Anurag Acharya is Google Scholar’s Principal Engineer. After Dr. Pauly’s article Web of Science vs. Google Scholar was debated on blogs/ listservs yesterday, I sent Anurag three questions:
1. Is the Google Scholar index the same searchable index, regardless of location?
Anurag: It is the same index and the same ranking [regardless of location].
My comments: To have any utility as a citation tracker, Google Scholar’s results must be reproducible. At the very least, that means reliable, stable content – not differing results for each search. If results were to differ across time zones and geographies, Scholar would only ever provide vague notions of citedness scores.
A priori – the index must be stable, complete and transparent to searchers!
**********
2. Have you tested Google Scholar’s performance from various geographic points globally? Are you aware of any country to country variations due to specific network issues [preventing it from performing similarly]?
Anurag: I don’t know of any filtering at the network level. The index at our end is the same.
My comments: This is very good news because regular Google has variants: Google.fr, google.de, google.it. For nations like Google China some results are blocked. Anurag seems to say that regional differences do not exist in Scholar.
**********</center
3. How often does GS change? daily? monthly?
Anurag: We cannot share update information, but our long-term goal is to update every day.
My comments: Daily updates (even hourly) should not be difficult. However, this response is classic Google, showing extreme reticence. The PubMed index changes daily; WoS changes weekly. Librarians know how to cope with differences in coverage between databases, and can modify strategies based on what they know about a database. Without this basic information, librarians cannot recommend Google Scholar (except for the most basic kinds of browsing; known-items).
Surely, this is not only a librarian thing though. I’ll blog a little more about how librarians cope with linguistic/semantic challenges in databases (what programmers call normalization – which is not the same thing as authority control in an index).


The

In the