LSTOWN-L Archives

LISTSERV List Owners' Forum

LSTOWN-L

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Topic: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Eric Thomas <[log in to unmask]>
Fri, 6 Jan 1995 16:34:22 +0100
text/plain (37 lines)
On Fri, 6 Jan 1995 07:27:15 -0600 Steve Hirby
<[log in to unmask]> said:
 
>> 3. don't blame LISTSERV's "auto-delete" function for something
>>    which is totally beyond its control.
>
>     Re: 3,  Seems to  me that  there is  a legitimate  question whether
>LISTSERV,  in   implementing  auto-delete,  should  knowingly   rely  on
>information that is unreliable. The  intention of auto-delete is to save
>listowners work; deleting in error based on reliable information has the
>opposite effect.
 
That's  all fine  and well,  but  unfortunately LISTSERV  hasn't got  the
beginning of a possibility to  classify delivery reports as "reliable" vs
"not reliable". We human beings may know from painful experience that Joe
Professor has a mail system that  lies, but how would LISTSERV know that?
This particular  example was  about name  servers. The  DNS distinguishes
between "authoritative" and "non-authoritative"  replies. The former mean
"I know for a fact that..." and the latter mean "I believe that..., for a
final answer you'll need to ask that server". For any given domain, there
is a  very small amount of  servers with authority to  make authoritative
statements.  They  are,  by  definition,  the  most  reliable  source  of
information about  a domain. Where  are you supposed  to turn to  if they
start lying? Maybe in  10 years the next version of  LISTSERV will make a
phone call to the  human person in charge of the name  server and say "Hi
Joe, this is LISTSERV at  UBVM.CC.BUFFALO.EDU here. Say, your name server
at 131.172.2.1 claims that there is no such host as XXX.YYY.EDU. However,
my  records indicate  a  plausible probability  that  this machine  might
indeed exist,  based on an  analysis of  common patterns for  your domain
adjusted  by historical  data.  Could  you please  find  our whether  the
machine exists or not, and call me at XXXXX when you have an answer? I'll
take care of removing the users if that is the case. Ah, and don't forget
to fix the name server if it is in error". But today computers don't know
how to talk to human beings.
 
  Eric

ATOM RSS1 RSS2