[foaf-protocols] Some basic questions prior to development ?

Story Henry henry.story at bblfish.net
Mon May 17 21:00:52 CEST 2010


On 17 May 2010, at 19:49, Seth Russell wrote:

> I have a potentually big problem with getting then parsing a RDF document
> and then using SPARQL. to retireve information from a profile that may or
> may not contain the information that i need.

You are using PHP I think, so Melvin and co will be the best people to help you here.

>  In this case i need email, but
> later when i use this for the ordering system i would also need name and
> address information.

Those are usually in the foaf file.

I suppose if it's not, THEN you ask the person, the way we do it currently.
This will give people an incentive to put the information in their foaf :-)

>  How fast will that process be?  

1 HTTP connection on the server should do it.

> How much development
> time?  

PHP RDF libraries are still a bit early days, but with Drupal embracing the semantic web, this should change fast. I don't think it's much coding.



> How many people will be putting their email address in their profile
> anyway?

A lot of us to start with, as we are already getting spammed. Later fewer, but then we will have authentication systems to help as explained with the photo printing service.

Applications like your are essential to help us along, as they are the motivator behind getting WebIds.

> 
> Then too, lurking in the back of my mind, is my worry that there is a big
> part of this process still missing.  Doesn't the customer need to select in
> their browser somewhere what information they want to share with a
> particular web site?

That is what the photo printing service sketch is about.
Clearly this is an important part of the protocol. We should develop it further

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


>  Where and when is that customer input to take place?

I think that is application specific.

> I don't see any place where that has been even specified yet, let alone
> implemented.

This is perhaps what we need to work on next. Really we have been waiting
for feedback to get an idea what we should be working on.

Currently we are still working a lot on the ease of use of creating webids.
But I think the next piece will be apps like yours, then more access control.

I think if you can see where it is heading you'll be happy to develop as things stand
now, and we will improve as things go along.

Henry

> 
> Seth Russell
> Alpha testing: tagtalking.net
> Facebook ing: facebook.com/russell.seth
> Twitter ing: twitter.com/SethRussell
> Blogging: fastblogit.com/seth/
> Catalog selling: www.speaktomecatalog.com
> Google profile: google.com/profiles/russell.seth
> 
> 
> On Mon, May 17, 2010 at 11:24 AM, Story Henry <henry.story at bblfish.net>wrote:
> 
>> 
>> On 17 May 2010, at 18:36, Seth Russell wrote:
>> 
>>> I did manage to implement your Identity Provider pretty easily.  You can
>>> start my simple application at:
>>> 
>>>     http://www.speaktomecatalog.com/login.php
>> 
>> Absolutely adorably cute! :-)
>> 
>>> 
>>> Do, please try it out and give me feedback.
>> 
>> I did not win anything!
>> 
>>> It will email me your spin but
>>> since my payoff is tied to the email i would have dispatched to you, and
>>> your service does not give me your email address, I can't yet put the
>> payoff
>>> coupon in your shopping cart yet.   Which brings me back to basic
>> questions:
>>> 
>>>  - How can a web developer easily get information from the customer's
>>>  profile?
>> 
>> Ok, So you could do an HTTP GET on the WebId. In my case that was
>> 
>> http://bblfish.net/#hjs
>> 
>> And you can parse the representation into an RDF graph (a set or
>> relations).
>> You need an RDF library for this. Not sure which language you are using.
>> Essentially you need to check the mime type of the returned representation
>> and use the appropriate parser. This is usually application/rdf+xml or
>> text/html or application/xhtml+xml
>> 
>> Your RDF library will return you a handle onto a graph object, which you
>> can query with SPARQL. As you are looking for an e-mail you could query it
>> for
>> 
>> PREFIX foaf: <http://xmlns.com/foaf/0.1/>
>> SELECT ?mbox
>> WHERE {
>> <http://bblfish.net/#hjs> foaf:mbox ?mbox .
>> }
>> 
>> You have not verified that mailbox at this point yet, but you could think
>> of your first email as being such a verification.
>> 
>> We were thinking of developing a ping mechanism, which Sebastian Tramp is
>> working on,
>> where you could POST a simple message to an enpoint. The endpoint would be
>> found in a similar manner. (we don't have a namespace yet)
>> 
>> SELECT ?ping
>> WHERE {
>> <http://bblfish.net/#hjs> fp:ping ?ping .
>> }
>> 
>> This would have the advantage of not requiring you to have an email. It
>> could
>> also itself be protected by foaf+ssl, so that your agent would have to
>> identify itself, and thus at least be somewhat locatable, and advantage over
>> e-mail.
>> 
>> 
>>>  - If all the information in the profile is public, and it sure seems to
>>>  me that it is, why would you expect that everyone wants to publish
>> their
>>>  emails, names, and address for all the world to see?
>> 
>> That is partly why we are looking at ping.
>> The next thing we can do is also develop something like OAuth, where in one
>> click we
>> give your agent access to the ping or email address.
>> 
>> See:
>>  http://blogs.sun.com/bblfish/entry/sketch_of_a_restful_photo
>> 
>> We just need to implement services that know how to make these access
>> control
>> decisions. It's not that difficult. If we get ping working, we could try
>> doing that
>> next.
>> 
>>>  - Is there a service now that will allow me to add information to my
>>>  profile now in a user friendly manner which does not assuming that i
>> know
>>>  how to write RDF?
>> 
>> I think http://webid.myxwiki.org/ allows you to create a very simple
>> account to do this.
>> It would be great if http://foafbuilder.qdos.org/ gave its users secure
>> WebIds, as it has a good editing interface. I think Melvin is working on an
>> Elgg implementation that does
>> this too.
>> 
>> This is where we need to work a lot more.
>> 
>>>  - Is there a URL to my profile that shows my picture which i can
>> publish
>>>  in my  emails?  The one below goes to a RDF document which is very
>>>  unfriendly to  non-geeks.
>> 
>> You mean of foaf.me? Not sure.
>> 
>> On http://webid.myxwiki.org your home page is the thing to point people
>> to.
>> 
>>> 
>>> PS:  I do like WebId and the logo much better than foaf+ssl which is
>> totally
>>> too geeky for to ever catch on :)
>> 
>> foaf+ssl was not meant to catch on :-) We just needed a handle to get
>> started.
>> We think of it as the geeky label, a bit like Mozilla used to be to
>> Netscape.
>> 
>> Great work Seth, and thanks for this very helpful non geeky feedback. We
>> need it.
>> 
>> 
>> Henry
>> 
>>> 
>>> Seth Russell
>>> WebId:http://foaf.me/seth
>> 
>> :-)
>> 
>>> Alpha testing: tagtalking.net
>>> Facebook ing: facebook.com/russell.seth
>>> Twitter ing: twitter.com/SethRussell
>>> Blogging: fastblogit.com/seth/
>>> Catalog selling: www.speaktomecatalog.com
>>> Google profile: google.com/profiles/russell.seth
>>> 
>>> 
>>> On Fri, May 14, 2010 at 11:28 AM, Story Henry <henry.story at bblfish.net
>>> wrote:
>>> 
>>>> 
>>>> On 14 May 2010, at 18:57, Seth Russell wrote:
>>>> 
>>>>>> FOAF is supported by google and there's at least one bridge to twitter
>>>>> http://semantictweet.com/
>>>>> 
>>>>> Wow! I am truly impressed with how much information that provides.
>> One
>>>>> big question:  if people use that information and write in their foaf
>>>>> profiles that they foaf:knows me as
>> http://semantictweet/SethRussell#mebut
>>>>> i am accumulating information about myself in my own profile where i
>>>>> consider myself to be http://foaf.me/seth#me  won't we all get just a
>>>> little
>>>>> bit confused?
>>>> 
>>>> That's a bit like the web is today. You have a Facebook profile, and
>>>> perhaps a
>>>> linkedin one, then other services on the web create a service to extract
>>>> the data
>>>> and they publish it somwhere else.  The difference is that most of your
>>>> friends will be linking to http://foaf.me/seth#me because that is the
>> one
>>>> you give out on a business card, or append to your emails.
>>>> 
>>>> I think of semantictweet as a demo of what twitter could be doing if
>> they
>>>> wanted to. Indenti.ca produces its own foaf so there is no need for
>>>> semantictweet.
>>>> 
>>>> 
>>>>> Shouldn't you have some way in your tool to ask me not only
>>>>> who i am to twitter but where i consider my official profile to reside?
>>>> 
>>>> As I said above, that is the one that you put in your emails.
>>>> Also when people link to it, they get to go to your parties. I think
>> we'll
>>>> see how this evolves in a pragmatic way.
>>>> 
>>>>> 
>>>>> Is there, or can there be such a tool for facebook?  For Google?
>>>> 
>>>> Yes, they would just need to follow the procedure indicated here
>>>> 
>>>> 
>> http://esw.w3.org/Foaf%2Bssl/HOWTO#HOWTO_foaf.2Bssl_enable_your_Web_2.0_application
>>>> 
>>>> 
>>>>> How are
>>>>> all of these different foaf identities of me ever going to ever get
>>>> equated?
>>>> 
>>>> You can link them in various ways. One way would be to
>>>> 
>>>> <http://www.facebook.com/russell.seth#me> owl:sameAs googleseth:me .
>>>> 
>>>> from one account to the other.
>>>> We can start looking into that as soon as we have enough apps so that
>> the
>>>> problem becomes more practical. There are a number of solutions to this
>>>> problem.
>>>> 
>>>> Hope this helps,
>>>> 
>>>> PS. Don't forget to send us a pointer to your app!
>>>> 
>>>> 
>>>>> Seth Russell
>>>>> Alpha testing: tagtalking.net
>>>>> Facebook ing: facebook.com/russell.seth
>>>>> Twitter ing: twitter.com/SethRussell
>>>>> Blogging: fastblogit.com/seth/
>>>>> Catalog selling: www.speaktomecatalog.com
>>>>> Google profile: google.com/profiles/russell.seth
>>>>> _______________________________________________
>>>>> foaf-protocols mailing list
>>>>> foaf-protocols at lists.foaf-project.org
>>>>> http://lists.foaf-project.org/mailman/listinfo/foaf-protocols
>>>> 
>>>> 
>> 
>> 



More information about the foaf-protocols mailing list