Fwd: [link-relations] NEW RELATION - annotation-server (request for review)
Ed Summers
IETF might be of interest...
//Ed
---------- Forwarded message ----------
From: "Martin J. Dürst" <due...@it.aoyama.ac.jp>
Date: Wed, Sep 7, 2011 at 9:43 PM
Subject: [link-relations] NEW RELATION - annotation-server (request for review)
To: link-re...@ietf.org
Hello Mark, Eran, Julian,
I just updated my http://tools.ietf.org/html/draft-duerst-anno-link
for the "annotation-server" link relationship. I would like to get
your (and everybody else's on this list) input on it. The registration
template is as follows:
>>>>>>>>
Relation Name:
annotation-server
Description:
Designates an annotation server used to store annotations
for the link's context.
Reference:
RFC YYYY [RFC Editor: Please replace with actual RFC number.]
Notes:
currently none
Application Data:
currently none
<<<<<<<<
I don't think I'm on this list, so please keep me in the cc.
Regards, Martin.
_______________________________________________
link-relations mailing list
link-re...@ietf.org
https://www.ietf.org/mailman/listinfo/link-relations
Bernhard Haslhofer
to find annotations for a given HTML website and it could make sense
to have something similar for OAC.
The problem with the current RFC draft is that it completely builds on
Annotea, including the Annotea publishing protocol, which we wanted
to get rid of with OAC because it is somehow obsolete since we have
REST and Linked Data. So the "annotation-server" is probably not what
we want.
One possibility could be to propose an "annotation" relationship type which
can be used to point to OAC annotations from an HTML website. Since each
OAC annotation has a dereferencable URI there is no need for providing
infos on the annotation-server. This however works only for pulling annotations;
creating and pushing annotations is currently out of the scope of OAC.
Best,
Bernhard
_____________________
Bernhard Haslhofer
Postdoctoral Associate
Cornell Information Science
301 College Ave.
Ithaca, NY 14850
Phone: +1-607-379-0831
Skype: bernhard.haslhofer
WWW: http://www.cs.cornell.edu/~bh392/
Ed Summers
On Mon, Sep 12, 2011 at 10:27 AM, Bernhard Haslhofer
<bernhard....@cornell.edu> wrote:
> The problem with the current RFC draft is that it completely builds on
> Annotea, including the Annotea publishing protocol, which we wanted
> to get rid of with OAC because it is somehow obsolete since we have
> REST and Linked Data. So the "annotation-server" is probably not what
> we want.
I don't necessarily agree that Annotea is out of step with REST.
POSTing a chunk of application/rdf+xml to the annotation server to
create an annotation, and getting a 201 back seems pretty RESTful to
me [1].
> One possibility could be to propose an "annotation" relationship type which
> can be used to point to OAC annotations from an HTML website. Since each
> OAC annotation has a dereferencable URI there is no need for providing
> infos on the annotation-server. This however works only for pulling annotations;
> creating and pushing annotations is currently out of the scope of OAC.
Yes, I Erik Wilde suggested [1] that it would be useful to say what
type of annotation server was being linked to. I don't know what the
best way to pursue this is. Is there any other way to add type
information to:
<link rel="annotation-server" href="http://example.com/annotations">
other than adding more rel values: "annotation-server-annotea",
"annotation-server-oac", etc?
//Ed
[1] http://www.w3.org/2001/Annotea/User/Protocol.html#PostingAnnot
[2] https://twitter.com/#!/dret/status/111774958081945600
Bernhard Haslhofer
On Sep 12, 2011, at 10:42 AM, Ed Summers wrote:
> Hi Bernhard,
>
> On Mon, Sep 12, 2011 at 10:27 AM, Bernhard Haslhofer
> <bernhard....@cornell.edu> wrote:
>> The problem with the current RFC draft is that it completely builds on
>> Annotea, including the Annotea publishing protocol, which we wanted
>> to get rid of with OAC because it is somehow obsolete since we have
>> REST and Linked Data. So the "annotation-server" is probably not what
>> we want.
>
> I don't necessarily agree that Annotea is out of step with REST.
> POSTing a chunk of application/rdf+xml to the annotation server to
> create an annotation, and getting a 201 back seems pretty RESTful to
> me [1].
Sure. I meant that it is not really necessary to define the publishing protocol as
part of an (annotation) data model specification because meanwhile there
are other specs that define how to publish (annotation) and retrieve data on the Web.
>
>> One possibility could be to propose an "annotation" relationship type which
>> can be used to point to OAC annotations from an HTML website. Since each
>> OAC annotation has a dereferencable URI there is no need for providing
>> infos on the annotation-server. This however works only for pulling annotations;
>> creating and pushing annotations is currently out of the scope of OAC.
>
> Yes, I Erik Wilde suggested [1] that it would be useful to say what
> type of annotation server was being linked to. I don't know what the
> best way to pursue this is. Is there any other way to add type
> information to:
>
> <link rel="annotation-server" href="http://example.com/annotations">
Wouldn't it be better to add direct links to the dereferencable annotations? Then the server-details
become transparent.
<link rel="annotation" href="http://example.com/annotations/1">
>
> other than adding more rel values: "annotation-server-annotea",
> "annotation-server-oac", etc?
What about using mime-type properties?
<link rel="annotation" type="application/rdf+xml; model=oac" href="http://example.com/annotations/1">
<link rel="annotation" type="application/rdf+xml; model=annotea" href="http://example.com/annotations/1">
>
> //Ed
>
> [1] http://www.w3.org/2001/Annotea/User/Protocol.html#PostingAnnot
> [2] https://twitter.com/#!/dret/status/111774958081945600
_____________________
Ed Summers
<bernhard....@cornell.edu> wrote:
> Wouldn't it be better to add direct links to the dereferencable annotations? Then the server-details
> become transparent.
>
> <link rel="annotation" href="http://example.com/annotations/1">
That's useful for discovering that there is an annotation, but doesn't
help a client figure out where it can create annotations does it?
> What about using mime-type properties?
>
> <link rel="annotation" type="application/rdf+xml; model=oac" href="http://example.com/annotations/1">
> <link rel="annotation" type="application/rdf+xml; model=annotea" href="http://example.com/annotations/1">
Yes, this generally seems like a good approach. Have you seen
mime-type properties used with application/rdf+xml before?
//Ed
Robert Sanderson
resource and an annotation server, or more accurately "service".
Surely one can annotate any resource, with any service? I can, and
do, post annotations to my Blogger blog, to PasteBin and to a
dedicated, personal annotation server, depending on who I want to
share the annotation with. None of which are known to, or co-located
with, the resources that I'm annotating.
Sure, one *might* use an annotation service associated with the
resource, so this is not an attack on the suggested link header, just
that it's not a requirement for annotation, and especially not for
Open Annotation.
Rob
Ed Summers
> Sure, one *might* use an annotation service associated with the
> resource, so this is not an attack on the suggested link header, just
> that it's not a requirement for annotation, and especially not for
> Open Annotation.
As you can tell I'm way behind on keeping up with what you all are
doing. So does OAC not describe a service model for creating
annotations, it's just about describing annotations that have already
been made?
//Ed
Bernhard Haslhofer
Yes, exactly. OAC describes a data model but not a service model.
> //Ed
Robert Sanderson
annotations... it just doesn't specify a required transport from the
creation environment to a storage environment.
(and note especially not "client" and "server" or "protocol")
Rob