Attention.XML suggestions and related discussions
-
Up to the
![[WWW]](https://cdn.statically.io/img/web.archive.org/web/20060206133653im_/http://developers.technorati.com/content/rightsidebar/img/moin-www.png)
attention.xml spec
Conceptual Discussions
Related and Similar Work
-
http://www.25hoursaday.com/draft-obasanjo-siam-01.html (if unavailable, try http://web.archive.org/web/20040206093727/http://www.25hoursaday.com/draft-obasanjo-siam-01.html )
-
Attention.XML would solve many of the problems/issues raised in that article.
Comments
Technorati Cosmos of Attention.XML
Posts:
-
[ftma://www.richardgiles.net/blog/archives/2004/12/revolution.html Richard Giles]
Discussed and to be added to v2004-10-01
-
A general mechanism for tagging/annotating URLs, similar to flickr and the del.icio.us bookmarking service. Thanks to
Jeremy Zawodny for the suggestion.
-
A single property for both sites and pages would be sufficient.
-
Perhaps "usertags" which would take an unordered list (could be minimized with a space separated attribute).
-
Perhaps this property should simply be a replacement for the "groups" property which can cause confusion with a top level groups object.
-
Danny Ayers suggested using or referencing dc:subject. Perhaps this property should be named "subject" and reference dc:subject as the normative definition. Though the term "subject" implies a classification within an authoritative taxonomy (perhaps by the author, perhaps by a library), whereas the intent of this property is to capture the users's keywords for this URL. Thus "subject" does not quite have the right connotation.
-
Timothy Appnel suggested the simple name "tags" for the property.
-
DECISION: new property "tags" on both blogs/sites and posts/pages.
-
Perhaps add "altURLs" as a property for items as well, so that sources can reference alternative URLs for items, which must of course be reflexive in order to be trusted.
-
DECISION: rejected/postponed for now. "same" posts/items should be determined by client applications using hashes as an optimization.
Free Discussion
-
The size of attention.xml files will get quite large. Clarify what portions of the HTTP 1.1 standard will be used to retrieve portions of this file. BradNeuberg
-
Additional discussion is needed on the topic of trimming the attention file, including the inclusion of a last trim date.
Eric Hayes
-
In addition embedding trim date information will allow separate attention.xml files from the same user to be aggregated without requiring HTTP headers. -
Tantek
-
Authentication will be an issue when retrieving these files; clarify what kind of HTTP authentication will be used to retrieve attention.xml files. BradNeuberg
-
Brad, I believe this will end up being an implimentation detail. There will have to be a privacy agreement between the user (creator/owner of the attention.xml file) and the service they are choosing to share their attention data with. This should cover in plain words use, access, and any potential propogation of the attention.xml file.
Eric Hayes
-
Should attention.xml use a more XML-like layout for its data-format, extend OPML, or stay with the XHTML-XOXO format? We should determine which one is easier for developers to grok and implement. BradNeuberg
-
Does it make sense to break attention.xml into several smaller standards, such as one for syncing, one for social networking, one for actual attention information, etc.? BradNeuberg
-
I'm interested in making putting music selection and presence information into attention.xml, allowing for future analysis of things like "he likes to play the following music when at the office". ChristopherAllen
-
Eric Hayes made a good argument for putting in unread items and deleted items, especially unread items that have been deleted. We need to consider adding this to the format. -
Tantek
* According to this page, attention.xml uses the XOXO format, but on 21 Dec. 2005 (yesterday) this page mentioned format for attention.xml which is the OPML format <outline text="title" description="" ...> for the post and item elements? Is that change because you are not sure about the format that you will use for data representation of the specification or what? In my research I'm trying to use attention.xml to represent attention users give to pages/documents they read, etc. It's not clear which format to use yet? Is there an XML schema for the attention.xml spec..
Another question: would not it be easier and more controllable to use the XML (at least the OPML format) like format to represent the attention data? This will enable making some calculations (metrics) on the data, also easy to store and retrieve these data in XML databases if needed. - Najjar