Robert_Charlton - 7:21 pm on Jun 2, 2001 (gmt 0)

Just to play devil's advocate...

>>...if the results were unsatisfactory with a dB of 30 million, then the addition of more inaccurate data will only serve to increase the errors.<<

There would also be additional accurate data. To me what's important is whether the percentage of inaccurate data were increased. A larger database would at least pull niche-searches up into view.

>>I was a user for most of last year and been back for another look. I still get many results that I *know* are plain wrong.<<

GoTo also gives results that are just plain wrong. With its increased size, is WT better or worse than GoTo? I'm not sure that it's relevent any more to ask how current WT compares to the older WT.

Points of comparison for me would be sizes of the GoTo and WT databases, demographics of searchers, proportion of anomalous data between the two databases, and fine distinctions in data (like singular/plural and misspelling distinctions, no longer available on GoTo).