I’ve been testing the English version of the major Russian search-engine Yandex for the last month. I’ve found it to be a useful alternative search engine, one with a surprisingly wide range and depth. In terms of the reach of its plain Web search I’d say it’s comparable with DuckDuckGo, though it can’t match the excellent relevancy-ranking of DuckDuckGo’s Image Search.

Other notable points are:

* it has ‘instant’ speed, as fast as Google Search.
* filters and search modifiers that are similar to Google.
* the user interface is easy to use and doesn’t annoy.
* there’s a pleasing lack of nag-nag-nag about EU privacy and cookies laws.
* it doesn’t throw a tantrum if you make more than a half-dozen site: searches.

For discovery of Web items from the “last 24 hours” / “last two weeks” Yandex is much less spammy than Google, making it much more useful for the few journalists and bloggers who actively go looking for timely new content. That’s true even when one runs Google Search with an extremely well-populated installation of Google Hit Hider by Domain (regrettably this Firefox/Greasemonkey addon, which automatically blanks spammy URLs in your search results in Google/Bing/DuckDuckGo, doesn’t yet support Yandex). A “last 24 hours” search in Yandex is also refreshingly free of Google’s “yeah, it’s actually from 2009, but we’re showing it because we crawled it again in the last 24 hours…” results.

When a 24-hour search is run on a timely in-the-news topic name or keyword, Yandex effectively becomes a very useful alternative to Google News or Bing News. It also accurately detects user-location (UK, city), without a sign-in, and correctly and deftly skews the news sources to the appropriate nation. Although there’s a caveat… on many keywords the first 10 results will often seem to skew toward the more left-leaning news outlets. I’d suspect the algorithm is just being unduly influenced by the clicks of a minority of heavy Yandex users, who will probably skew strongly toward people inclined to click on anti-western news headlines? I half-expected that the English version of Russia Today (RT) would rank highly, in that scenario, but it doesn’t seem to rank at all in such 24-hour search results. Not in the UK, anyway.

So Yandex’s Web search is impressive, but a little quirky in coverage. The largest gaps in Yandex’s coverage are due to sniffy sites that only allow the Googlebot to index their files. When Yandex is allowed to get in to such sites, the indexing and title identification of PDFs and articles seems directly comparable to Google Search.

I looked hard for browser add-ons for Yandex, but there are hardly any in English. A couple are worthy, though, and a search veteran working with Yandex would ideally want to…

1. Stop Yandex’s URL obfuscation in search results, with the browser add-on Google Search Link Fix (also works with Yandex).

2. Get a double-column layout for search results, if working on a desktop PC. There’s no version of GoogleMonkeyR that works with Yandex, but the Firefox add-on Stylish has a double-column UserStyle for Yandex that works fine.

Interestingly Yandex offers a sort-of Custom Search Engine. But your Yandex CSE and its topic have to pass moderation, and you have to admin the CSE from a single fixed IP. That’s no good at all for those who have a western-style ISP which dynamically assigns an IP address each session. I presume that Russian Internet users must be forced to have a fixed IP address.

Update, April 2016: a new and utterly dim-witted spelling auto-correction has been introduced. For instance “modal” is corrected to “model”, with no way of forcing the original term. eg: “modal” “wind turbine tests”. At a stroke, this makes Yandex unusable as an academic search engine.