wfinger: classic finger with WebFinger support
Chuck Houpt
A new WebFinger client implementation: I added WebFinger support to
the classic command-line finger program, and called it wfinger:
http://extechops.net/2010/08/28/wfinger/
Cheers - Chuck
Kingsley Idehen
Cool!
Some observations:
Using my non GMAIL based Webfinger Service:
kidehen$ /Applications/wfinger kid...@id.myopenlink.net
[id.myopenlink.net - web finger]
Account: kid...@id.myopenlink.net Profile: http://id.myopenlink.net/dataspace/KingsleyUyiIdehen
OpenID: http://id.myopenlink.net/openid
Using GMAIL:
kidehen$ /Applications/wfinger kid...@gmail.com
[gmail.com - web finger]
Account: kid...@gmail.com Name: Kingsley Uyi Idehen
Org: OpenLink Software Title: Technology & Executive Management
Email: Phone:
Address: USA
Profile: http://www.google.com/profiles/kidehen
OpenID: http://www.google.com/profiles/kidehen
Links:
Picasa Web Albums - kidehen: http://picasaweb.google.com/kidehen/Banff2007#
Google Reader: http://www.google.com/reader/shared/09057581521099060896
Twitter: http://twitter.com/kidehen
Del.icio.us Bookmarks: http://del.icio.us/kidehen
Slideshare Presentations: http://www.slideshare.net/kidehen
YouTube Channel: http://www.youtube.com/user/kidehen?feature=mhw5
Latest Tweet:
Traditional #UNIX Finger meets #Webfinger: http://extechops.net/2010/08/28/wfinger/ . Neat!!
Here is another view of my non GMAIL Webfinger's XRDS profile:
1. http://bit.ly/c3hgKH -- .well-known/host-meta resource
2. http://bit.ly/bu9VDC -- actual Webfinger profile .
-- Regards, Kingsley Idehen President & CEO OpenLink Software Web: http://www.openlinksw.com Weblog: http://www.openlinksw.com/blog/~kidehen Twitter/Identi.ca: kidehen
Chuck Houpt
On Aug 31, 2010, at 9:09 AM, Kingsley Idehen wrote:
> Some observations:
>
> Using my non GMAIL based Webfinger Service:
>
> kidehen$ /Applications/wfinger kid...@id.myopenlink.net
> [id.myopenlink.net - web finger]
>
> Account: kid...@id.myopenlink.net Profile: http://id.myopenlink.net/dataspace/KingsleyUyiIdehen
> OpenID: http://id.myopenlink.net/openid
Thanks, I think I see the problem, wfinger is following the Alias, rather then webfinger.net-profile-page relation. I'll check what the right order of precedence is.
Possibly there is a typo in your account XRD? Shouldn't the Alias point to the same URL as the webfinger.net-profile-page relation? The Alias doesn't have the /person/ part of the path.
Cheers -Chuck
Kingsley Idehen
Chuck,
Yes, the <Alias/> should hold my WebID (an HTTP scheme URI) rather than Profile Document URL. I'll have that fixed. Once fixed though, shouldn't you be reading the links in the XRDS en route to data similar to what you get via GMAIL?
Chuck Houpt
On Aug 31, 2010, at 10:26 AM, Kingsley Idehen wrote:
> Yes, the <Alias/> should hold my WebID (an HTTP scheme URI) rather than Profile Document URL. I'll have that fixed. Once fixed though, shouldn't you be reading the links in the XRDS en route to data similar to what you get via GMAIL?
I misspoke slightly in my last email. Currently, wfinger is basically using WebFinger to locate the account's hCard contact info (which is typically the profile page). If the hCard has rel="me" links, then those will be printed too. If no hCard relation-link is found in the account XRD, then just a stub description is printed (account, profile, OpenID).
So the differences between your Gmail and MyOpenLink wfinger-output is because MyOpenLink's XRD has no hCard-relation link.
All the foaf-OnlineAccount links in your account XRD seem to be duplicate data from your foaf-file, should they really be in the XRD? If I were to add FOAF support to wfinger, I assume I would get foaf-OnlineAccount links from a foaf-file, not the account XRD, no? I'm certainly not an expert on all these formats and how they interact...
No great thought went into wfinger's use of hCard. Since WebFinger is a greenfield protocol, I just followed the "cow paths" produced by existing WebFinger producers/consumers (for example, gmail.com and webfinger.org use hCard). Obviously, there are a bunch of ways to get basic contact info via WebFinger (hCard, FOAF, portable-contacts, etc), but hCard seemed the most widely supported.
Cheers - Chuck
Kingsley Idehen
> On Aug 31, 2010, at 10:26 AM, Kingsley Idehen wrote:
>> Yes, the<Alias/> should hold my WebID (an HTTP scheme URI) rather than Profile Document URL. I'll have that fixed. Once fixed though, shouldn't you be reading the links in the XRDS en route to data similar to what you get via GMAIL?
> I misspoke slightly in my last email. Currently, wfinger is basically using WebFinger to locate the account's hCard contact info (which is typically the profile page). If the hCard has rel="me" links, then those will be printed too. If no hCard relation-link is found in the account XRD, then just a stub description is printed (account, profile, OpenID).
>
> So the differences between your Gmail and MyOpenLink wfinger-output is because MyOpenLink's XRD has no hCard-relation link.
>
> All the foaf-OnlineAccount links in your account XRD seem to be duplicate data from your foaf-file, should they really be in the XRD?
Why not?
XRD resource is just another representation of metadata associated with
a given Subject.
> If I were to add FOAF support to wfinger, I assume I would get foaf-OnlineAccount links from a foaf-file, not the account XRD, no? I'm certainly not an expert on all these formats and how they interact...
>
You can choose to follow my WebID URI (a Generic URI) which will resolve
to a representation of a FOAF based profile document. Doing that will
provide you with the same data in the XRD.
> No great thought went into wfinger's use of hCard. Since WebFinger is a greenfield protocol, I just followed the "cow paths" produced by existing WebFinger producers/consumers (for example, gmail.com and webfinger.org use hCard). Obviously, there are a bunch of ways to get basic contact info via WebFinger (hCard, FOAF, portable-contacts, etc), but hCard seemed the most widely supported.
But you don't really have to think in terms of "widely supported"
vocabularies etc., just project the links as expressed in the <link/> +
@rel associations, no more no less :-)
Kingsley
> Cheers - Chuck
>
>
--
Regards,
Kingsley Idehen
President& CEO
Chuck Houpt
On Sep 1, 2010, at 7:18 AM, Kingsley Idehen wrote:
>> All the foaf-OnlineAccount links in your account XRD seem to be duplicate data from your foaf-file, should they really be in the XRD?
>
> Why not? XRD resource is just another representation of metadata associated with a given Subject.
Well, I think there are reasons not to include duplicate or extraneous data in the account XRD:
One reason is a simple network rule of thumb: "don't transmit more then needed." An account XRD will be used by many different tools, accessing different linked resources. One tool might only need OpenID info, another needs contact info, yet another needs calendar info -- why make all these tools download FOAF-OnlineAccount data, when they don't need it? This is particularly important for low-bandwidth use (phones, etc), where every byte counts.
Another reason is that, at least to me, it doesn't make sense to include fragments of a data records in the XRD. What good is a foaf-OnlineAccount link, if the rest of the foaf-record is not present? From your account XRD, I can see that you have a YouTube account, but I don't know your full-name, so I have to fetch your foaf-file anyway. My impression is that the intended use of the account XRD is as a dispatch-table directory of links to different data records for the account. The links should be to whole/complete data records (hCards, open-contact JSONs, FOAF record, etc), not links to the users online accounts (those should be in hCard urls fields, FOAF OnlineAccount fields, etc).
Anyway, those are just my thoughts -- I realize the issue of what account-XRDs should contain is not well-defined.
- Chuck
John Panzer
Well, I think there are reasons not to include duplicate or extraneous data in the account XRD:
On Sep 1, 2010, at 7:18 AM, Kingsley Idehen wrote:
>> All the foaf-OnlineAccount links in your account XRD seem to be duplicate data from your foaf-file, should they really be in the XRD?
>
> Why not? XRD resource is just another representation of metadata associated with a given Subject.
One reason is a simple network rule of thumb: "don't transmit more then needed." An account XRD will be used by many different tools, accessing different linked resources. One tool might only need OpenID info, another needs contact info, yet another needs calendar info -- why make all these tools download FOAF-OnlineAccount data, when they don't need it? This is particularly important for low-bandwidth use (phones, etc), where every byte counts.
Kingsley Idehen
Your personal views. In a sense this is part of the problem the more
Semantic Web oriented keep on try to solve (more self describing data so
that machines can figure stuff out for themselves). You've just
described a personal perspective (which you are 100% entitled to) which
manifests in your code (your right too), but that doesn't really define
the realm of structured profile discovery and use.
You have a microformats bent in your wfinger and as such it skews my non
GMAIL profile. If you are capable of processing a FOAF profile, then my
WebID will give that to you in the data serialization formats that it
supports. All you (or your agent) has to do is leverage in-built HTTP
functionality re. content negotiation :-)