What are you looking for?

It’s harder when you don’t know, but can be more rewarding.

Over the weekend our astronomer didn’t get around to his normal procrastination, and went through a pile of scientific literature (both paper and electronic) that he’d had set aside for quite some time.  This was not a Literature Search, in which one seeks what relevant, or at least important, work has been done in a narrow field.  For instance, when he reports his observations on Planetary Nebulae he’ll find a recent survey of the field, to place his work in context; any specific papers on the particular objects he observed; any papers that support or disagree with his conclusions.  That’s relatively straightforward, using certain keywords in the search engines and following references in other papers.

This weekend’s literature was more general.  He marks articles and papers in general journals, those not limited to astronomy, that are. . . interesting.  It’s hard to say why, comprehensively.  They’re most often not in his field.  He finds recent results on prehistory done through DNA fascinating.  Dynamical systems, thermodynamics, the odd finding in vision or optics is marked for later perusal, but not everything in these fields, and not necessarily the most prominent ones.

The important point is that he would have come upon almost none of this by a directed search.  The relevant keywords  simply would not have occurred to him.  Indeed, he was unaware that many of these sub-fields even existed before stumbling upon the article or paper.

His habit of non-directed searching was probably formed early in his scientific career, when he was less adept at finding particular things (and search engines were less prevalent).  Leafing through all the pages of a paper journal, or going through whole shelves of books in a library, he would sometimes get distracted by something he was not looking for.  No doubt it delayed some of his own work.

For the main drawback of such general browsing is that it takes time.  Most of each journal does not find its way to his pile of interesting things, but he goes through it all.  And even when something is marked as promising, often it doesn’t turn out, when read closely.  And reading closely takes time.  Hence the pile is a prime candidate for procrastination.

We see no way to reliably automate the process.  Even if he were to train some AI program to pick out things like those he’s chosen before, it would probably miss something quite new and interesting.

Most of the time, when you seek information you know what you’re looking for and you want to get it done quickly.  But the most interesting results are the unexpected ones.