Proposer: Iain Stewart, room 220a Huxley, e-mail ids@doc.ic.ac.uk
This project was inspired by Crit - a marvellous suite of "critical discussion tools for the web" developed by Ka-Ping Yee for the Foresight Institute. Crit allows users to annotate the web by adding comments of their own to any web page - theirs or anyone else's, stored anywhere in the world - for all to see.
When people first hear of this their reaction is usually one of disbelief. Surely that's impossible without an act of malice - hacking into the web server (or file server or whatever) hosting the page in question and tampering with its contents?
Neither Crit nor this project is about annotating the web that way! Crit achieves its apparent miracle without malice or coercion by arranging for the annotations to be seen only by those who choose to see them. The target page remains quite undamaged and un-tampered with. So people who view it in the regular way continue to see it just as its author intended.
At the moment, the way you choose to see the web in annotated form
(and to make any annotations of your own) is to view the web through a
Crit server.
This is achieved by adding a prefix (pointing to the Crit server)
to the URL of a web page you want to view.
For example, this project proposal web page has a URL of
"http://www.doc.ic.ac.uk/~ids/dotdot/misc/projects/AnnotatingTheWeb/proposal.html".
If you're reading this from within IC DoC you can view it in annotated form
using the URL
"http://www.doc.ic.ac.uk/lab/nph-crit.cgi/http://www.doc.ic.ac.uk/~ids/dotdot/misc/projects/AnnotatingTheWeb/proposal.html"
instead.
Annotations, if there are any, appear as delicate little arrow-like symbols
which you can click on to see the content of.
The facility of adding your own annotations is currently offered in the form of
a clickable "comment" button
leading to a form-based submission process.
The Foresight Institute hopes that critical discussion tools like Crit will improve the quality of the worldwide scholarly process of creation and critical evaluation of research ideas. That hope may err on the side of optimism, given the obvious possibilities for less than optimal use (not to say abuse) of this kind of system, but annotating the web is certainly a fascinating concept. This project would be about adding your contribution to the ongoing development of web annotation concepts and software worldwide. This could mean either building on the existing Crit software (or similar efforts by others), or developing your own - or a bit of both of course. This is your chance to go down in cyberhistory by helping to make a real difference to the whole future look and feel of the web. If you find the web annotation idea intriguing and think you have the coding and software engineering skills to contribute to progress in this area, this could be the project for you!
Here are some sample avenues for further development.
Decentralizing the annotations structure and making annotations "free as air". The description above makes it sound as if all the annotations have to be stored on the Crit server. This is not in fact the case: pages that comment on other pages can be stored anywhere on the web, and merely viewing them through the Crit server is enough to bring the annotation links in question to the server's notice and have them seen by other users of that server from then on. However, what is the case is that the "meta-level" information - saying which pages comment on which other pages - currently has to be stored on the server. This is obviously not scalable, and indeed the original Crit server died some time ago, after showing ever greater signs of strain towards the end!
A better arrangement for the longer term might be to treat annotation meta-level information something like Usenet news. That is, it could propagate from annotation server to annotation server and be collected by clients (browsers) from whichever server was nearest or most convenient. If one's local server was having trouble storing the whole table of annotation information, it could perhaps store a subset and tell browsers to collect further annotations from a server elsewhere which stored a different subset. - Or something like that anyway!
Annotation-aware browsers. The Crit folk have achieved the magnificent feat of having their system work straight away with existing browsers. This lets people see the concept in action with the minimum of fuss. In the longer term, though, it would presumably be helpful if browsers were to become "annotation-aware". If nothing else, this would take some of the load off the annotation server by having as much as possible done by the client - currently even little things, like peppering the page with the appropriate annotation marker images in the right places, have to be orchestrated at the server end. But beyond that, it might well be that annotation-aware browsers open up entirely new possibilities for the treatment of critical comment and discussion on the web.
Pete Zaborszky, who did this project in 2005-06,
created an annotation-aware extension to the Mozilla Firefox browser
which talks to an annotation server also of his creation.
You can see his work, including any further development since then,
at his website:
.
The treatment of abuse. It's perhaps inevitable that some people and companies will create "junk annotations" by analogy with the spam that currently infests e-mail and Usenet news. There are various possible approaches to combating this. Annotation-aware browsers could support "kill-files": instructions not to show annotations from certain sources. This has had only limited success in the case of mail and news spam, and it might be necessary to advance to server-level filtering, whereby junk annotations are detected and removed from the database (or not included in the first place).
I mention these issues of abuse simply for completeness: it's maybe a bit dispiriting to get bogged down in this side of things as a choice of project area at such an early stage in the game, and probably this year we can just forget about it and trust to goodwill. Time will tell to what extent abuse proves a problem.
General incremental improvements to current systems. For example, at the moment Crit sometimes renders a page with frames or divisions very incompletely.
Here are some web pages related to the general topic of critical discussion and web annotation which you might like to look at.
(My lawyers told me to put this in. :-) ) I'd better explain here that I certainly wouldn't claim to be an expert myself in the kind of coding skills and the like needed for this sort of work - I simply find the whole idea of annotating the web very interesting. So if you do this project you'll be in some sense "on your own" (though obviously I'll try to keep up!). In short: this project is "for the brave". If you're still interested, do come and see me about any aspect of the project and all the various forms it could take!
This corner of the web maintained by
Iain Stewart
<ids@doc.ic.ac.uk>,
Department of Computing,
Imperial College,
London, UK
- add your own comments to this web page or any other for all to see!
(Only supported by some browsers, sometimes via an extension or the like.)
(If you're reading this from within IC DoC you can try Crit,
an earlier way to annotate the web.
)