… and annotations (page comments, highlights, stuck and floating notes) in the Diigo sense?
independent of the XML document structure
A document is a sequence of items
is composed of an XML document and a set of
annotations.
Annotation keys and values are strings
each item conceptually has its
own annotation map
more efficient to
have just one annotation map for each consecutive run of items with
the same annotations
serialization of the document without annotations
into a string is not formally an XML document
current annotations update, which is a map of annotation
keys to pairs (old-value, new-value), where old-value and
new-value are either null or an annotation value
After the final component, the annotations update must be empty
Document operation components can be divided into four classes
do not directly affect the document or
the cursor
annotation boundaries (annotationBoundary) change the current
annotations update
interaction
with annotations
Appendix A.
Protocol Schema
message AnnotationBoundary {
// This field is set to true if and only if both ends and changes are
// empty. It is needed to ensure that the optional annotationBoundary
// component field is not dropped during serialization.
optional bool empty = 1;
// MUST NOT have the same string twice.
repeated string end = 2;
// MUST NOT have two updates with the same key. MUST NOT
// contain any of the strings listed in the 'end' field.
repeated KeyValueUpdate change = 3;
}
statistical natural language processing,
document classification, clustering, topic modeling, information extraction,
and other machine learning applications to text
It is only the new version (to be deployed in the coming weeks, most likely in January 2010) that markdown is the pivot format and you will be able to use the "latex to markdown" conversion of pandoc to upload texts.
With the present version, the best solution is either:
- Through OpenDocument and upload, which is a pain when you start from lyx, as lyx to OpenDocument does not work well.
- By generating HTML from lyx, displaying it in a browser, creating a new text in co-ment (http://www.co-ment.net/text/add/) and cutting and pasting from the browser to the edition window in co-ment. This is ugly as a method but works well if you don't have images.
or write it directly with your browser
work privately on your text with a few chosen collaborators
or open the commenting process to the public
use comments to improve your document and create new versions of your text
export your text (and all comments) in any format (MS Word, OpenOffice Document, etc.)
Polling sucks. We think a decentralized pubsub layer is a fundamental, missing layer in the Internet architecture today
a fundamental design that looks like this: This picture leaves out multiple publishers and subscribers and the subscriber registration process, but you get the basic idea
fine for server subscribers (like, say Google Reader) but not for client subscribers (like, say TweetDeck).
the only way to enable client subscribers to play in this async messaging world is via some type of relay service
In this approach, the client subscriber makes an outbound connection to some type of relay infrastructure
technically feasible
Yes, having to relay messages sucks. But the question is
I work at CENTRIM — the Centre for Research in Innovation Management. Beyond work: I enjoy solitude, letting go, camping getaways and dreaming. I have great family and friends, and two gorgeous cats. Take not too much notice of my bookmarks — hundreds are unread!