On 23.12.2006 15:16, Trans wrote:
> Trans wrote:
>> Robert Dober wrote:

>> In a post script, Jeremy McAnally had written,
>>
>> "It would also be mildly entertaining to have an auto-answer FAQ bot
>> that parsed the language of a message and if it could decently figure
>> out what they're saying, post an answer pointing  to a (currently
>> non-existent) Ruby FAQ..."
>>
>> I realize that automated replies don't have the niceties of human
>> interaction, but I think something like this could go a long way toward
>> improving list usage. Rather then automated answers just against a FAQ,
>> automate them against the mailing list itself via the archives. An
>> automated answer could then provide a list of links to related old
>> posts. Even as an experienced user I would find this kind of
>> representation helpful! Also, with these automated replies, experienced
>> users who are already setting up "weak" list filters, could create
>> stronger ones based on whether a new topic received an automated reply
>> or not, and what the automated reply concluded about it. (In fact I
>> imagine a Bayesian filter would be part of the bots implementation.) I
>> think it would be better for us to try something like this, and see how
>> it fairs, before taking the leap to two lists.
>>
>> What do others think? And also, is anyone up to the challenge?
> 
> I'm curious. Did anyone recieve/read this? I'm surprised no one has
> replied to it, even if it's just to say "dumb idea".

I received and read it - but since I could not provide good feedback... 
  It sounds like a good idea but I wonder how it will work in practice. 
  This might not be a problem in a newsgroup but for a mailing list this 
automated reply could easily trigger spam filters etc.  Generally I have 
a tendency against stuff that automatically increases traffic but it 
still might well be worth a try.  Dunno...

Regards

	robert