Re: Clax86 and the New SPAM prevention measures

From: Siberian Husky (alaskanmalamute_at_hotmail.com)
Date: 06/27/04 

Date: Sun, 27 Jun 2004 18:21:08 +0000 (UTC)

Dr John Stockton <spam@merlyn.demon.co.uk> wrote in message news:<oUC25qI$Sf3AFw2n@merlyn.demon.co.uk>...

Discussions with you seem to become longer and longer. Before Charles
kills our discussion, I would suggest you ask your questions in
another newsgroup, somewhat for general questions on Usenet and how it
works (those in news.* hierarchy).

I advise us to stop continuing a long discussion on moderation in CLAX
because we are not addressing anything on x86 assembly languages nor
issues directly concerning CLAX. In the newsgroups I myself moderate,
I would not allow such a discussion to continue for such a long time.

That is, this message is the last one I respond to yours in this
thread unless I feel you are still directly addressing CLAX issues.

> JRS: In article <4cf8dc14.0406252251.2d0fc383@posting.google.com>, seen
> in news:comp.lang.asm.x86, Siberian Husky <alaskanmalamute@hotmail.com>
> posted at Sat, 26 Jun 2004 07:02:46 :
> >You are right that moderators.isc.org does not limit what it might do.
> >
> >But then when this address relays a submitted message to one of the
> >sites actually keeping track of the moderator's submission address,
> >these sites *CAN* do something whether you like it or not. For
> >instance, the ucsd.edu site is known to filter virus.
>
> It is a moderator's duty to receive all mail that might be approvable;
> that may mean that the moderator must make alternative arrangements for
> receiving mail or turn the job over to someone else. Spam-filtering is
> acceptable, if sufficiently reliable that it kills a smaller proportion
> of good messages than human error would otherwise have killed.

There are factors beside a moderator's control when you insist he
receives all mail that might be approvable. Charles has already
posted how many sites (and each of them could have done something to
your message) your message traveled through before reaching him.

The only way you can ensure nobody else messes around your message
before reaching him is you directly send your message to the moderator
submission address -- but the ISP Charles uses can still decide it is
a spam and leads it to /dev/null, or Charles' ISP can determine your
domain is a spam machine and ignore any emails from your domain.

> >Back to Chuck's original issue about AOL, Yahoo and Microsoft (MSN)
> >proposed ideas to curb spams. Suppose there are 6 such sites. If one
> >of them decides to adopt the spam-curbing mechanism (so a munged email
> >address is not accepted if it is not a valid domain, say), then it
> >means your submission to CLAX has a 1/6 chance of being hit with the
> >issue raised by Chuck.
>
> MY submissions should go nowhere near any of those services. If someone
> wants to post, then he needs to choose a service that does not stop him
> doing so. A munged E-mail address should be an address that one has
> permission to use for the purpose.

If you submit your message to a moderated newsgroup (not just CLAX),
it will go through one of these 6 or 7 sites. It is not up to you and
it is not up to Charles.

If you insist your message not go anywhere near these services (like
ucsd.edu), you have a choice -- you post your message in alt.lang.asm,
the unmoderated newsgroup. This is up to you and completely up to
you. Whether your message is buried under the noise or not is also
completely up to you.

> >Usenet is a decentralized medium. There is no such thing called
> >"Usenet-wide device". The moderator of a moderated newsgroup is the
> >closest approximation to that idea.
>
> It might be the closest existing approximation, apart from
> moderators.isc.org and other similar addresses (IIRC, m.i.o is not the
> only such; I believe there is a uk-based service which redirects
> directly for news:uk.* groups and to m.i.o for those that it does not
> recognise); but clearly it does not have to be.

Newsgroups in uk.* hierarchy is completely another issue not addressed
here.

> >Not really. You are right that you can compose your article and put
> >the time string in your message body (like 14:01 BST) then click
> >"Send" button exactly 1 second before that time instant.
>
> I do not have a "Send" button.

It does not matter -- there is something equivalent. In "tin" on
shell accounts it is a press of the key "p" (for post). In Google it
is the web button "Post message".

> >But that is *your* time and *your* zone. The news server your news
> >reader program connects to is free to accept the Date: header entry
> >your computer and/or your newsreader program put in, but the server
> >can also ignore your Date: information and put the time stamp
> >generated by the server.
>
> The news that I post to non-moderated groups leaves this room, as I
> wrote, by NNTP. It then goes to the Demon service, which peers it in
> various directions. Demon has no authority to change such a header line
> in stuff posted by its customers (it does have the power to do so, of
> course). Likewise, news.maxwell.syr.edu for example, which is in the
> Path of your article as received here, has no authority to change such
> header lines.

There could be a reason why Demon wants to change your Date:
information in your post to an unmoderated newsgroup. I can think of
two.
(1) The clock on your computer is incorrectly set. For instance,
    for whatever reasons it says 1996 (say, the battery on your
    motherboard no longer works but you don't know).
(2) For some reasons demon.co.uk temporarily can not keep
    processing Usenet posts and your submission goes into a
    queue waiting to drain. It takes hours to clear.

> > Remember at this moment nothing is involved
> >with moderation yet.
> >
> >Then when the server realizes your message is submitted to a moderated
> >newsgroup, so the server will modify your message body somewhat so it
> >can be mailed to the moderator as an email (for instance, try to put a
> >"From " in the beginning of a line in a Usenet message.
>
> That is, in the words of WSC, a terminological inexactitude. Evidently,
> while you may understand how your system works, you do not understand
> other legitimate ways of working. I use Turnpike software, which knows
> the moderation status of each newsgroup. For moderated groups, Turnpike
> mails m.i.o; for others, it uses NNTP.

Turnpike software "knows" the moderation status of each newsgroup
because the news server it connects to tells it so. If the news
server is clueless, then your Turnpike software is also clueless.

For instance, now you compose a message and crosspost it to
a.b.c.unmoderated and s.t.u.moderated_in_Russia while your news server
doesn't carry s.t.u.moderated_in_Russia. The result is your Turnpike
software will treat your message as unmoderated and use NNTP, and in
Russia nobody reads your message.

> > I think it
> >will be added with a ">" in emails. Not sure about Usenet). At this
> >stage again, the mailing program at the news server is free to use the
> >Date: information passed from the Usenet agent, or use its own Date:
> >information. That could mean a few seconds added to the Date:, or
> >even a few minutes.
>
> However, I'm pretty sure that I can myself locally mark a newsgroup as
> moderated or otherwise; and that, if I mark clax as not moderated, posts
> to clax from here will go by NNTP to Demon, who will mail them to m.i.o,
> who will send them on to crayne.org. Demon will not alter any header
> line (except to add what is needed for mail, and to deal with any not
> allowed in mail); likewise, m.i.o will alter nothing except as is
> required for their function.

False. When you submit your message to demon.co.uk, the news server
there finds CLAX is moderated and turn it into an email sent to
Charles. Whether your own computer thinks CLAX is moderated or not is
irrelevant. Your computer talks to news server at demon via NNTP
protocol (through port 119, IIRC). Whether CLAX is locally moderated
or not is irrelevant.

> >So, to sum up, it is not *your* time and *your* zone. It is the news
> >server time and its zone.
>
> Rubbish. You do not understand how posting of News by a dial-up
> Internet host _can_ be done. For example, my dial-up service has an
> access number for receiving calls from overseas (it has better
> arrangements for those with UK land-lines) : if I take my overseas and
> configure it for the foreign locality, then Turnpike will generate the
> appropriate foreign time in the Date header, and that it what Demon will
> pass on.

I read the O'Reilly book "Managing Usenet" by Henry Spencer and David
Lawrence several years ago. I agree perhaps I forgot most of it, and
what I learned was not updated. However, if you have also read that
book (or anyone about managing Usenet), it is not too late for you to
call my words "rubbish".

> > If the news server forgets to adjust its
> >time based on Daylight Saving Time (a scheme in the U.S.), then you
> >can probably expect another discrepancy of one hour.
>
> Well, we have Summer Time here. The time will be what my PC thinks the
> time is; it resynchronises whenever it dials up, Windows handles the
> change to Summer Time on the correct date (at the wrong time; but I
> never use the computer while it will be wrong), and I can correct the PC
> clock as needed (it's about half a minute slow at present ... and now it
> is within a couple of seconds of correct.

Perhaps. But Usenet is not designed for you alone, to assume
everyone's computer has a clock as accurate as yours.

> >Repeat after me. Usenet is an asynchronous medium inherent in its
> >flooding algorithm. You don't make any point to insist what
> >information is put in Date: (and whether the moderator can introduce a
> >few minutes or a few hours).
>
> Rubbish. That header line is generated when the article is released for
> transmission at the next opportunity, and is good data which should be
> preserved. I participate in an astronomy group, in which sometimes
> people report observations as they happen. If an observer chooses to
> post "ISS is overhead ... NOW!", then the originating time should be
> within a couple of seconds of the event, and is real-world data.

And you should know that a message you write in U.K. might spend 3
seconds when a netter in Los Angeles read it; but it can also happen
that it takes 3 days before your message reaches another Usenet reader
in Dublin or Liverpool. And perhaps it never reaches a site. There
is a reason to call Usenet "asynchronous" by its flooding algorithm.

> If a moderator wishes to record when he acted, he can do so as a comment
> in the Header Line of Approval, or he can add an X- line. He should not
> destroy data.

Adding an "X-" line is itself "changing" data.

> This is part of the header of an article I wrote just before 10 pm UK
> time on Thursday, as transmitted :
>
> Message-ID: <Gsr9nQHeC02AFwJS@merlyn.demon.co.uk>
> Date: Thu, 24 Jun 2004 21:59:10 +0100
> To: comp-lang-asm-x86@moderators.isc.org
> From: Dr John Stockton <jrs@merlyn.demon.co.uk>
> Reply-To: Dr John Stockton <reply0406@merlyn.demon.co.uk>
> Sender: Dr John Stockton <spam@merlyn.demon.co.uk>
> Newsgroups: comp.lang.asm.x86
>
> These are the corresponding lines of the article received in News :
>
> Message-ID: <Gsr9nQHeC02AFwJS@merlyn.demon.co.uk>
> Date: Thu, 24 Jun 2004 23:10:49 +0000 (UTC)
>
> From: Dr John Stockton <jrs@merlyn.demon.co.uk>
> Reply-To: Dr John Stockton <reply0406@merlyn.demon.co.uk>
> Sender: clax86@wotan.crayne.org
> Newsgroups: comp.lang.asm.x86
>
> I see that Sender has also been replaced. I believe that the RFCs are
> clear about source address lines - From gives the nominal composer of
> the work, Reply-To if present gives where replies should go, and Sender
> if present names the minion who pressed the buttons. Putting wotan
> there is not right; if the deity must be recorded, that should be done
> in the header line of approval.

Then you should help Charles find an ISP which 100% trust him.

A moderator is the God in the newsgroup he moderates. But he does the
moderation job through a computer account. This account could be a
private *.com account you pay to have service on. It can be an
educational one or a government one. Anyway, a moderator is a nobody
in the ISP/school/government domain his account is unless he is also
the system administrator of that domain. His acts, including
injecting an approved message to Usenet, have to abide by the rules
his domain imposes.

Therefore Charles cannot preserve your NNTP-Posting-Host: information
-- perhaps his ISP or the news server he uses will overwrite that
information. He can move that information to X-NNTP-Posting-Host:,
but if his news server does not honor any X- headers, then he cannot
help.

Your demon.co.uk could have inserted something like
X-Abuse-Complaints-To: in your submitted message and perhaps Charles
cannot preserve it because his own ISP will put in its own and delete
all the X- headers.

Again, all these are up to the ISP Charles uses. You are not happy
about that? You can help Charles find an ISP which trust him 100% and
accept everything he injects in CLAX.

> BTW, my system generates an X-Mailer line. The spam-scanner adds its
> own X-lines, after that. But the X-CLAX lines are added before. ISTM
> that the header would be easier to understand if the extra lines were
> added after the existing ones.
>
> Where it is known that X-header lines are added by a transit agent such
> as UCSD, ISTM that they should NOT be preserved. They refer to the
> article as it was when it passed UCSD; it has since been processed by
> crayne.org, for which AFAIK UCSD is not responsible; they are of no
> interest to readers of News.

Maybe.

You used two "rubbish" on my words and usually that was enough to
convince me a reasonable exchange of views is no longer. Let me
summerize here:
(1) Charles as comp.lang.asm.x86 moderator is the God.
    His words are final.
(2) If you don't like his moderation policy, nobody
    stops you from posting in alt.lang.asm.
(3) If you are enthusiastic enough, you can even
    produce your own moderated newsgroup to be a
    God of your own; you can also set up a web chat
    forum for x86 assembly discussions. If your
    new site or new newsgroup does a better job than
    Charles', people will migrate there.

Enough rubbish told to you. Adios.

Loading