Web annotation systems are meant to give electronic documents some of the same note-taking possibilities as paper documents. Unlike paper system, annotating on a web page will not modify the web page itself. The problem is, when page loaded, how to identify the target of each annotation. It is not a trivial matter, for the content and structure of web page may change from time to time. In this article, we proposed a feed based annotation method. Unlike web page, feed document is stable. We try to solve this problem by mapping web page annotations to their back-end feed document.
Page 1 of 5
Web Annotation on Feed
1. Background
How many times have you, when reading paper documents, taken notes in the margin of the page, or just underline/circle some important text on the page? Web annotation systems are meant to give electronic documents some of the same note-taking possibilities as paper documents.
When annotating on a paper document, the annotations, in fact, have become part of the document. So when you re-open the document,
annotations on their right place. But for web annotation, things are quite different. We are discussing public annotating service. People can annotate on any web pages from any web application server. Annotating on a web page will not modify the web page itself. Annotations are stored centralized on a separated annotation server. If later someone re-open the web page and ask to see these annotations, annotation service is responsible to find the exact elements that theses annotations are associated with, and display them near these elements. The problem is: how to identify the annotation target, such that each annotation can be displayed on its right place always?
Here are some known solutions:
1. Pixel offset against the browser window.
Used by many existing web annotation tools (including World forum , Trailfire, Fleck, Stickis , MyStickies , DrawHere ). They do not care about what element does people annotating on. They just remember the position of annotations (sticky notes) inside the browser window. The drawback is obvious. People is always annotating on html content instead of browser window. Most web pages adopting floating layout. Some of the page content will automatically wrapped to another line when the browser window becoming smaller. That is when resizing browser window, the pixel offset of html contents against the browser window may change from time to time, but annotations do not! Just resizing the browser window, the annotations may shiftaway from the html contents they are annotated on.
2. Location in the document.
Used by Marginalia and Annotea. For these annotation services, each annotation is associated with a location in the document -- when someone annotate on part of a web page, it is remembered the offset of the target html element against the root element in the html document (use XPointer or like). The drawback is that unlike paper document, web pages may change from time to time. If the structure of the page (the page layout)is changed, the annotations may no longer display properly.
3. Text match.
Some annotation service (Diigo, Gibeo , Linebuzz, etc.) will stores the text you selected, when you doing annotation. To locate annotations,they searches the content for matching text.
The drawback:
Firstly, what if there are multiple matches on the page? There's no good solution yet.
Secondly, what if I annotate on a paragraph just tocomment on its topic? (Even the content of the paragraph has been changed, the topic is not change...
