Lawrence D'Oliveiro <[EMAIL PROTECTED]> wrote: > In article <[EMAIL PROTECTED]>, > [EMAIL PROTECTED] (Alex Martelli) wrote: > > >... let's try some google searches and see the number of million hits...: > > But how reliable are those estimates of numbers of hits, anyway? More > than once I've got a page showing something like "Results 1 - 10 of > about 36 hits", only to find that there were no more pages after the > second one. If it could get estimates so wrong with such small numbers, > how can you trust the large ones?
If your experience in the matter dates from a few months ago (and most of us may have done Google searches for years, contributing to form our impressions) it may perhaps not be all that applicable any more, according to rumors from early this year (e.g. <http://www.eweek.com/article2/0,1895,1916406,00.asp>) and blogging notes by Matt Cutts at <http://www.mattcutts.com/blog/bigdaddy/> . If you can give examples of a query which estimates the (small) number of results very badly, net of course of Duncan Booth's observation regarding Google's attempts to filter out near-duplicates, I'll be glad to look further into the matter (my group is Production Systems, just about as far from such issues as you can imagine -- rather, we write software to keep our clusters, networks &c smoothly running, monitor them, and so forth -- but, of course, I do have friends over in Search Quality and related groups... fortunately, Google has so far managed to avoid much of the "silos syndrome"... so I'll be glad to check it out, but I do need specific examples!-). Alex -- http://mail.python.org/mailman/listinfo/python-list
