"Message has a suspicious header"

Pete Batard pete at akeo.ie
Wed Jan 25 07:53:38 EST 2012


Hi David, thanks for the reply.

On 2012.01.25 05:28, David Woodhouse wrote:
> On Mon, 2012-01-23 at 11:15 +0000, Pete Batard wrote:
>> Maybe something to look into as well.
>
> That was the 'thread hijacking' filter — your message had the
> References: and/or In-Reply-To: headers which marked it as a reply to a
> previous message, but didn't have 'Re:' in the subject.
>
> This usually happens when some lazy person hits 'reply' to an existing
> thread and changes the subject (hence 'thread hijacking', because it
> injects a completely separate topic in the middle of an existing
> thread).

OK, then I guess I will have a problem with that for the following reasons:

1. The error is too vague. At least the mailing software should mention 
something about it not accepting "thread hijacking" so that people 
confronted to the issue are not mislead to think that it's just poorly 
configured mailing-list software and give up on trying to post their 
message.

2. Since I tend to use thread-hijacking on regular basis to reframe 
existing discussions, I can tell that none of the other mailing lists I 
participated so far seem to implement such a feature (IMO, with good 
reason). Ergo, I suspect at least some subscribers of this mailing list 
will be taken aback by this unexpected behaviour.

3. The choice of implementing it should really be up to the list 
administrator(s) (which it possibly already is, in which case the points 
I make will be aimed at the libusbx admins, though my guess it is not) 
and not the hosting provider. The reason for this is, if I was one of 
your customers, the first thing I'd ask would be for a way to lift this 
restriction, because I find it unnecessary restrictive, penalizing for 
users, and likely to prevent very relevant data from ending up on the list.

4. While a subset of people may use thread-hijacking because they are 
"lazy", others may see very legitimate use for it such as reframing a 
debate or going on a tangent *while* wanting to ensure that readers of 
the new thread can find out very precisely where the tangent originated 
from. I can very much see cases with thread-hijacking forbidden where 
subscribers reading the new thread might wonder whether it spawned from 
something that was previously discussed, which may be very relevant, but 
are left with no possibility of figuring where that was (especially if 
the original thread was lengthy). IMO, mailing list software that allows 
thread-hijacking and displays the new thread with a different subject 
(and possibly no quotation of previous post) right in the middle of the 
existing one, are doing the right thing from a user's perspective.

5. The first duty when implementing a mailing-list should be with 
fostering as much contribution as possible, rather than exclude it. This 
means, the less you restrict a subscriber's ability to post, the better. 
Else, one might find that some very insightful posts were prevented for 
pure arbitrary restrictions. Again, I'd very much prefer a mailing list 
where thread-hijacking is allowed, if it gets someone to post very 
relevant and valuable information, than find that such a post never 
occurred, because the contributor gave up after getting a "suspicious 
header" message.

6. Finally, if I wanted to take a jab at the "lazy" contributor quip, 
and if the restriction is implemented at the software level, I'd say 
that, from an external perspective, mailing-list software that prevents 
thread-hijacking very much seems to be due to "lazy" developers, who may 
have found that allowing such a feature would result in additional work 
for them, and therefore chose to disable it. At the very least, if the 
goal is to prevent new threads that may or may not be relevant to an 
existing discussion from appearing in the middle of it, since human 
processing of header + subject can easily figure this out, it must be 
possible to automate such a task. Therefore the one reason not to 
include automation that allows thread-hijacking may very well be because 
someone didn't want to spend time implementing it...

I hope I'm not being too harsh here, but all in all I see a no 
thread-hijacking restriction being detrimental to the libusbx mailing 
list, and not just for "lazy" people. Thus I would kindly ask for you to 
consider lifting it, or if this is controlled at an individual list 
level, for the admins of libusbx to do so.

Regards,

/Pete



More information about the libusbx mailing list