Re: [Asrg] draft-irtf-asrg-criteria (was Re: request for review for a non FUSSP proposal)

--On 27 June 2009 14:37:52 +0200 Alessandro Vesely <vesely@tana.it> wrote:

> J.D. Falk wrote:
>>> However, I think an it could, and should, go beyond that. For
>>> example, why is it not in the scope of that document "to attempt to
>>> distinguish or justify any more detailed definition of [the term spam]"?
>>
>> Because attempting to define "spam" is the very best way to ensure that
>> a document is never finished.
>
> Finishing a document, by itself, is not a progress in the anti-spam
> endeavor, unless it contains useful knowledge for countering spam. I
> suspect that if we are not even able to agree on a definition of spam,
> there will never be much efficient anti-spam research by our group. And
> if people were to judge such research by the spam levels out there,
> they'd conclude it's all hot air.
>
> Could we start again, please?

The mistake below is to assume that "spam" is a single category that can be 
tidily defined. It can't.

The best one could do is to try to enumerate some types of spam. I'd 
suggest two, that look very different from each other: (a) submission of 
large quantities of electronic messages to an electronic messaging service 
with the deliberate aim of reducing the usability of that service, and (b) 
transmission of unsolicited communications by means of electronic mail for 
direct marketing purposes.

These are very different - the content is irrelevant in the first, but 
highly relevant in the second. The threat to the system is the motivation 
of the first, but would be harmful to the sender in the second. A recipient 
might welcome the second, but is unlikely to welcome the first.

One might wish to add other types of spam (phishing, viruses, chain 
letters...) but it's unlikely that even the longest list constructed today 
will cover all future types of spam.

> Let me try the following variation on UBE, in order to see if we can
> classify the objections against adopting it as a working definition.
>
>   *Spam* is a message or a class of messages composed automatically,
>   possibly using templates and databases of names or words, using a
>   list of destination mailbox addresses, and failing to meet both the
>   following conditions:
>   * There is a record or evidence whatsoever, even implied, that the
>     recipient has registered, subscribed, or otherwise solicited the
>     message; and
>   * it is obvious from the message content, including the headers,
>     who is in charge of controlling the processing, and how the
>     recipients can amend or delete their addresses from the list (as
>     well as any other part of their personally identifiable data from
>     any database used to compose the message), where "obvious" means
>     the relevant entity is indicated, exists, is normally reachable by
>     the recipient, and is effectual to the purpose it is referred for.
> _______________________________________________
> Asrg mailing list
> Asrg@irtf.org
> http://www.irtf.org/mailman/listinfo/asrg

-- 
Ian Eiloart
IT Services, University of Sussex
01273-873148 x3148
For new support requests, see http://www.sussex.ac.uk/its/help/