GranularityOfScholarlyWriting

ThoughtStorms Wiki

This is some old writing. Today I think what I say on AIAndEducationExploded is more up-to-date and urgent. That the granularity / UI for knowledge which is about to dominate today, is ChatBots enabled by LanguageModels. These are so obviously better than almost anything else that I think they'll soon become the dominant media. So it's essential that people with something to say, learn to use them to say it.

Now back to what I said previously

Genres of Writing

One way of looking at it is like this. Academia currently has a number of traditional, institutionalized units of delivery of research :

  • books
  • papers
  • lectures
  • theses

Unless you can deliver your research packaged into one of these formats, it's unlikely to accepted or referenced. Furthermore these packages come with further constraints and requirements which made sense traditionally but may now be redundant.

When documents are hard and expensive to track down and acquire copies of, it makes sense that they contain a lot of context. So the author is obliged to include a fair amount of introduction which describes where she's coming from. And to accurately cite and guide the reader to further material which is relevant but beyond the scope of the document. If you go to effort to tracking down and acquiring a document which fails to do this responsibly, you have every right to feel cheated. Similarly, you have the right to expect that the author has done a requisite amount of checking for errors; and that the publisher has cross checked with expert referees.

But of course, as documents are bulked out this way, the burden on the reader increases. Each document is an even larger investment of time, because the reader needs to check through all this writing to make sure she's getting the benefit. Of course, she learns certain rules of thumb and learns to assess a document by it's abstract and citation list. There might even be a HandicapPrinciple at work, where the weight of the citation list signals the fitness of the author.

On the web things are different. Locating a document is very quick. And acquiring a copy is as simple as clicking the link. Hence context no longer really needs to be folded in to the document. The document can assume the reader has acquired most of the context on the navigational route she used to find it in the first place. Furthermore, citations can be instantly followed to set further context.

Context may also be supplied by third parties who produce their own guiding reading lists, curricula and indexes. There can now be a multiplicity of these, available in whichever flavour is convenient to the problem that the reader is researching. Multiple targetted indexes, perhaps available from a mentor or trusted other researcher are much more useful than an exhaustive citation list which attempts to be all things to all people, prepared by the author at the end of the paper. We can dispense with this ritual.

Now, of course, individual documents can be smaller, finer grained chunks which are skimmed quickly, and which give up their content easily.

More OnContext

The web publishing changes the economy of publishing - since the cost is very low it makes sens to publish not finished thoughts in the hope that other people critics would polish them quicker than lonely thinking. It does not mean that finished works are something from the past - in the ocean of low quality publications their value sticks appart.

Another thing to consider is the linking between publications. The scientific citation conventions are designed for printed material, much work is put into making it as efficient as possible, and very much of the strenght of the science model lies in it. On the web it is all obsolete. I think we shoule think what could replace it. – ZbigniewLukasiak

Counter Arguments

Short books with optional supplements : (OnBooks) http://jwikert.typepad.com/the_average_joe/2008/07/a-fan-of-short-books.html

See also :