The RISKS Digest
Volume 20 Issue 68

Tuesday, 14th December 1999

Forum on Risks to the Public in Computers and Related Systems

ACM Committee on Computers and Public Policy, Peter G. Neumann, moderator

Please try the URL privacy information feature enabled by clicking the flashlight icon above. This will reveal two icons after each link the body of the digest. The shield takes you to a breakdown of Terms of Service for the site - however only a small number of sites are covered at the moment. The flashlight take you to an analysis of the various trackers etc. that the linked site delivers. Please let the website maintainer know if you find this useful or not. As a RISKS reader, you will probably not be surprised by what is revealed…


o RST discovers defective crypto in Netscape mail password saver
Gary McGraw
o Canada Post has "electronic post" on line
Alan DeKok
o buy now, pay never
David Shaw
o A Tale of Two Web Sites: Calling it secure doesn't make it so
Steven J. Zeil
o IDs in color copies and prints: confirmed
Lauren Weinstein
o BBC Censorship!
Peter McWilliams via Lindsay Marshall
o Melissa perpetrator faces five years in prison
o Oh, no! Y2K virus competitions
Ross Stewart via Peter de Jager
o Re: No bounds checking in Microsoft RTF controls
o Slashes in spreadsheets
Christopher Warnock
David Empson
o Risk of APC Power Chute
Geoffrey Coram
o Risks of e-mail monitoring
Thomas Roessler
o Re: Counterfeit Japanese coins and resulting risk...
Henry Spencer
o Re: Ladbroke Grove
Mark Brader
o USENIX Security Symposium 2000 - A Call for Papers
Moun Chau
o Call for Papers - Safecomp 2000
Gemma Windt-Krose
o Info on RISKS (comp.risks)

RST discovers defective crypto in Netscape mail password saver

Gary McGraw <>
Mon, 13 Dec 1999 17:18:18 -0500
Because remembering your passwords is a pain (you do have more than one,
don't you?), many programs are set up to remember them for you.  Exactly how
they do this is a risky business.  Netscape didn't do it right.  Beyond simply
stealing e-mail passwords, our discovery provides a gateway to other accounts
and systems since people generally use the same password
everywhere. Netscape has been notified of the flaw.

The POP3 and IMAP protocols are often used to read e-mail on a home or
office PC from a central mail server.  As a convenience to the user, many
programs offer to remember the user's password.  When Netscape offers to
save your e-mail password, it is encrypted before being stored in the
registry or preferences file on your computer.

Unfortunately, the encryption algorithm used by Netscape to scramble
passwords is exceptionally weak.  Tim Hollebeek, an RST Research Associate,
and John Viega, a member of the RST Software Security Group, were able to
deduce the algorithm after only eight hours of work.  No reverse engineering
of the software was involved.  Instead, a few hundred carefully chosen
passwords were analyzed using pencil and paper.  The algorithm turns out to
be a simple combination of XOR with a constant key and a substitution cipher
weaker than those found in puzzle magazines. For more details, see

Once the cipher is known, recovering a POP3 or IMAP password stored on a
machine is trivial.  Any attacker with physical access to the victim's
machine or the ability to run code on it can use our exploit.  Additionally,
passwords can be stolen from some versions of Netscape remotely using

RST has created a working password snagging attack in the lab.  A successful
attack allows the bad guy to download and read a victim's e-mail from a
remote machine. Since careful use of the hack would not leave too many
obvious clues, a victim's e-mail could be snooped indefinitely.  The only
workaround is to turn off the ``remember password'' feature.

Though stealing mail alone is a very serious security/privacy problem, more
dangerous scenarios exist. The largest risk is that people use the same
password for POP3 and other logins to remote machines (and maybe even their
PGP passphrase).  In particular, many people use IMAP or POP3 to access work
related e-mail from home, and their mail password is also the login password
they use at work.  In fact, the login account and the mail account are often
the same.  Home computers are notoriously insecure and easy to penetrate.  A
malicious attacker can read the POP3 password stored on an insecure home
computer (often over the net) and use it to log in to a more secure machine
run by the victim's employer.  The attacker can then take control of an
account, read sensitive information, attack more privileged accounts, and
set up remote monitoring systems inside a corporate network.  Our exploit
code could also be used as a payload in a much more insidious version of

  Quote of the day: ``We didn't do this with just a pencil and some paper.
Lots of our notes are in pen.  We didn't need to erase much.'' Tim Hollebeek
& John Viega

  Other quote: ``This is another illustration of how bad closed,
proprietary, cryptography is.  What makes this vulnerability particularly
nasty is that people tend to use the same passwords over and over again.  If
you can attack someone's mail server password, you're likely to also have
their login password, PGP password, etc.  Software security is important.''
Bruce Schneier

Gary McGraw, Ph.D., Vice President, Corporate Technology
Reliable Software Technologies

Canada Post has "electronic post" on line

Alan DeKok <>
Tue, 30 Nov 1999 19:02:41 -0500
Canada Post has recently been announcing it's new on-line mail service,
"epost", at  They claim it's a secure on-line service
for handling mail and bills via the web, as if it's something new and

Why would I need an on-line central post office if I can PGP encrypt/sign my
messages, or use a corporations SSL-enabled web site to view and pay my

In addition, they still fall into the standard security traps.  Your
"secure" e-mail is stored on their servers, where it is vulnerable to non-web
based attacks.  This is equivalent to having the paper version of the post
office open and sort all of your mail for you.

Their security statement at

has the following wonderful quote:

   If you are accessing your ELECTRONIC POST OFFICE BOX[tm] using a Web
   browser, you do not need to worry about viruses since no data are
   exchanged between the computer you are using and our system.

Umm... right.  How many security related browser bugs have their been in the
past year?

Alan DeKok buy now, pay never

David Shaw <>
Tue, 30 Nov 1999 16:00:58 +1100
The following is a summary of an article found on page 101 of the
*Sydney Morning Herald*, Saturday November 27, 1999.'s web-site,, an online audio CD shop, was
launched on October 18, 1999 with a significant flaw. It was possible to
order CDs and not have to pay for them.

Reports of the flaw resulted in Internet users swamping the web-site,
prompting an official to say "We're close to having a meltdown."

The problem with the site began with a design flaw which was originally
contrived to make life easier for consumers. Customers are given the
option of ordering a CD online but are not required to enter their credit
card details online if they feel uncomfortable about security.

The intention is of course that these customers would phone, fax or
e-mail their credit card details before receiving their goods. However,
the site does not tell consumers the phone numbers or e-mails to use to
pass on credit card details.

Customers who omitted their credit card details soon discovered they
received the CDs anyway. Invoices obtained by the *Herald* showed that
the "payment by" field was left blank.

Global Fulfilment, a US-based company, which provides the technical,
financial and ordering system for the site, denied any technical error,
but did reveal that the website had not installed an automated
credit card processing system in the first few weeks of operation. has cut its relationship with Global Fulfilment for the
technical design of the site, which will now be done on-site. Global
Fulfilment will still provide coordination and fulfilment of CD orders.'s shares are set to debut on the Australian Stock Exchange (ASE)
today (Tuesday November 30, 1999) after an AUD$8.4m float of 14%. maintains it lost no substantial revenue due to the free CDs and
is therefore under no obligation to file a prelisting disclosure to the ASE. has stated that the problem has been fixed and that no customer
would be retrospectively billed for any free CDs.

David Shaw, Alcatel Australia Limited

  [I eschewed any spelling corrector's suggestions of Fulfillment
  or Fulfilament.  PGN]

A Tale of Two Web Sites: Calling it secure doesn't make it so

Thu, 9 Dec 1999 13:19:59 -0500
My wife was recently engaging in some on-line Christmas shopping, and
stopped short of placing orders with two merchants because the usual "secure
site" icon was not being displayed on the web browser's status line (both
the latest versions of Netscape and Internet Explorer flagged the pages as

The first point of interest was that both sites featured prominent "This
site is secure" assurances, with links to pages explaining that they use
secure servers to protect credit card and other customer information.

A little investigation showed that the first site,,
really was using a secure https server, but the secured page was appearing
as a frame inside a larger page from an unsecure server, thus fooling both
browsers into displaying the "unsecured" icon.

The second site,, showed no signs at all
of using any security mechanism despite their stated "Privacy and Security
Policy".  We sent a message to their customer service address complaining of
this fact and asking if they would honor the sale prices advertised on their
web site even if we were to call their toll-free number and place the order
by telephone.  Their response absolutely floored us:

  "Thank you for writing to

  We apologize for any problems you are having using our site and appreciate
  your concerns and criticism regarding security.  Please try placing your
  order again, as there has been ongoing adjustments to the pages.  All our
  orders in the online store are placed via the Internet, even if you were
  to call us." [remainder omitted]

In other words, whether you click or whether you call, your credit card
number is going out over the net via plain-text!

So, one site that is secure masquerades as insecure. Another is truly
insecure, whether you use it or not!

Steven J. Zeil                 
Dept. of Computer Science  Old Dominion University  Norfolk, VA 23529

IDs in color copies and prints: confirmed

Lauren Weinstein <>
Wed, 08 Dec 99 14:52:59 PST
Greetings.  In my latest PRIVACY Forum Digest, I've presented
a report confirming the fact (well known in the copier trade,
but something of an urban legend for everyone else) that xerographic
color copiers/printers include "steganographically" encoded IDs.
Essentially, these IDs are encoded repeatedly as part of the
background "noise" in images, and cannot be decoded without
knowledge of the proprietary algorithms involved.

The full report can be viewed at:

In that report, I mentioned the possibility of such IDs being implemented
within inexpensive, consumer-level equipment.  This could include both
copiers and such widely-used devices as inkjet printers.  Those readers who
might question this possibility might wish to take a look at:

which is a Bureau of Engraving and Printing procurement offering
dealing with precisely these issues.

Lauren Weinstein
Moderator, PRIVACY Forum -
Co-Founder, PFIR: People For Internet Responsibility -

BBC Censorship!

Fri, 3 Dec 1999 13:11:54 +0000 (GMT)
This item was in rec.humor.funny. If you go to
<> and try the "E-mail a friend" button,
you can verify that this indeed happens.

----- Original Message -----
From: Peter McWilliams <>
Newsgroups: rec.humor.funny
Sent: Friday, December 03, 1999 3:30 AM
Subject: BBC Censorship!

> The BBC web pages have links which allow you to send web pages to e-mail
> addresses and to include a text message.
> I just sent myself a page with the word Saturday in the body of the
> message. It arrived with the word changed to Sa****ay!
> Hmmm. So I tried sending...
> >"I hope you still have your appetite for scraps of dickens when I bump
> >into you in class in Scunthorpe, Essex on saturday"
> Yes! It replied...
> >"I hope you still have your appetite for s****s of dickens when I ***p
> >into you in class in S****horpe, Es*** on sa****ay"
> Missed the t**, the a** and the d*** though.  [* This line PGN-ed]

> Peter Mc*****ams

> Selected by Jim Griffith.  MAIL your joke to
> Attribute the joke's source if at all possible.  A Daemon will auto-reply.
> Jokes ABOUT major current events should be sent to
> (i.e., jokes which won't be funny if not given immediate attention.)
> Anything that is not a joke submission goes to
> For the full submission guidelines, see
> This joke's link:

    [* The other parts (of speech!) have been truncated in
    the name of Internet Decency.  PGN-Petered out, I guess]

Melissa perpetrator faces five years in prison

"NewsScan" <>
Thu, 09 Dec 1999 09:32:30 -0700
David L. Smith, the New Jersey computer programmer who has pleaded guilty to
having created the Melissa virus that infected more than 100,000 computers
worldwide and caused an estimated $80 million in damages, is likely to face
a sentence of five years in federal prison.  Security expert and former
prosecutor Mark D. Rasch says, "The federal government is trying to send a
message that they take these kinds of Internet crimes very seriously. This
is also a recognition that a single individual has the ability to cause tens
of millions of dollars of damage." (*The New York Times*, 9 Dec 1999,; NewsScan Daily, 9 Dec 1999)

Oh, no! Y2K virus competitions

Ross Stewart <>
Wed, 08 Dec 1999 13:23:21 +1300
   [Via: Peter de Jager <> (tel 1-905-792-8706)]

>I don't have a lot of patience nor tolerance for those who use their
>talents to try and stuff up my life nor those of my staff by developing
>ever more sophisticated and warped viruses.  All it does is cause us grief
>and hinders us from finding good jobs for smarter, more honourable and
>usually more intelligent IT professionals.
>We have just received a warning from an anti-virus specialist (poacher
>turned gamekeeper ?) about some idiots in Holland who have established a
>competition for the best Y2K virus.  The p[i?]rates involved have set up
>"game rules" with the only requirement being that each entry has to be a
>virus/trojan or backdoor, and preferably related to the Millennium Bug or
>the year 2000.
>Quoting (after editing) from the site :
>"The viruses that will be sent for the Y2K infection feast [sic] will be
>sent 4-5 days before the release to AV companies. The one that gets
>detected last will be the winner."
>Words fail me.
>I'm told there are almost 50,000 viruses floating around the world
>now.  It's a real pain that some stupid ***** ***** is going to further
>waste my time over Y2K, just when I don't need the hassle.  I'm sure you
>don't either.
>What to do?
>We're disconnecting our e-mail servers from the WWW and setting up an e-mail
>service on a standalone PC.  This standalone machine will do ALL e-mail
>downloads from real soon now until well after New Year.  We'll keep that
>machine loaded to the gunwales with the latest anti-virus software and will
>endeavour to ensure that any virus outbreaks are thus contained.
>We strongly suggest you seriously consider doing the same.
>We'll pass CVs internally across to other "more sensitive" machines in .rtf
>format (using rtf avoids Word-type macro viruses).
>Remember also that the major anti-virus companies have all agreed to
>provide FREE 90-day versions of their software to cover this period.  GET
>SOME !!!  before 31 December via
>Lastly, remember to pull the power cord/s out of the wall socket/s when you
>go home before New Year, a simple method of avoiding any possible surge
>damage to electronic equipment.
>Thinks : Wouldn't it be nice if we could have a quiet Y2K period ????
> A R (Ross) Stewart, Wilson White Group, Auckland, New Zealand
> ph: +64(9)307-3869 (posted at

Re: No bounds checking in Microsoft RTF controls (Downes, R-20.66)

meeroh <meeroh@MIT.EDU>
Wed, 01 Dec 1999 18:28:12 -0500
>I can think of hundreds, thousands, hundreds of thousands of loops I
>have written and seen over the years, everyone of course having a bounds
>check built in. I mean, this is very _basic_ programming, isn't it?
>  for (cp = buf; cp < buf + BUFSIZE; cp++)
>    /* * */

It may be worth pointing out that a more correct way of implementing this
loop is

  size_t size = BUFSIZE;
  while (size-- > 0) {
     /* ... */

The original code fails in the subtle case when buf + BUFSIZE extends past
the end of the address space. Note that the ANSI C guarantee that you should
be able to compute the address of the first element beyond the end of a
named array doesn't apply if buf is dynamically allocated.

This common idiom, while it's guaranteed to work if buf is a named array of
characters, will fail if buf was dynamically allocated — yet almost every
programmer I've seen uses the same idiom for both cases.

(See Writing Solid Code by Steve Maguire, p 132.)


PS. Yes, I do realize that in the context of a local buffer with hardcoded
size, which was the topic of the original post, this idiom is okay. | <> | MIT I/S Mac developer

Slashes in spreadsheets (Re: Quirk, RISKS-20.64-65)

"Christopher Warnock" <>
Mon, 22 Nov 1999 08:04:31 -0500
The "slash" usage in Excel can be changed quite easily to another character
or to nothing at all, effectively being disabled.  In Excel, under
Tools/Options, select the "Transition" tab and you will see the text
"Microsoft Excel menu or Help key:" and this will have a text box to the
right of it containing a "/" character.  In this text box, type the
character that you wish to use for this function and then select the "OK"
button.  Leave this text box blank to disable the functionality.

Christopher Warnock, CCNA, Senior Network Architect
Resource One Computer Systems V: 614.228.8165 F: 614.241.5810

Slashes in spreadsheets (Re: Quirk, RISKS-20.64-65)

David Empson <>
Tue, 23 Nov 1999 01:21:57 +1300
> From: Kent Quirk <>
> [Lotus] 1-2-3's menu system (before any attempt at GUI standards and
> before the mouse was common) was invoked by hitting slash. Once you got
> used to it, you really really liked it. So /fr was "menu file retrieve",

It is even older than that: the slash key was used in the same manner by the
original spreadsheet program: VisiCalc running on an Apple II (circa 1980).

David Empson <>
Snail mail: P O Box 27-103, Wellington, New Zealand

Risk of APC Power Chute

<Geoffrey Coram>
Tue, 7 Dec 1999 18:19:28 -0500
I recently installed APC's Power Chute Plus for Win98/NT, which allows
control of various features of their UPS devices.  I was surprised when I
started the program that it listed two possible devices, mine and another
machine in my building.

It turns out that the default installation of PC+ sets up the installation
directory as shared to everyone in your workgroup - perhaps a fine idea for
corporate networks, but not the right idea for cable modems and campus dorms
(my case).  I don't want my neighbors to have the ability to simulate a
power failure - with the loud warning beep - at 4 AM, or otherwise mess with
the settings.

I happened to have file sharing turned off, so the properties of the
directory were ignored.  I was surprised that this was not mentioned during
the install process, but instead buried in the help file.  APC tells me they
will consider my recommendation to switch the default, and let the
(presumably) more knowledgeable corporate network administrator enable the
feature if he wants it.

Just another risk of shared networks.

Risks of e-mail monitoring

Thomas Roessler <>
Wed, 8 Dec 1999 00:46:48 +0100
In some small firms, a simplistic approach may be applied to monitor
employees' electronic mail: A responsible person inside the firm receives a
copy of each incoming message, and possibly carbon copies of outgoing

The privacy problems of this approach are obvious, however, they are not
what this is about.  In fact, there is another risk for the person who
*receives* the message copies which is not completely obvious.

This risk is closely linked to the way in which most e-mail clients in
common use today display message overviews to users.  In most cases, users
are pesented a list of message senders and subjects.  The list of recipients
is not shown, and no indication of the recipient is given, since the
software assumes that every message in the user's inbox is intended for him.
(There are exceptions from this, for instance the pine and mutt e-mail
clients for Unix.)

Given no help from the software, the user will make up his own mapping from
the information presented to him (i.e., subjects and message senders) to
recipients. In particular, when there are some message senders who
correspond with one or two employees on a regular basis, the person doing
the monitoring will mis-recognize messages from these senders which are
actually directed to her, and not to the senders' usual correspondence
partners.  The monitoring personnel may even delete these messages unread
when the correspondence between the parties in question is not considered
worth the effort of closer monitoring.

Given that the monitoring personnel might frequently be senior employees,
this effect may even lead to the loss of particularly important e-mail
messages sent by clients to these senior employees.

The cure?  Use e-mail clients which tell you whether your e-mail is personal
or not.  Or, even better, use a reasonable e-mail monitoring policy which
(1) avoids the privacy implications of the simplistic scheme and (2) makes
it possible to easily distinguish surveillance output from personal
messages.  Finally, senior employees may just be forced to read *all*
messages which reach their e-mail inbox.  However, this may not be an option
depending on the amount of noise introduced by the monitoring.

Re: Counterfeit Japanese coins and resulting risk... (Opie, R-20.67)

Henry Spencer <>
Tue, 7 Dec 1999 18:59:14 -0500 (EST)
>The lesson to comp.risks?  Of course, that assumption that a coin is a coin.

Unfortunately, merely returning the customer's own coin is not sufficient.
Unless there are other restrictions, it's almost as easy for the crook to
make some trivial purchase, requiring most of his supposed money to be
returned as change.  (This is a common counterfeiter tactic, although in
the past it has usually been applied to humans rather than machines.)

Henry Spencer  (

Re: Ladbroke Grove (RISKS-20.62)

Mark Brader <>
Mon, 6 Dec 1999 17:37:44 -0500 (EST)
In the first item on the Ladbroke Grove rail accident, the number of deaths
was given as "at least 70".  It turned out that that number included a large
number of people that simply *might* have been on the train and could not be
immediately located.  Based on information I got from Clive Feather, there
were only 30 deaths that day (including the two drivers), plus one more who
died from burns on 3 Nov 1999.

USENIX Security Symposium 2000 - A Call for Papers

Moun Chau <moun@usenix.ORG>
Mon, 6 Dec 1999 19:39:27 GMT
9th USENIX Security Symposium 2000 Conference
August 14 - 17, 2000
Denver, Colorado, USA
Conference URL:

The USENIX Security Symposium brings together researchers, practitioners,
system administrators, systems programmers, and others interested in the
latest advances in security and applications of cryptography. The keynote
speaker is Dr. Blaine Burnham, Director of the Georgia Tech Information
Security Center (GTISC) and formerly Program Manager for the National
Security Agency (NSA) at Ft. Meade, Maryland.

We are currently seeking submissions for Refereed Papers, Works-In-Progress
Reports, Talks/Panel Session proposals, and Tutorial presentation proposals
for this event. If you are working in any practical aspect of security or
applications of cryptography, the program committee urges you to submit a

Please see the detailed author guidelines, which include a sample abstract,
for more information.

* Paper submissions due: 10 Feb 2000

USENIX Security Symposium 2000 is sponsored by USENIX, the Advanced
Computing Systems Association, in cooperation with the CERT Coordination
Center. USENIX is an international membership society.

Call for Papers - Safecomp 2000

"Windt-Krose, Gemma" <>
Thu, 2 Dec 1999 16:11:49 +0100
SAFECOMP 2000 Call for Papers

          19th International Conference on
     Computer Safety, Reliability and Security

  October 25-27, 2000    ROTTERDAM, The Netherlands

Papers are invited on all aspects of state-of-the-art, experiences and new
trends in the area of computer safety, reliability and security regarding
critical applications of computer systems.  Special topics include security
relevant to safety, human factors, hardware solutions, verification &
validation, distributed systems, and safety-critical y2k experiences.
Special themes are on medical systems, transport & infrastructure systems,
and software process improvement. Contributions from research, industrial
applications and experiences, and on licensing questions are welcomed.

====> Paper submission DEADLINE: FEBRUARY 15, 2000 <====


Please report problems with the web pages to the maintainer