A post on rel=me

[Bob Ngu]

Since the first proposed DIY project is rel=me implementation, I
finally got some time today to look further into it and wrote a post
about it.
http://ungeekdapo.wordpress.com/2008/04/29/a-simple-data-portability-project-or-is-it/

Bottom line, IMO, rel=me and XFN are not ready for prime time, I
welcome your thoughts and feedback.

[Bob Ngu]

Phil asked the following questions after reading my post, I am happy
to share my perspectives and welcome any and all feedback.
Q. How do you find and recognize opportunities for using rel=me in
your business?
Like I said in my post, at this point, rel=me proliferation is limited
and experimental at best, so there isn't much to capitalize on for
business. By itself, rel=me is a very simple tweak to standard HTML
and it is only good for announcing one multiple online identities or
profiles. As such, MyBlogLog has implemented something similar, About
Me widget, though not using XFN. They also recently added support for
FOAF - a topic for another post. In the long run though, I see rel=me,
XFN, microformats, FOAF, etc as basic building blocks for data
portability efforts and there are definitely opportunities there for
the taking.

Q. Does the rel=me attribute define the relationship between the text
in the link and the url? or between the originating url and the
destination url?
If I understand you correctly, rel=me attribute defines the
relationship between the originating url and the destination url. Note
that this can be problematic because it assumes that there is only one
person attached to the originating url. For example, if we jointly
write a post, we can't put both our online profiles in the post with
rel=me, it just isn't designed to work that way.

Q. What are the risks associated with consolidating someone's identity
this way?
At this point, zero risk or rather explicit consent assumed because a
person has to physically tweak HTML to add rel=me unless there is
implicit support built into certain sites without their knowledge. I
wonder if a DiSo WordPress plugin already does something like this, I
am not sure, if it does, it should clearly state that this information
can be publicly consumable by XFN readers such as Google Social Graph
APIs and request the user to explicitly opt-in. The exact text should
be worded properly because most users won't understand what XFN and
Google Social Graph APIs mean.

Q. Should you ask for a person to opt-in on rel=me before applying it
to their identity?
See my answer to the above question.

On Apr 29, 4:56 pm, Bob Ngu <bob_...@yahoo.com> wrote:
> Since the first proposed DIY project is rel=me implementation, I
> finally got some time today to look further into it and wrote a post

> about it.http://ungeekdapo.wordpress.com/2008/04/29/a-simple-data-portability-...

[J. Trent Adams]

Tangential Comment:

I get why the rel=me concept is a good first step since it's basically
something anyone with a blog can do. I wonder, though, how it plays
with the more context-rich FOAF representation that can contain the
same information.

Many of the questions posed by Phil are more easily addressed within
FOAF, though I do understand it (currently) takes a bit more mojo to
build it out.

My $0.02,
=jtrentadams

[danielabarbosa]

My understanding is that it could benefit users that does it to their
one sites but that vendors (twitter,friendfeed etc) are also
utilizing. so the message needs to be for both.

see this an comments:
http://www.scripting.com/stories/2008/04/28/overlookedDetail.html#disqus_thread

looking forward to the education and discussion tonight!

[danielabarbosa]

ok that was the worst spelling and grammar EVER let's try it again....

My understanding is that it could benefit users that use it on their
own sites but that vendors (twitter,friendfeed etc) are also

utilizing. so the message needs to be for both.

see this and especially the comments:
http://www.scripting.com/stories/2008/04/28/overlookedDetail.html#dis...

looking forward to the education and discussion tonight! (and
perhaps some spelling and grammar lessons ;-)

[Kevin Marks]

Bob, the post is a good summary of the chicken versus egg problems
that are represented by standards adoption. The Social Graph API is an
attempt to help eas adoption by caching the distributed graph in a
useful form for developers to take advantage of.

rel="me" and the other XFN types are useful to avoid the 'please re-
enter all your public connections' problem for signing on to websites.

On Apr 29, 10:40 pm, Bob Ngu <bob_...@yahoo.com> wrote:
> Phil asked the following questions after reading my post, I am happy
> to share my perspectives and welcome any and all feedback.
> Q. How do you find and recognize opportunities for using rel=me in
> your business?
> Like I said in my post, at this point, rel=me proliferation is limited
> and experimental at best, so there isn't much to capitalize on for
> business. By itself, rel=me is a very simple tweak to standard HTML
> and it is only good for announcing one multiple online identities or
> profiles. As such, MyBlogLog has implemented something similar, About
> Me widget, though not using XFN. They also recently added support for
> FOAF - a topic for another post. In the long run though, I see rel=me,
> XFN, microformats, FOAF, etc as basic building blocks for data
> portability efforts and there are definitely opportunities there for
> the taking.

I'd encourage mybloglog to adopt this as twitter, friendfeed etc
already have.

> > Q. Does the rel=me attribute define the relationship between the text
> in the link and the url? or between the originating url and the
> destination url?
> If I understand you correctly, rel=me attribute defines the
> relationship between the originating url and the destination url. Note
> that this can be problematic because it assumes that there is only one
> person attached to the originating url. For example, if we jointly
> write a post, we can't put both our online profiles in the post with
> rel=me, it just isn't designed to work that way.

right, rel="me" is for the subset of pages used as representations of
individuals. You can express joint authorship with hcard in hAtom, but
that is a different problem.

> Q. What are the risks associated with consolidating someone's identity
> this way?
> At this point, zero risk or rather explicit consent assumed because a
> person has to physically tweak HTML to add rel=me unless there is
> implicit support built into certain sites without their knowledge. I
> wonder if a DiSo WordPress plugin already does something like this, I
> am not sure, if it does, it should clearly state that this information
> can be publicly consumable by XFN readers such as Google Social Graph
> APIs and request the user to explicitly opt-in. The exact text should
> be worded properly because most users won't understand what XFN and
> Google Social Graph APIs mean.

Wordpress already supports XFN relationships by default. I think that
adding XFN need not be opt-in when the intent of declaring a
connection publicly is clear (as in the twitter 'enter your url'
field) This is more problematic in other contexts (for example in the
FOAF SHAsum of email case, where email addresses can be leaked). The
advantage of annotating links in thsi way is that the user can already
see that the connection has publicly been made.

>
> Q. Should you ask for a person to opt-in on rel=me before applying it
> to their identity?
> See my answer to the above question.

No.

>
> On Apr 29, 4:56 pm, Bob Ngu <bob_...@yahoo.com> wrote:
>
> > Since the first proposed DIY project is rel=me implementation, I
> > finally got some time today to look further into it and wrote a post
> > about it.http://ungeekdapo.wordpress.com/2008/04/29/a-simple-data-portability-...
>
> > Bottom line, IMO, rel=me and XFN are not ready for prime time, I
> > welcome your thoughts and feedback.

What would be sufficient to make them ready for it in you estimation?
That will help with setting goals.

[Bob Ngu]

Kevin,

> Bob, the post is a good summary of the chicken versus egg problems
> that are represented by standards adoption. The Social Graph API is an
> attempt to help eas adoption by caching the distributed graph in a
> useful form for developers to take advantage of.

SG APIs is a good thing but to be honest, I am not comfortable
consuming (and showing) data that is potentially stale or missing
because it hasn't been crawled or updated.

> I'd encourage mybloglog to adopt this as twitter, friendfeed etc
> already have.

I subsequently found out that mybloglog supports XFN as a producer
like most everyone else.

> Wordpress already supports XFN relationships by default. I think that
> adding XFN need not be opt-in when the intent of declaring a
> connection publicly is clear (as in the twitter 'enter your url'
> field) This is more problematic in other contexts (for example in the
> FOAF SHAsum of email case, where email addresses can be leaked). The
> advantage of annotating links in thsi way is that the user can already
> see that the connection has publicly been made.

Thanks for pointing out Wordpress support for XFN, it is great to
learn about all these sites with XFN support that I am not aware of.
Note that Wordpress like many of the other sites only support XFN as
producers, not consumers. I still think it is worthwhile to provide
plain English verbiage to explicitly let users know that this
information can be publicly consumed in a certain way. They may know
(what XFN is) and not care, in which case it's all good but for those
that don't know (what XFN is) and care, then they can choose not to
enable it.