Abora Applications

The Abora server will provide a document server that focuses on the connections between content and their changes. A number of interesting and unusual capabilities arise from the design. These apply at a relatively low level, and it can be confusing to relate these new possibilities with applications that make full use of them.

The aim of this page is to highlight a number of possible applications that could gain from using an Abora server. Extensions to existing well known applications and services will be sketched out, together with potential new applications. The hope is to help answer the question, "but what can you do with these neat features?"

Extending An Existing Wiki Server

The family of Wiki web servers are one of the more interesting examples of websites supporting collaboration. They are based on the initially improbable idea of having open access to write and modify any of the text on a page, and create new pages. One can edit an exiting page, inserting your comments into someone else's original text, or even rewriting it. Wikis succeed due to their simplicity and freedom to comment and rework.

The simplest view is tha an Abora server should provide enough capabilities to allow one to write a stardard wiki web server that could serve web browsers.

The family of Wiki servers ranges from versions created in an evening from a few short Perl scripts, to pretty sophisticated implementations with JavaScript text editors, XML access and other features.

Starting from the relatively well known area of Wikis, what kind of additional features or changes to its abilities might an Abora backed implementation enable.


Focused Comments

One of the typical scenarios with a wiki page is that someone else has written a few screens of text, and you want to make a comment on a couple of sentences in the middle of it. You would either insert a comment line at this point and interrupt the original flow or you could add your comment at the end of the original text, but you would then have to describe the context of your comment by referring back to the line or two you are commenting on before you could make the actual comment.

With the abora server you could select the sentence you want to comment on, then make your comment, and under the covers your comment would be associated with the original sentence. Effectively there is a link between the sentence and your comment. It would then be up to the wiki server to render that comment in whatever way it liked; possibly an in-line link, or showing your comment in the margin beside the original sentence. Anything is possible with the rendering, but the critical thing is that you have tied your comment to the specific original text. This both saves you from having to describe the context due to association with original text, plus this relationship will survive feature editing and quoting by others. You could also associate a type with the comment, so in this case you may have thought that there was a mistake in the original and so your comment could be clarified as a correction. This provides a compact summary of the comment, and also allows the active reader to better filter additions to a page to help with numerous comments. This technique can also be used to build up a structured dialogue of pro/con arguments and references else where.

Filtered Notifications

Say one is interested in "acceptance tests". Perhaps there is already a wiki page with this title. Most Wikis would allow me to set a trigger such that I receive an email if that page is changed. This allows me to stay up to date with documents and arguments that I find interesting or would like to add as others extend it.

Extending this the Abora server would also be able to trigger notifications if anyone made a new link onto the acceptance tests page, or if someone copied-and-pasted any text of the acceptance tests page to another page. These triggers can be associated with specific parts of the page, so if you are interested in additional comments and modifications only to a certain paragraph of the page, then triggers will only be fired for that paragraph rather than for a change anywhere else on the page. You could also filter by type of comment, to only show agreements with a paragraph of the page.


Comparing different versions of the same wiki page would highlight changes made by users such as copying/moving/duplication/quoting changes as those kinds of things rather than just being forced down to primitive inserts or deletes operations.

The comments that have been made to the acceptance tests paragraph will stay connected to it, even if the acceptance tests page has grown too large and people end up splitting the original paper into 4 separate pages. The differencing stuff also functions at this point, and any links that had made into the original page still stay valid - as the original page is kept - plus you will be able to move forward through the revisions and track the page split up and where the paragraph you were interested in had moved off to. Comparing the original whole page and one of the split off pages you would accurately see the text that had been split off.

Beyond Text

Wikis are focused on text. The abora server, not implemented yet, but the udanax-gold server that is the inspiration for my abora work, also works for non text types. So I could imagine a wiki page which includes an image from an advertising campaign. I would now like to comment on the image to show how the viewers eyes moves around the page, and how they balanced the focus on part A of the image with B and C parts. Links can be made from the text description to the specific parts of the image. A suitable rendering could show that. Backfollowing links still works for images, so I could get all the comments on a specific part of the image, or could be informed if anyone modifies or linked to a specific part of the image. I could also create a new page focused on only a couple of sections of the image, so I could transclude those sections of the image onto my new page - so I see just the nose and left ear of the original face, but again the comparison between the full image and my two sections would highlight the similarities, plus anyone could backfollow from the sections to the original image, and comments on either side could be found if the comments were made on the sections. Another example could be a music score rather than an image, plus comments and quoting on sections of that musical score.