[OT] Conjectures? on the "constant pointer on Turbo C programming"junk

[Francois Grieu]

Hi,

I'm getting increasingly puzzled at what that "constant pointer on Turbo
C programming" junk flooding the group can *really* be.

The messages are composed much as if originating from real humans with
varying (generally low) knowledge on C and programming, instructed to
post a question on (Turbo) C programming on the group comp.lang.c thru
Google groups, stating affiliation to the Islamic University of
Technology, department, name and 6-digit student number. Some message
seem to forget one part of the assignment, use a name not matching the
google account, are duplicated... just like would happen for real
students doing that task. All messages seem to originate from
Bangladesh, typically either from an IP block apparently assigned to
(within a one-letter typo-like variation) the Islamic University of
Technology, or an ISP in Bangladesh, again just like would happen in a
university with some students using the universitie's computers.

To me, if this is a robot, it passes some reduced Turing test.

The nicely worded complaint emails sent to webmaster, abuse, and other
published addresses listed on the website of the Islamic University of
Technology, by me and others, have not been answered. Even more
strangely, there is AFAIK never any follow-up on the genuine answers
made to (a sizable fraction of) the original posts, which very much
contradicts the apparent objective of the posts.

I fail to find any rational explanation to all this; one-sided wormhole
from a kafkayan universe to our spot in usenet does not count.


Francois Grieu

 

[Seebs]

The messages are composed much as if originating from real humans with
varying (generally low) knowledge on C and programming, instructed to
post a question on (Turbo) C programming on the group comp.lang.c thru
Google groups, stating affiliation to the Islamic University of
Technology, department, name and 6-digit student number. Some message
seem to forget one part of the assignment, use a name not matching the
google account, are duplicated... just like would happen for real
students doing that task. All messages seem to originate from
Bangladesh, typically either from an IP block apparently assigned to
(within a one-letter typo-like variation) the Islamic University of
Technology, or an ISP in Bangladesh, again just like would happen in a
university with some students using the universitie's computers.

It does seem awfully plausible.

The nicely worded complaint emails sent to webmaster, abuse, and other
published addresses listed on the website of the Islamic University of
Technology, by me and others, have not been answered.

Which is pretty strange!

Even more
strangely, there is AFAIK never any follow-up on the genuine answers
made to (a sizable fraction of) the original posts, which very much
contradicts the apparent objective of the posts.

Actually, I don't think that's strange.

It seems to me that the kids presumably have an assignment to post to Usenet.
They do not have an assignment to retrieve answers, or interact in any way.
Most of them don't seem to have very good English. My guess would be that
the name and student number are there for the TA who is grading the assignment
by confirming that there was a post from each name and number pair in the
class which asked a basic question about C -- possibly from a predefined
list of "possible questions".

This looks EXACTLY like what you'd expect from a bunch of lazy students who
don't care at all about learning and are doing exactly the bare minimum
necessary to obtain a passing grade.

I fail to find any rational explanation to all this; one-sided wormhole
from a kafkayan universe to our spot in usenet does not count.

I would think it more likely we were the kafkayan universe.

I can explain the students easily. They don't have any interest in learning,
they're just following a badly-written course plan. The lack of any responses
from the various people at IUT is a bit weirder. It seems odd to me to
imagine that the entire staff of a university would consist of people who
simply don't care *at all* about any queries receieved via email. Not even
enough to respond with "we think it's a great assignment" or "bugger off,
it's none of your business" or "it's not us, it's some guy".

And that part means that, even if the posters are just some troll or trolls,
I should probably view IUT on a resume as a disqualifier, because a university
run by people who can't be bothered probably does not produce good students.

-s

 

[Seebs]

A zillion miles from the truth. No response has ever shown any
indistinguishability from what a human might respond. Yes, that's
a vacuously true assertion, as there have been no responses.

Well, imagine a Turing test reproduced to "no responses, just generates
a series of posts".

It seems to me that they do a fairly good job of imitating a series of
students posting stuff.

-s

 

[Phil Carmody]

^^^^^^^^^^^
Not a poster people should be including in broad killfile filters, IMHO.

Hi,

I'm getting increasingly puzzled at what that "constant pointer on
Turbo C programming" junk flooding the group can *really* be. ....
To me, if this is a robot, it passes some reduced Turing test.

A zillion miles from the truth. No response has ever shown any
indistinguishability from what a human might respond. Yes, that's
a vacuously true assertion, as there have been no responses.

Even my game-playing bot responds more intelligently than the
UIT drones. (Mostly by saying "I'm just a dumb bot, I didn't
understand that".)

Phil

 

[Francois Grieu]

It does seem awfully plausible.

Which is pretty strange!

Actually, I don't think that's strange.

It seems to me that the kids presumably have an assignment to post to Usenet.
They do not have an assignment to retrieve answers, or interact in any way.
Most of them don't seem to have very good English. My guess would be that
the name and student number are there for the TA who is grading the assignment
by confirming that there was a post from each name and number pair in the
class which asked a basic question about C -- possibly from a predefined
list of "possible questions".

This looks EXACTLY like what you'd expect from a bunch of lazy students who
don't care at all about learning and are doing exactly the bare minimum
necessary to obtain a passing grade.

My experience with teaching computer science (admitedly old and limited in time, location, and cultural variety) is that some fraction (1/10 maybe) of students will become deeply interested in the subject and go beyound the assignment, especially when it comes to communicating. I find the oberved total lack of followup very disturbing for the simplest theory (mechanical execution of an ill-advised assignment by numerous students).

I would think it more likely we were the kafkayan universe.

How true...

I can explain the students easily. They don't have any interest in learning,
they're just following a badly-written course plan. The lack of any responses
from the various people at IUT is a bit weirder.

That could have a sound explanation e.g. a technical reason and/or webmaster gone for lack of subsidy; but any student able to post on a google group is able to read a reply from it and reply on that AFAIK.

It seems odd to me to
imagine that the entire staff of a university would consist of people who
simply don't care *at all* about any queries receieved via email. Not even
enough to respond with "we think it's a great assignment" or "bugger off,
it's none of your business" or "it's not us, it's some guy".

And that part means that, even if the posters are just some troll or trolls,
I should probably view IUT on a resume as a disqualifier, because a university
run by people who can't be bothered probably does not produce good students.

Anyone had sucess at grabbing some of the scientific articles referenced on the "Journal Publications" page of the IUT website?
<http://www.iutoic-dhaka.edu/journals.php>


Francois Grieu

 

[Francois Grieu]

Le 18/05/2010 21:56, Phil Carmody a écrit :

^^^^^^^^^^^
Not a poster people should be including in broad killfile filters, IMHO.

I use gmail (decent free webmail usable around the globe with a good spam filter); but I seldom post thru google groups (hopeless mangling of content).

To expand: by "reduced" I meant reduced to unidirectional channel.

A zillion miles from the truth. No response has ever shown any
indistinguishability from what a human might respond. Yes, that's
a vacuously true assertion, as there have been no responses.

Even my game-playing bot responds more intelligently than the
UIT drones. (Mostly by saying "I'm just a dumb bot, I didn't
understand that".)

All text-generating bots that I ever met look dull in comparison to this one, if it is a bot. Which I doubt. Some rare posts, like that one
Date: Sun, 16 May 2010 10:11:24 -0700 (PDT)
Subject: Warning in structure programme
Message-ID: <[email protected]>
show some genuine effort. I hope it does not degenerate into a bad mark for failing to include the student number.


What if c.l.c was a guinea pig for some experiment run by true students in phycho/sociology, or Google staff, or anyone with the ability to forge usenet headers (not hard), and they are laughing like mad that I could believe a second real students would do what I consider these IUT students could be doing?

Francois Grieu

 

[Francois Grieu]

One characteristic that I've noticed is that sometimes
several posts will show up at about the same time
with the same name in the Sender line,
but when the posts are opened up
they turn out to be from people with different names.

I account for this effect by conjecturing that a single gmail account is created and used to send several messages. It could well be that the same person sends several messages with different names, but the apparent variations of skill level (in english and computer programming) and kind of IP sources makes me doubt that all messages originate from the same individual.

I'm still thinking that they are the work of a Markov text generator.

If that's true, it is one cleverly written to mimic human errors, and largely avoids the pitfall of all Markov text generators that I watched, which is that there is no recognisable line of though in their production.

Dann Corbit once posted a text generator called "Malbrainspam"
which generated text in the style of posts from Karl Malbrain.

Francois Grieu

 

[Seebs]

My experience with teaching computer science (admitedly old and
limited in time, location, and cultural variety) is that some
fraction (1/10 maybe) of students will become deeply interested in
the subject and go beyound the assignment, especially when it comes
to communicating. I find the oberved total lack of followup very
disturbing for the simplest theory (mechanical execution of an
ill-advised assignment by numerous students).

What if the students taking the assignment seriously don't do it so blatantly
that we all immediately spot them?

Alternatively, I'm forming a secondary theory that perhaps IUT is just a
really crappy school.

That could have a sound explanation e.g. a technical reason and/or
webmaster gone for lack of subsidy; but any student able to post
on a google group is able to read a reply from it and reply on that
AFAIK.

Technically able, true. Hmm.

Anyone had sucess at grabbing some of the scientific articles referenced on the "Journal Publications" page of the IUT website?
<http://www.iutoic-dhaka.edu/journals.php>

Haven't tried yet.

-s

 

[Seebs]

What if c.l.c was a guinea pig for some experiment run by true
students in phycho/sociology, or Google staff, or anyone with the
ability to forge usenet headers (not hard), and they are laughing
like mad that I could believe a second real students would do what
I consider these IUT students could be doing?

Then they have obviously never met real students, because real students do
stuff like this all the time.

-s

 

[Phil Carmody]

Well, imagine a Turing test reproduced to "no responses, just generates
a series of posts".

I tried that, but it came out as imagining a dog's tail is a leg.

It seems to me that they do a fairly good job of imitating a series of
students posting stuff.

Creating original material in a particular style has never been an
important component of a Turing test. That requires little more than
a Markov Model, or equivalent, nothing more. (I think the first ever
O'Reilly perl book included a script which would do that when fed a
source corpus - travesty.pl, I believe it was called.) Less, even,
as even a trivial generative grammar would do.

Phil

 

[Francois Grieu]

Creating original material in a particular style has never been an
important component of a Turing test. That requires little more than
a Markov Model, or equivalent, nothing more. (I think the first ever
O'Reilly perl book included a script which would do that when fed a
source corpus - travesty.pl, I believe it was called.) Less, even,
as even a trivial generative grammar would do.

The Markov model theory does not explain well messages like this one

From: Rafe <[email protected]>
Subject: Tubo C
Date: Sat, 15 May 2010 01:58:50 -0700 (PDT)
What is the relation of C programming with CIVIL engineering?


The capitalization of CIVIL is very pertinent; it seems exactly like the
student writing that is bored with the task that he suspect is
irrelevant to what he wants to learn. This rethoric effect, or text with
wording similar to this one, is found in none of the other posts. I have
a few other examples favoring the hypothesis of that there is
intelligent and diverse life on the originating side of the link.

Also: anyone with the capability to write such a good Markov model
generator would be able to ask much brighter questions and be bored to
death by the ones we see.



Among the serious conjectures that I form or have seen in other posts
for the whole mess (some overlap)

a) misguided assignment mostly mechanically executed, perhaps because of
a culture of obedience to the teacher / teaching that was long lost /
overcame in universities that I know.

b) semi-organized protest of the students against unpopular teacher /
teaching program / misguided assignment by purposedly executing almost
verbatim one of the assignment given.

c) deliberate attempt to increase the credibility of / awaremess on the
university and its activity on (use)net by any mean
c1) with extremely poor execution,
c2) along the theory that bad publicity remains puclicity and thus
better than none (which is a win in our era of search engines that only
touch the surface of things).

d) attempt to decrease the credibility of the university by making us
believe c) is attempted


The deliberate use of Wikipedia for the purpose of promotion of the
university shows that at least the desire to use (if not abuse) the
power of internet towards goal c) exists. So far I see nothing to
substanciate d).

I hope it was b), but a)+c) is simpler/better.


Francois Grieu

 

[Nick Keighley]

I tried that, but it came out as imagining a dog's tail is a leg.


Creating original material in a particular style has never been an
important component of a Turing test. That requires little more than
a Markov Model, or equivalent, nothing more. (I think the first ever
O'Reilly perl book included a script which would do that when fed a
source corpus - travesty.pl, I believe it was called.) Less, even,
as even a trivial generative grammar would do.

what difference C wanna IUT a
(095411) student Mohammad CEE structure What about Student Turbo of
IUT
Mahmudul Your on union? Hi, everyone, Hi, Hi, everyone, everyone, the
Department in that,I of question Department.I've shovon a Tawsif is
between Jahid Programming. know Hasan University CEE Department of
095429 Islamic Scanf? and Technology is string of c CEE & the hasan
I'm faithfully, what difference C wanna IUT a (095411) student
Mohammad CEE Printf What about Student Turbo of IUT Mahmudul Your on
union?

 

[Alan J]

The Markov model theory does not explain well messages like this one

From: Rafe <[email protected]>
Subject: Tubo C
Date: Sat, 15 May 2010 01:58:50 -0700 (PDT)
What is the relation of C programming with CIVIL engineering?


The capitalization of CIVIL is very pertinent; it seems exactly like the
student writing that is bored with the task that he suspect is irrelevant
to what he wants to learn. This rethoric effect, or text with wording
similar to this one, is found in none of the other posts. I have a few
other examples favoring the hypothesis of that there is intelligent and
diverse life on the originating side of the link.

Also: anyone with the capability to write such a good Markov model
generator would be able to ask much brighter questions and be bored to
death by the ones we see.



Among the serious conjectures that I form or have seen in other posts for
the whole mess (some overlap)

a) misguided assignment mostly mechanically executed, perhaps because of a
culture of obedience to the teacher / teaching that was long lost /
overcame in universities that I know.

b) semi-organized protest of the students against unpopular teacher /
teaching program / misguided assignment by purposedly executing almost
verbatim one of the assignment given.

c) deliberate attempt to increase the credibility of / awaremess on the
university and its activity on (use)net by any mean
c1) with extremely poor execution,
c2) along the theory that bad publicity remains puclicity and thus better
than none (which is a win in our era of search engines that only touch the
surface of things).

d) attempt to decrease the credibility of the university by making us
believe c) is attempted


The deliberate use of Wikipedia for the purpose of promotion of the
university shows that at least the desire to use (if not abuse) the power
of internet towards goal c) exists. So far I see nothing to substanciate
d).

I hope it was b), but a)+c) is simpler/better.

My guess:

e). None of the above; it's "performance art".