Attention.XML Frequently Asked Questions
This page documents questions and answers regarding ![[WWW]](https://cdn.statically.io/img/web.archive.org/web/20051217093958im_/http://developers.technorati.com/content/rightsidebar/img/moin-www.png){kind=small}
Attention.XML.
Q: When I query the Attention API, the response is "No Attention Data Found". Do I need to create my own Attention.xml and post it before a populated response will be received? Or in what other manner will my Attention.xml by populated?
Q: If the 'dateremoved' of a blog/feed is received by a server after a 'dateadded' but the date/time specified by the "remove" client precedes the "add" client, who wins?
A: The "add" client. This is a typical dueling offline clients scenario. Ironically, in this case you can't trust the "when the client contacted the server" time. Because you may happen to connect the "more up to date" client to the server before you connect the "older" client.
Q: I want to share my Attention.XML with other users so they can see what I read. How do I do that?
A: This makes a lot of sense. There are at least two ways that this could work. First, you could designate your Attention.XML to be read-only to the general public (or perhaps just other members of a service providing AttentionQuery support like Technorati). Second, you could indicate that you wanted your social network, or some particular slice thereof, to have access to your Attention.XML. Either way, this depends on an Attention.XML hosting/syncing service providing a way to view/edit such preferences. We are looking into doing this.
Q: Could you support a more secure method for sending member user/pw info in the AttentionQuery API?
A: This is a great suggestion. We're looking at changing AttentionQuery to first at least use HTTP-Authentication for user/pw, and eventually supporting SSL (e.g https).
Q: What is the value of supporting XFN in Attention.XML?
A: An (X)HTML blogroll, as seen in nearly every blog today, is essentially a degenerate Attention.XML structure. Blogs and blogrolls are forming a loose decentralized social network that many are enhancing and making explicit through XFN relationships, thus it makes sense to preserve those annotations in Attention.XML. In addition, this feature enables Attention.XML implementations to more readily filter and prioritize your attention using your social network.
Q: While certainly interesting, I can't help think that this is a tremendous amount of information that is being recorded regarding user habits. So much that I can see privacy groups possibly raising concern regarding this. Has there been any thought of having the data storage be done on a third machine (e.g. a user hosted domain?) While this would not solve all of the potential privacy issues, it would go some way to ensure that users have some level of control over their own content (provided the trust between themselves and their current attention.xml provider remains comfortable.) --jrc
A: jrc, you're absolutely right, there are privacy implications with attention information, as with any information that is user-related. We have thought about the privacy issues, and there are two fundamental aspects to Attention.XML that help its privacy "awareness". The first is that Attention.XML format is completely anonymous. There is no name, address, phone number etc. or any other directly identifying information in the format. The second is that the format is designed such that the user is in control of their own data. The user explicitly chooses where to store it (whether locally, with a service like Technorati, or some other domain), and the user explicitly chooses which applications to use it with.
Q: Could you Integrate this with the (RDF) menow schema? Or could you work with them, at least?
A: At first blush, "menow" appears both premature/arbitrary (there doesn't seem to be any methodology to what is or is not in the spec), and hugely overdesigned (like most RDF efforts). Attention.XML focuses on current feed reading use cases, and is designed to be simple, minimalist, and easy to use and implement. Given the different apparent design centers, it will likely be much more beneficial to develop independent specifications and then perhaps contrast them.
Q: How should I track read items where no specific information such as duration or lastread is available?
A: By its mere presence in the data, an item is implied to have been read. If other information is not available, the respective fields should be omitted.
Q: May I troll here?
A: No, trolling is not welcome, and trollcomments will be deleted.
Q: How do I add a question?
A: Click the "Edittext" link down below, and then type in "Q:" followed by your question at the end.