In a fascinating recent paper a team of Google-based researchers have shown that the way academic literature is disseminated has gone through dramatic changes over the last two decades. 'Elite' journals in most fields have shrunk in relative significance, as both highly cited papers, and citations in general, are increasingly dispersed across a wide variety of outlets. Whatever other factors are driving this process, the biggest surely is the rise of internet databases as a research tool, with Google's own 'Scholar' search engine first in that class. Papers no longer need to be in the major journals that people read, or even that libraries subscribe to, to be discoverable. This could perhaps be seen as part of a broader societal trend, driven by the internet, away from monolithic and powerful broadcasters, which is sometimes referred to as the long tail.
There does seem to be a mismatch between this drift away from elite journals and the pressure I feel under as a postdoc bouncing around the world on short-term contracts, which is to publish in eye-catching places to increase my chances of impressing hiring committees and attaining something more permanent. Could having a paper in a famous and hard-to-get-into venue become more about advertising oneself in the labour market than getting one's work read? In theory, perhaps, everyone's incentives should align, as the most selective outlets accept only the most significant papers, thus sending a credible signal about the quality of their authors. But the waning of elite journals, despite the academic market being tougher than ever, seems to belie this. Perhaps attitudes simply take time to change, and the current obsession with venues will fade.
Setting this aside, it is clear that Google Scholar has changed how research is done. Last year I launched this blog to have a play with the 'Updates' tool for the Google Scholar database, which uses one's own research as a 'search term' to discover pertinent new research, wherever it is released (often it is 'published' nowhere more prestigious than the homepage of the author). My review of the tool is given here; in short it is remarkably good for discovering papers relevant to your 'old' work, but obviously useless for discovering papers relevant to your ongoing work that is not yet archived by Google. As I shifted jobs from Australian National University to Aarhus University in Denmark I have been furiously learning new stuff to catch up with the interests of my new group, which is great fun but for which the Updates tool, and my blog about it, would have been a distraction.
Now that I have got my bearings a bit in Aarhus (although there always so much more to learn!) I thought I could bring back this blog with a slightly different focus. Instead of looking at the new research recommended to me, which is sometimes very interesting but often a bit obscure outside a very narrow community, and so not necessarily great 'blog material', I would briefly review some 'classic papers' useful for a slightly wider world of theoretical computer scientists and logicians. But how do I choose the classics in question? To keep consistency with my blog's theme to date, and make sure I choose papers relevant to my corner of the research world, I will use the Updates tool again, looking at the papers most cited by those new papers recommended to me. Having done the tedious task of compiling 1133 citations from 34 papers, and hence divined the papers considered most relevant by the authors that are considered most relevant to me by Google, I will post my first review next Monday.
