A while ago, I may have linked a post by Dan Lemire about how Google and Bing handle issues with searching for Kurt Godel vs Gödel vs Goedel. In teaching my grad class this semester, one of the issues we talked about in low-level processing is compound words. The title is an example – do we write quicksort without a space or with a space?
In this case, there is variation but “quicksort” seems more accepted. I definitely prefer it without a space – after all, it’s the quicksort not a quick sort. How about “mergesort” vs “merge sort”? And so on. Other common examples include “powerpoint” vs “power point”. Even if one form is more common, I don’t always know which to type into a search engine.
So I googled both versions and compared the top of the result. I found that they’re handled a bit differently.
The results make me curious about the underlying algorithms. At a high level, the big change is a slightly different ranking. The video search vertical is shifted around a little (though that wasn’t the case when I looked a couple of days ago). The NIST page and the Kent page swap places. So at the least, the two queries aren’t completely equivalent.
The biggest shock for me is the snippet for the NYU page, which is maybe less useful. The strangest thing is that it picks snippets with “quicksort” even in the results for “quick sort”, but yet it changes this particular snippet.
The variation in the “Something different” widget is interesting too – it seems to group sorting algorithms differently for the queries. But it becomes even more strange – the form “merge sort” is suggested for “quicksort” and the form “mergesort” is suggested for “quick sort”.
oh also, there’s a discrepancy in the number of pages, which may be problematic in some cases, but the top of the ranking is pretty similar regardless.