[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Bug#684128: down the memory hole

On Thu, 4 Apr 2013 13:09:30 +0200
"Didier 'OdyX' Raboud" <odyx@debian.org> wrote:

>> If it were thought that the criticism was unfair, or inaccurate, then
>> it could be allowed to remain in place, so that other people might
>> judge its lack of merit for themselves.
>> In the case of bug #684128, post #108, however, the fact that the
>> offending message was promptly vaporized* (as will be this one also),
>> of course suggests that the opposite is true.
> Are you talking about that mail ?
> http://article.gmane.org/gmane.linux.debian.devel.bugs.general/1040315

Yes. Or here:


> If that's the case, I'm not surprised that it got flagged as spam and
> removed from the bug tracking system (as that's what I suppose
> happened): I needed to read until the end of it to notice that this
> long mail was vaguely related to the bug at hand: both the subject and
> the content look like spam to me. That said, I don't find that this
> mail is unfair criticism, just that it is not sufficiently related to
> the bugreport (more in shape than in content) to be unambiguously
> non-spam.

When read in the context of that particular bug report, I don't see how
it could possibly be any more relevant, since it refers directly to the
discussion above.

I would be surprised to discover that there are spam filters in
operation whose criterion for accepting an email is that it be
"unambiguously non-spam", since that determination would seem to be well
beyond the competence and authority of automatic systems to decide.
Instead, they look for the usual evidence of spam: HTML, embedded
images, keyword lists, links to spamvertising websites, etc. I fail to
see how the message in question would be flagged on those grounds, or
how it would differ in that respect from millions of other messages sent
to the bugtracking system which are not so categorized.

It's certainly the first time any message I've sent to the Debian BTS
has been dropped, especially after receiving a confirmation message.

> Dropping obvious spam from public archives has nothing to do with
> "Hystorical Revisionism" or whatever else: don't assume malice here: I
> think that in this case either the automatic filters or the human
> triagers have slightly overlooked your mail.

I suppose it's possible, but I doubt it, for several reasons:

1 - the fact that it was archived on the "debian-boot" list, as above,
suggests that it was accepted by whatever automatic filters actually
looked at it.

2 - as I mentioned in my previous message, I received a confirmation
email from the bugtracking system. I very much doubt that those are sent
for incoming email which has been determined to be spam.

3 - one would imagine that a spam filter attached to a bug tracking
system would be sophisticated enough to know that messages sent to a
particular item in their database, from the same email address as the
original bug report, are highly unlikely to be spam.

4 - the message in question did, in fact, appear at the bottom of the
relevant bug page for a day or two (as post #108), and then mysteriously
vanished. It seems unlikely that the automatic spam filters would wait
around for as much as 48 hours before examining a message which they had
already allowed to be posted, and then retroactively remove it.

5 - you mention human triagers; given the volume of mail that the Debian
bugtracking system must receive, I cannot believe that more than a tiny
fraction of it is sent to human spam reviewers. Presumably they only
examine those messages for which the automatic systems have found enough
spam evidence to be suspicious, but not enough to meet the threshold for
automatic rejection; the decision must happen automatically in the vast
majority of cases. For the reasons mentioned above, I really doubt that
the message in question fell into the "suspected, but not conclusively
proven" category.

> So please, next time something puzzles you similarly, ask for
> clarification in a neutral way instead of publicly accusing "Debian of
> operating Historical Revisionism", which is incredibly rude.

It's only rude if it turns out not to be true. If I'm wrong, I will of
course retract my comments. But I don't think that I am.

-- Ian Bruce

Reply to: