatnik. It should be really
easy to edit and publish you foaf file, just as it should be easy to
explore the foaf relations. Is there a webdav server somewhere you are
thinking of in particular?

One really cool feature that should not be too difficult to implement,
would be to make it easy to specify with Beatnik where you happen to
be currently. This would then publish this with information in your
foaf file, so that people could find out where you are located. I am
now in Sao Paulo Brazil for example. This can help people avoid waking
other people late at night with a phone call when they are on the
other side of the world.

> Can you make it so that if you don't have a FOAF file, it creates
> one, asks you where you have write access?

Yes. That is exactly how BlogEd, the blog editor I worked on before,
worked. There was a wizard panel to step you through this setup
procedure.

So before I can get this to work I need to add the editing
functionality, so that someone can edit their info. But I agree:
getting the publication functionality in should be very high on the
agenda. I'll bump it right up there.

Henry

Michael Wechner

unread,
Dec 21, 2007, 7:36:00 AM12/21/07
to social-networ...@googlegroups.com, us...@sommer.dev.java.net
Henry Story wrote:

>
> On 20 Dec 2007, at 23:27, Tim Berners-Lee wrote:
>
>> Brilliant.
>>
>> I hope you get the save back working soon.
>> Actually I think it should save back your FOAF file whenever it
>> makes a change, without asking, with webdav.
>
>
> Yes. This is the most important part of Beatnik. It should be really
> easy to edit and publish you foaf file, just as it should be easy to
> explore the foaf relations. Is there a webdav server somewhere you
> are thinking of in particular?


as soon as I manage to get Beatnik compiled, which I am very close now
;-), I will be able to help on this and make a ("demo") server
available, but also important to me is that "people" can install their
own open social servers very easily (just as people have email servers
running decentralized).

Re WebDAV, I think it certainly makes sense to support this and Yanel
already supports this, but to be honest I don't think it's flexible and
simple enough and is another show blocker (I know "PUT" is simple, but
that's not the only operation which needs to be supported by
decentralized social networks :-)

Cheers

Michael

Michael Wechner

unread,
Dec 21, 2007, 7:41:56 AM12/21/07
to social-networ...@googlegroups.com
Danny Ayers wrote:

>On 21/12/2007, Julian Bond <julia...@voidstar.com> wrote:
>
>
>
>>>The Tabulator plugin helps:
>>>
>>>http://www.w3.org/2005/ajar/tab
>>>
>>>
>>Well that killed my browser. Oh, well, Alpha software. :(
>>
>>
>
>Yeah...get on #swig and give the man a bug report ;-)
>
>
>
>>I don't mean to be rude about the FOAF and RDF community but I just
>>don't see how it's going to become mainstream.
>>
>>
>
>Sure, I know you've been around this stuff long enough to know what
>you're talking about, got a lot of respect for your position.
>
>

I am probably not around long enough, but I like FOAF and I think it has
nice stuff in there (and also some stuff which I think is not so nice,
but maybe just don't understand well enough yet).

But I don't think this has much to do with becoming mainstream. But
rather, just don't talk about FOAF when talking to end-users. My mother
will not understand it, but she understands that I am her son and that
she would like to talk to me with one click on my animated icon on her
desktop (which is just another visualization/implementation of a FOAF
client)

Cheers

Michael

Julian Bond

unread,
Dec 21, 2007, 7:55:08 AM12/21/07
to social-networ...@googlegroups.com
Danny Ayers <danny...@gmail.com> Fri, 21 Dec 2007 12:14:22

>> - Building real world aggregators and tools for the Social data that's
>> already out there.
>
>There's the rub. Each of the the above has it's own quasi-proprietary
>view of the data.

I was really thinking here about the large quantities of FOAF and hCard
data that's already out there. There's enough data that we should be
able to get some real benefit from real applications right now.

>They may be open, but there's only minimal reuse of
>standards and commonality in the way the information is represented.
>There's no real harm as such in having lots of different, largely
>incompatible APIs or even in the kind of reinvention found in
>Attribute Exchange. But to work across this stuff, a *lot* of glue
>will be needed.

Aug 17 and Nov 1 2007 to now has been an incredibly frustrating but also
hopeful time. The hope comes from Brad's paper, almost all of the big
players announcing support for Google with Google proposing a People
Data API, along with Open ID 2.0 and AX 1.0 being announced. (Actually
Twitter's tipping point in Mar was important too) The frustration comes
from how much of this is still incomplete, barely thought through or
still just PR-ware. It's probably too much to hope for that there's one
global People/Friends API that everyone supports. But in content
syndication we've settled on two so perhaps we can hope that in the
social graph area it resolves down to a handful.

While the stock is being reduced and concentrated, the development
process still needs outliers like Twitter to demonstrate by example. It
was amazing how fast developers latched onto the minimal people data API
they provided.

BTW. OpenID AX bothers me. I'm amazed that they're using plain text
key:value pairs and slowly recreating the MS Outlook CSV format. It's
hard not to think that a more structured layout would have been better.
And a straight jump to an existing data schema rather than creeping up
on it.

>Just because FOAF isn't ubiquitous doesn't mean the approach is
>flawed, or even that there's a more compelling alternative (I can't
>see one). BradFitz's Social Graph piece was a bit of a wake-up call
>that "We're not in Kansas any more". FOAF's been sorting out the
>Munchkins in this space a long time now.

There's a perception that RDF and RDF-XML are too complicated while XML
and JSON are understandable. My own feeling is that this is the
difference between mesh thinking and tree thinking. Even though the real
world is full of 2D mesh relationships we constantly try and reduce it
to 1D trees and outlines that can be expressed as a linear flow down a
page. And our programming styles are all about iterating down lists that
don't jump sideways. This pretty much happened in FOAF with one major
style being that a FOAF file is about a single person but contains lists
of links to people that person knows with minimal cross links internal
to the file. So even though the source data is a huge mesh, the API into
it only returns trees.[1] Given that, you can pretty much always treat
and parse FOAF files in this layout as plain old XML. In which case,
where's the complexity problem?

The other source of complexity is to do with the way that FOAF grew up
as a documentation of existing usage in hand coded files. Here lies the
problem of standardisation and the relationship between data source
developers, data consumer developers and the documentation that links
them. Grapple with OWL and RDF and the relationships can be seen even if
you can't interpret the tag names used. Use a dumb XML parser and it
will break your carefully crafted interpreter or just ignore information
in the file.

[1] I bet there's a computing science thesis in there somewhere. Can any
arbitrary mesh be exploded into a set of trees and then recombined into
the original mesh? What is the effect on duplication of data in the
exploded version and is there a lower limit to that duplication.

--
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

Use It Ritually

Julian Bond

unread,
Dec 21, 2007, 8:05:44 AM12/21/07
to social-networ...@googlegroups.com
Michael Wechner <michael...@wyona.com> Fri, 21 Dec 2007 13:41:56

>But I don't think this has much to do with becoming mainstream. But
>rather, just don't talk about FOAF when talking to end-users. My mother
>will not understand it, but she understands that I am her son and that
>she would like to talk to me with one click on my animated icon on her
>desktop (which is just another visualization/implementation of a FOAF
>client)

I'm sure there are a lot of people out there who use Google reader and
have little or no understanding of RSS/Atom. We really shouldn't be hand
crafting FOAF, or asking the user where their FOAF file is.

I want to get to the point were I sign up for YASN du jour and it says
"Hi, I found your details on A,B,C,X,Y,Z and I've filled your profile
from your master web page at M. I also found your friends on A,Y,Z and
have made links to the ones that are already here. Can you give me the
URL of a profile and friends anywhere else that I've missed?" Behind the
scenes, the work will be done using protocols like FOAF but as a
customer I don't need to know that.

--
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

Use It Ritually

Michael Wechner

unread,
Dec 21, 2007, 8:10:03 AM12/21/07
to social-networ...@googlegroups.com
Julian Bond wrote:

>
> Michael Wechner <michael...@wyona.com> Fri, 21 Dec 2007 13:41:56
>
>> But I don't think this has much to do with becoming mainstream. But
>> rather, just don't talk about FOAF when talking to end-users. My
>> mother will not understand it, but she understands that I am her son
>> and that she would like to talk to me with one click on my animated
>> icon on her desktop (which is just another
>> visualization/implementation of a FOAF client)
>
>
> I'm sure there are a lot of people out there who use Google reader and
> have little or no understanding of RSS/Atom. We really shouldn't be
> hand crafting FOAF, or asking the user where their FOAF file is.
>
> I want to get to the point were I sign up for YASN du jour and it says
> "Hi, I found your details on A,B,C,X,Y,Z and I've filled your profile
> from your master web page at M. I also found your friends on A,Y,Z and
> have made links to the ones that are already here. Can you give me the
> URL of a profile and friends anywhere else that I've missed?" Behind
> the scenes, the work will be done using protocols like FOAF but as a
> customer I don't need to know that.


exactly :-)

Danny Ayers

unread,
Dec 21, 2007, 1:14:40 PM12/21/07
to social-networ...@googlegroups.com
On 21/12/2007, Tim Berners-Lee <ti...@w3.org> wrote:

> Eg. Linked Data = interop, RDF, etc,
> compare with Linked Open Data = Public linked data drive, dbpedia, etc.
>
> They are obviously connected -- you can't have LOD without LD standards.
> I think "Linked Public Data" might have been a better name.

Yup. I'm not sure who coined LOD, but I've a feeling there was
initially some intentional leveraging on the Open Data bandwagon,
something like:

http://en.wikipedia.org/wiki/Open_Data
+
http://www.w3.org/DesignIssues/LinkedData.html
=
http://esw.w3.org/topic/SweoIG/TaskForces/CommunityProjects/LinkingOpenData

The observation being there that it's one thing to make data open,
another to make it useful - which leads straight back to your point on
interop.

Bob Wyman

unread,
Dec 21, 2007, 10:29:57 PM12/21/07
to social-networ...@googlegroups.com
On Dec 21, 2007 8:05 AM, Julian Bond <julia...@voidstar.com> wrote:
> I want to get to the point were I sign up for YASN du jour and
> it says "Hi, I found your details on A,B,C,X,Y,Z and I've filled your
> profile from your master web page at M. I also found your friends on
> A,Y,Z and have made links to the ones that are already here. Can
> you give me the URL of a profile and friends anywhere else that
> I've missed?"
Errrr... Clearly I've missed something in this discussion. I can't imagine why this wouldn't be trivial if you're using OpenID and have an XRDS file. Wouldn't you just have pointers in your XRDS file that identify your FOAF data and all the other bits of needed data? Why isn't an XRDS file all you need?

bob wyman

Julian Bond

unread,
Dec 22, 2007, 4:29:46 AM12/22/07
to social-networ...@googlegroups.com
Bob Wyman <b...@wyman.us> Fri, 21 Dec 2007 22:29:57

Off the top of my head
- Implementation. Lots of OpenID 2.0 providers and consumers
- A standard for describing profiles. Which OpenID AX 1.0 is moving
towards.
- A standard for identifying a master profile. Perhaps XRDS has this.
- A standard for describing lists of friends. Currently not in OpenID.
FOAF?
- A standard for identifying people. foaf:mbox_sha1sum ?
- Implementation. "Standards without implementation are just academic
wanking"

Reply all
Reply to author
Forward
Message has been deleted
0 new messages