Lists Home |
Date Index |
On Monday 10 March 2003 11:50 am, Alessandro Triglia wrote:
> How expensive would it be for the authors of this paper to create a new
> set of test documents vaguely similar (in terms of some metrics) to the
> "unpublishable" ones, and then re-run the tests and publish the new
> results? Actually, the new test data wouldn't need to resemble the
> current test data - how much it does is up to the authors. But the new
> results would be more useful and interesting than the current ones.
Seems like a perfect job for a perl script...