Wednesday, 12 October 2011

the trouble with indexing ...

One of the problems I find with all note structuring applications is indexing and categorisation. Most of them have moved away from the strict hierarchical categorisation model (if it's this it sits in this bucket, and if it's that that bucket) by using tags but even so you do tend to end up with a pile of thematic buckets.

This is absolutely fine when collecting material with a purpose - I'm going to write a paper on X - but not so fine when collecting ideas - what I describe as post-its on a wall.

Post-its on a wall is a technique I've used a lot. Write down an idea or concept on a post-it. Stick it on a white board Write down another on another post-it. If your'e clever you can use tricks like using different colours if the idea or concept comes from somewhere else. Draw a line between the two post-its describing the relationship between the two. Do it again. Draw a line. And so on.

You end up with what I used to call a connectedness diagram, but is really an informal representation of linked data. It's a technique I find really useful for understanding and organising material. It's also not a new technique, I used it, with sheets of butcher's paper and coloured pencils at the end of the seventies when revising for my finals and finding links and references across and between modules (We can say this about foraging behaviour in prosimians because their visual systems have this characteristics, and the environment in which they live lacks distinct seasons, etc)

I havn't really seen an alternative to the post it technique - mind mapping tools like freemind for some reason seem to lack the flexibility required, and what one wants to do is to arrange and diagram the relations between objects.

One alternative which does seem to do the job well is LORE - the literature object re-use and exchange tool developed as part of the Aus-e-lit project.

I'm going to guess that conceptually it started out as an annotation tool to allow the linking of notes and material together, but crucially what it allows is for you to develop and diagrams sets of links between objects and share them with collaborators (or the whole world should you want to) but also to creatively organise material.

Such a model also delivers what I call 'active curation'. Texts in other languages can often have ambiguities in translation, especially as when the language is something like Middle English.

One could take two versions of the same text, link the two and compare the readings and perhaps reference similar less ambiguous bits of text in other documents, etc, etc, to show why a particular interpretation should be preferred over another.

And of course this is not just for Middle English, the same approach could be taken to analysing witness statements, when investigating criminal cases, especially where we are talking about cases such as fraud or other forms of financial malfeasance which can be extremely difficult to prove but where the case is built of little facts and inconsistencies ...

No comments: