Friday 12 December 2008

Wish list

It's getting close to Christmas, and for many this means time to make a wish list. Now, I'm not expecting anybody to go shopping for me for the kind of things I've been wishing for; they're too ambitious and far-reaching. OTOH, this wouldn't just be for me - it should benefit researchers everywhere, in the future.

Item #1 on my list is for e-Santa, or Google, or the open-source community, or scholars themselves, to pave the way for a better means to identify authors in scholarly publications. Right now, most journals still follow the tradition of showing just one or two initials and surname for each author. Generally this is supplemented with info on institutional affiliation, and lately with an email address as well. However, journal search engines such as Google Scholar only key on the name; while they do what they can with this limited info, it is really not adequate to find all and only works by a specific person. The worst problem is with very common surnames, whether Smith, Scott, Wang, or Ramesh. Authors who don't have or don't go by any middle initial make the odds that much worse.

The easy first step would be for all scholarly journals to change editorial policy and begin listing full first names for authors. Citation style sheets will need to be brought up to date to conform.

But ultimately there will still be duplicate full names. I guess I can't wish for civilization to reform to ensure every child born from now on is given a name unique throughout history (ha!), and then waiting for a generation or two to age out the current cohort of duplicate names - that's not going to help me in this lifetime.

So I'll settle for somebody coming up with a good strong proposal to assign unique identifying strings to each author who gets published, much as we now do for for publishers and books with ISBN and ISSN numbers, and for articles with DOI numbers.

I don't think it would be too difficult to design such a scheme. We'd need a central worldwide registrar to track the unique number strings, just as Arxiv.org does now for DOIs. We'd need a way to ensure that an ID continues to point to a person even as they move through their career (and between jobs and cities, as young scholars seem so prone to do). Other confounding variables this could help overcome include uneven use of one vs. two initials, people with non-alpha characters in their surname including spaces, hyphens and apostrophes (van der Waal, O'Brien, Levi-Strauss) and name surname changes after marriage (or divorce).

Obviously I'm not the first to notice this problem, so maybe there's hope for a movement to build around this kind of idea. Here's one existing proposal and here's another.

I hope they get their wish.

Happy holidays to everyone.

No comments: