[seqfan] Re: The Demichel errors, continued

[Marc LeBrun]

I may not be following this closely enough, but isn't the intent of all
these sequences to cover all possible compositions of certain elementary
functions up to a certain complexity?

If so, and if some entries submitted as part of this program have incorrect
values, shouldn't the priority be on simply computing and entering the
correct values, rather than trying to debug what went wrong with the old
values?  

Aren't the incorrect values kind of immaterial?  I suppose it may bear on
whether we fix the old A-number or create a new one.  (Or perhaps I simply
misunderstand this whole discussion).