[seqfan] Re: Can we create a workable taxonomy for classifying all the various kinds of sequences in the OEIS? - not manually

[Frank Adams-watters]

There are a lot of properties of sequences that could be 
automatically recognized, except that they would almost 
all be conjectures.* Most of those conjectures could be 
verified fairly quickly, on an individual basis - but then you 
aren't gaining anything much on doing them manually 
in the first place. (We already have this problem with 
the nonn/sign distinction, but that's well established and 
so we only need to look at new  sequences.)

* One-way, anyhow. For things like monotonically 
increasing, one could automatically recognize almost 
all the sequences that do not have the property, with 
no errors; but those that appear to be increasing 
could only be conjectured.

Franklin T. Adams-Watters