The World of Computer Software/World_Of_Computer_Software-02-386-Vol-2of3.iso/a/aot-4.txt

Text File | 1993-01-30 | 43KB | 1,028 lines

         The Art of Technology Digest #4   Friday, September 4th, 1992

 %%AoT%%%%AoT%%%%AoT%%%%AoT%%%%AoT%%%%AoT%%%%AoT%%%%AoT%%%%AoT%%%%AoT%%%%AoT%%

  Editor: Chris Cappuccio ([e-mail])
  BBS Archivist: David Mitchell ([e-mail])
  E-Mail Archivist: Mike Batchelor ([e-mail])


[AoT Digest] Contents #4 (Fri, September 4th, 1992)

 Article 1: Notes From The Editor
 Article 2: Legalities Re: New Encryption Program
 Article 3: New Journal
 Article 4: UFC (Ultra Fast Crypt) on CM200
 Article 5: AT&T USL vs. BSDI/UCB, Mach3, OSF/1, GNU HURD, Linux
 Article 6: Important N.S.W. Report Released
 Article 7: Fido-Usenet Gateway
 Article 8: Linux 0.97pl2 Information
 Article 9: Announcement: VMS NEWS v1.24 released...
 Article 10: Cardinal 14.4 S/R Fax Modem
 Article 11: Microsoft Windows NT Device Driver Developers Conference

The Art of Technology Digest is distributed in the following ways:

By E-MAIL, send e-mail to [e-mail] and, to subscribe to
Art of Technology Digest, leave the subject blank and enter:  SUBSCRIBE aotd.
To get a back-issue of Art of Technology Digest, leave subject blank and
enter:  GET aotd/vol<number>.zoo UUENCODE (Example:  To get AOT-D number 2,
use GET aotd/vol2.zoo UUENCODE).  To get an index of Art of Technology Digest,
leave subject blank and enter:  INDEX.  To get AoT-D by BBS, Call
[phone], Live Wire BBS.  This system maintains a complete collection
of AoT Digest.  Speeds are 1200/2400/HST-9600/HST-14,400.

Or, if you have Internet FTP Access, the anonymous FTP site is:
crs.cl.msu.edu (35.8.1.10)
under /pub/cappucci/aot/
Note: When you anonymous ftp to crs, you MUST use some password!

The Art of Technology Digest is an open forum dedicated to sharing
information among computerists and to the presentation and debate of
diverse views. AoT-D material may be reprinted as long as the source
is cited.  Some authors do copyright their material, and they should
be contacted for reprint permission.  It is assumed that non-personal
mail at the moderators may be reprinted unless otherwise specified.
Readers are encouraged to submit reasoned articles relating to
computer culture and communication.  Articles are preferred to short
responses.  Please avoid quoting previous posts unless absolutely
necessary.  All articles for submission should be sent to:

                        [e-mail]

DISCLAIMER: The views represented herein do not necessarily represent
            the views of the moderators. Digest contributors assume all
            responsibility for ensuring that articles submitted do not
            violate copyright protections.

"Dan Quayle may be the secret weapon for the Republicans"
                -Reporter, 1992 Republican Convention


---------------------------------------------------------------------------

From: Chris Cappuccio, AoT-D Editor <[e-mail]>
Date: 8/14/92
Subject: Article 1--Notes from the Editor

Anonymous FTP site for Art of Technology Digest:
 crs.cl.msu.edu (35.8.1.10)
 under /pub/cappucci/aot/

Note: When you anonymous ftp to crs, you MUST use some password!

As you may have noticed, my e-mail address has changed from
[e-mail] to [e-mail].
All articles for submission should now go my new address.
It's been a little while since the last AoT-Digest, but, with
school and lots of other stuff, I've had no time to work on this.
Anyways, in the works are more anonymous ftp sites for AoT-Digest
and a newsgroup. I want to thank Computer Underground Digest
for printing an article on AoT-Digest, and now we have hundreds
(maybe even thousands!) more e-mail subscribers. I want to remind
people that all articles that deal with Technology (not just computers)
are welcome and they have to be somewhat recent (nothing before August
25th will be accepted in AoT-Digest #5).

------------------------------

From: [e-mail] (Dane C. Butzer)
Subject: Article 2--Legalities re: new encryption program
Date: 1 Sep 92 02:54:15 GMT

First of all, even though I'm sending this posting through OSU (where I'm a
graduate student), I'm making it in my capacity as a partner in an independent
company - we don't have internet access yet...

This posting is a request for some information from the security world out
there.  Basically, we'd like to find out what legal requirements/hurdles exist
for the sale of a new data encryption algorithm.

More specifically, we've developed a data encryption program based on a new
type of random number generation.  This generator produces a stream of numbers
with the following basic properties:
     1) The cycle of repetition is at least 10^34 (ten to the thirty-fourth).
     2) The correlations between any sections of the numbers is statistically
    zero (with a very high level of confidence).
     3)    2^128 (two to the one hundred twenty eighth) different number streams
    can be generated (ie. there are 2^128 different keys).
     4)    Changing the key be even one bit is guaranteed to change over 99% of
    the numbers.
     5) The cross correlations between each different number stream is almost
    zero. (With a normalized correlation of 1 being complete correlation,
    and a sample size of 8000 numbers, the typical correlation values are
    around 0.01; sometimes they are as high as 0.05, and sometimes they
    are as low as 0.00001).

The applications for such a generator in encryption are obvious.  Now, we're
curious about the legal side of selling something like this.  Are there
special certification, registrations, etc. that have to be taken care of?  Can
such a program even be distributed/sold legally?  Will we get squashed by the
government (Agents burst into the apartment and take everything at bazooka
point ;-> )?

Also, we're thinking about distributing a free copy of a limited version of
the program for the internet world to try out.  This version would only be
able to encrypt files of size 10k or less, would be time-bombed (ie. die off
after a certain date), and would probably not incorporate the most efficient
implementation of the generator (ie. it would be slow).  We're also thinking
about releasing an encrypted message and most of the corresponding clear text,
with a fee of $500 offered to whoever can "break" the encryption first (This
is just an idea right now - the rules to such an experiment would have to be
spelled out in detail - not something to bore you with now.)  Any comments on
the legality/risks involved with doing this?

Please respond here or send e-mail to [e-mail].

Thanks, and if this is the wrong place to have posted this, or if this message
seems too commercial, sorry - I'm new at posting to the internet :-) Also, I
had no clue as to where else to go to get this information.


/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\
| This posting is of a purely personal nature.  It is not a posting as |
| student, or a graduate, of OSU.  The views and opinions expressed    |
| are mine.  OSU has no control over them.                 Dane Butzer |
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/

------------------------------

From: [e-mail] (Ross Anderson)
Subject: Article 3--New Journal
Date: 2 Sep 92 10:41:15 GMT

Keeping up with research in computer and communications security is
becoming a very time-consuming activity. A conscientious researcher in
the field would have had to read well over 600 papers last year; and
as the amount of research activity grows, the problem can only get worse.

We are currently developing, with Cambridge University Press, an abstracting
service designed to solve this problem.

Computer and Communications Security Abstracts will summarise
research in computer security topics such as access control, database
security, formal methods, distributed systems, biometrics, security
management, risk management, contingency planning, legal issues, audit, and
applications; and in communications security topics including stream and
block cipher techniques, public key cryptography and computational number
theory, complexity and theoretical cryptography, cryptanalysis,
authentication, protocols, and applications.

Our mission is to provide abstracts of as much published research and
development work as possible. This includes not just conference and journal
papers, but also research reports and theses. We will make a particular
effort to report work which is published in languages other than English,
or which for other reasons might escape the notice of the research
community.

We expect that the first issue will be published in March 1993. It will be
quarterly to begin with, and become bi-monthly once a sufficient flow of
abstracts has been established. These should be a dummy issue out at the
end of 1992, which will be circulated with subscription information.

If you could be able to help us with abstracting work (particularly in
foreign languages), then we would be very keen to hear from you.

Material published in the main periodicals (IEE and IEEE journals, Journal
of Cryptology, Mathematics of Computation, Journal of Computer Security,
Computers and Security, Cryptologia) and the main conference proceedings,
will be reviewed automatically. However, if you are publishing material
elsewhere - such as in the form of a departmental research report, or in a
provincial journal - we would suggest that you sent us an offprint to
ensure coverage. This should be mailed to:

Ross J. Anderson ([e-mail])
University Computer Laboratory
Pembroke Street
Cambridge CB2 3QG
England

------------------------------

From: [e-mail] (Jorgen Bo Madsen)
Subject: Article 4--UFC (Ultra Fast Crypt) on CM200
Date: 2 Sep 92 19:36:57 GMT

Dear fellows,

The project: The implementation of UFC (Ultra Fast Crypt) on UNI-C's
CM200 is now finished.

The work was done by Michael Glad, who is also the auther of UFC.
The CM200 is a small Connection Mashine (parallel computer)
with 8K CPUs from Thinking Machines Corporation.

Below are the results:

  - The C* (C-star) version of UFC can crypt 55,000 words pr. second.
    It is higly optimized and uses special PARIS functions in all the
    time comsuming parts.

  - The Fortran version af UFC can "only" crypt 25,000 words pr. second.

  - The C* version implemented as the standard crypt() function
    can crypt 53,000 words pr. second.

The task is embarrassing parallel and is therefore linarly scalable.
So on the largest CM200 (64 K processors) the crypt() function can
crypt 424,000 words pr. second.

At the moment, the Fortran version of UFC can crypt 85,000 words pr.
second on a small CM-5 (128 nodes) with vector units and it is NOT
optimized yet.

Does anyone know who is running the fastest password cracking program
(crypt() function) on a SINGLE machine?

Please do * NOT * respond if your answer cant'be documented or verified!
There are just too many rumors around . . .


Thanks in advance
Jorgen Bo Madsen

+-----------------------------------------------------------------------+
!   Jorgen Bo Madsen,  Security Consultant                              !
!   UNI-C Lyngby,  Danish Computing Centre for Research and Education   !
!   DTH,  Building 305,  DK - 2800 Lyngby,                              !
!   Phone  : [phone]                                              !
!   Telefax: [phone]                                              !
!   E-Mail : [e-mail]                                  !
+-----------------------------------------------------------------------+

------------------------------

From: [e-mail] (Piercarlo Grandi)
Subject: Article 5--AT&T USL vs. BSDI/UCB, Mach3, OSF/1, GNU HURD, Linux
Date: 30 Aug 92 22:59:02 GMT

The USL suit has direr implications that most people realize. The FSF
has decided to abandon the BNR2SS server as it is derived from the
disputed NET2 sources, and CMU have withdrawn the BNR2SS sources
from public distribution. The FSF is considering building a Mach3 server
based on Linux.

Unfortunately, I reckon this is pointless. Let's look at USL claims
in the BSDI/UCB suit:

1) NET2 and thus BSDI/386 contain USL copyrighted text, or its
   derivative.

2) NET2 and thus BSDI/386 embody USL trade secrets licensed to
   UCB, and by UCB disclosed to BSDI or its employees.

The defense by UCB/BSDI is that:

1) all text copyrighted by USL has been removed from NET2 and thus
   from BSDI/386, and substituted with original text.

2) no USL trade secret was employed in writing the replacement text,
   but only original or publicly known techniques.

The interesting aspect of this matter is that the above points, both
USL's and UCB/BSDI's, apply *exactly* to Mach3 itself.

The lineage of Mach3 starts from 4.1BSD+Accent IPC, and by a process of
various releases all USL copyrighted text has been removed, and new text
written in its stead, until Mach3 was obtained, just like NET2 was.

Not only that, all the authors of Mach3 have been exposed to the very
same USL trade secrets to which the UCB/BSDI people have been, as both
groups of researchers have worked for (over) ten years on various
releases of USL source text.

It can actually be argued that if USL wins the UCB/BSDI case, then
*all* staff and students of *any* institution or company that has an
USL Unix source license must be presumed to have been exposed to the
trade secrets contained therein, unless it can proven otherwise.

  Doing so requires proving that the sources have been kept secure and all
  people who have accessed them have been logged -- once the license has
  been granted, all staff and students of the licensee are bound by it,
  and the burden of proving that they never had access to the secret text
  is on them, and even if they can I have doubts that it matters).

So, if USL wins the UCB/BSDI suit by having the court accepting its
contentions, then Mach3 most probably falls too as the next step, and
(less probably) so any and all Unix like code developed by staff or
students of any organization with a Unix source license.

It would astonish me if, having set a precedent w.r.t. UCB/NET2/BSDI,
USL were not to use it immediately w.r.t. CMU/Mach3/FSF. By doing so
they would effectively prevent the OSF, with their current staff, from
producing an OSF/n that does not require an USL Unix source license.

So, giving up on BNR2SS buys essentially nothing, if one continues to
use Mach3 as the substratum

The sure way for the FSF to have an unenncumbered OS would be to adopt
Linux, if it can be proven that Linus Benedict never worked or studied
at an organization with a Unix source license, or if he did, that he
never was exposed to it.

Or one could continue to use BNR2SS with Mach3, because they are bound
to stand or fall together.
--
Piercarlo Grandi                   | JaNET: [e-mail]
Dept of CS, University of Wales    | UUCP: ...!mcsun!ukc!aber-cs!pcg
Penglais, Aberystwyth SY23 3BZ, UK | InterNET: [e-mail]

------------------------------

Date: Sat, 22 Aug 1992 09:32:08 EDT
From: Roger Clarke <[e-mail]>
Subject: Article 6--Important N.S.W. Report released

A long-running 'Independent Commission Against Corruption' enquiry in
N.S.W. has finally reported on an investigation into leakage of personal
data to private enquiry agents, and the leading Sydney daily had over 2
large pages devoted to the matter.  Here's the lead article.

Roger Clarke

____________________________________________________________________________

SYDNEY MORNING HERALD
August 13 1992

HUGE TRADE IN PERSONAL FILES
By MALCOLM BROWN

Westpac, National Australia Bank, NRMA Insurance Ltd, Custom Credit and
Citicorp are some of the big names in a damning report by the ICAC
Assistant Commissioner, Mr Adrian Roden, QC, on the unauthorised release of
confidential government information.

Mr Roden found that there was a multi-million-dollar trade in such
information which involved public servants, including police, and private
inquiry agents.

"Information, from a variety of State and Commonwealth government sources
and the private sector has been freely and regularly sold and exchanged for
many years," he said. "NSW public officials have been heavily involved."

Mr Roden heard 446 witnesses in public and private hearings over 168 days
before compiling his 1,300-page report.

Even so, he said, it was necessary to be selective; thousands of private
and commercial inquiry agents had not examined.

Mr Roden found that more than 250 people had participated in the illicit
trade or had contributed to it.

Of these, 155 had engaged in corrupt conduct. A further 101 had engaged in
conduct which allowed, encouraged or caused the occurrence of corrupt
conduct.

Many are NSW and Commonwealth public servants who sold information
collected by the agencies where they work, including the Roads and Traffic
Authority (RTA), police force, Telecom and Sydney County Council.

The Attorney-General, Mr Hannaford, announced that the Director of Public
Prosecutions had set up a task force to consider laying charges against
more than 100 people named in the report.

He said many of the public servants named could expect to lose their jobs
and that the heads of all the government departments involved had been told
to examine the report and take action against those involved.

The Assistant Police Commissioner, Mr Col Cole, confirmed yesterday that
five police officers had been suspended and announced that three task
forces had been set up and computer security upgraded.

Mr Hannaford foreshadowed the introduction of privacy legislation to make
the unauthorised use of confidential information a criminal offence.

The major banks said that they could not condone what their staff had done
but said the staff had believed that they were acting in the best interests
of their employers and the community.

None of the banks was planning to sack staff found to be corrupt although
several said the staff had been counselled or "educated".

Mr Roden said the trade involved banks, insurance companies and other
financial institutions which had provided "a ready market".

The link was provided by private and commercial inquiry agents. With some
banks, codes had been used to conceal the nature of the transactions.

"As they have gone about their corrupt trade, commercial interest has
prevailed over commercial ethics, greed ha~ prevailed over public duty;
laws and regulations designed to protect confidentiality have been
ignored," Mr Roden said.

"Frequently the client, generally an insurance company, bank or other
financial institution, ordered the information from the agent with a full
appreciation of how it was to be obtained.

"The evidence disclosed that in the collection and recovery departments of
a number of those institutions, it has long been standard practice to use
confidential government information . . . as a means of locating debtors."

Some finance and insurance companies had directed agents to keep all
references to the trade off invoices and reports.

"Some even directed that the agents falsely state the source of the
information in their reports," Mr Roden said.

"Some solicitors in private practice have sought and purchased confidential
government information in circumstances in which they must have known that
it could not have been properly obtained."

Mr Kevin Rindfleish, an unlicensed private inquiry agent, had sold
Department of Motor Transport/Roads and Traffic Authority and social
security information "on a large scale". His principal client had been the
ANZ Bank.

A private investigator, Mr Terence John Hancock, and his company, All
Cities Investigations Pty Ltd, had sold confidential government information
to the National Australia Bank and Westpac on a regular basis.

 Two employees of the NAB had used prior contacts to provide the bank with
access to RTA, social security, Australia Post and immigration information.
Between them, the employees also provided silent numbers and information on
electricity consumers.

The Advance Bank had "over a period of years" obtained information
improperly released from the RTA, the Department of Social Security and the
Department of Immigration. The practice was "known and approved at least to
senior management level".

New Zealand Insurance and Manufacturers Mutual had bought confidential
government information from private investigators.

NRMA Insurance Ltd and the Government Insurance Office were "found to have
participated as freely in the illicit trade in confidential government
information as their more commercially orientated competitors".

"Evidence relating to NRMA Insurance Ltd established not only that it
purchased confidential government information through private
investigators, but also that investigators were required to obtain relevant
government information by unauthorised means if they were to retain the
company's work."

Esanda Finance Corporation Ltd had bought confidential information over at
least 23 years. Custom Credit Corporation Ltd which had engaged in the
illicit trade over "many years", had maintained false records to conceal
how it obtained information.

Alston de Zilwa, former underwriter and operations manager of Citicorp Ltd
and later, Toyota Finance Australia Limited's credit operations manager,
had established for each of the two companies a system for obtaining
confidential information.

The companies would seek information directly from employees of the DMA and
RTA and pay a private inquiry agent, Mr Kevin Robinson, who would "launder"
it, then invoice the companies for the corresponding sum.

Mr Roden said that hundreds of thousands of dollars had changed hands in
the trade uncovered. One agent had estimated that he had paid $40,000 to
$50,000 a year for Social Security information alone.

Another had said he received $100,000 over two years for government
information.

Yet another had, according to records, charged a bank $186,000 for "inquiry
services" over a period of 18 months.
____________________________________________________________________________


Simon Davies and Graham Greenleaf know a great deal about these matters;  I
know a bit too, so if there's valuable info in here to support your own
work, let one of us know and we'll track down the refs.  If there's
interest, I could also get the rest of the articles scanned in and put them
on an archive.

Regards

Roger Clarke
Reader in Information Systems
Department of Commerce
Australian National University
Canberra   ACT   2600
Fax:      [phone]  or  249 3942
Email:    [e-mail]
Tel:      [phone]  or  249 3664
Home:     [phone]

------------------------------

Article 7--Fido-Usenet Gateway

Part 1: Need to know about Fido-Usenet gatewaying!
Part 2: Re: Need to know about Fido-Usenet gatewaying!

From: [e-mail] (Mike O'Connor)
Subject: Need to know about Fido-Usenet gatewaying!
Date: 2 Sep 92 14:09:40 GMT

Situation:

I would like to gateway existing local Fidonet echoes into a local
Usenet system.  I would like for this to be a bi-directional gateway,
such that the Usenet people can post and have their posts "make it"
into the Fidonet universe.  I also want mail to work, such that a Fido
person can mail someone that posts from the Usenet easily and vice
versa.  Low-fuss maintenance is a priority.

What I have:

A.  A user with a PC who's more than willing to transmogrify his
echoes in a reasonable fashion such that they make it to the Usenet.

B.  A site that talks UUCP that'd very much prefer to receive the
Fidonet echoes as news batches.

C.  Someone who knows a bit from both the Unix and MS-DOS end who
wants to see this thing happen and is willing to sweat over a hot PC,
gallantly sacrificing time and energy to do the initial setup.  :)

What I think I need:

A.  Software at the MS-DOS end that mutates the Fido echoes we want
into properly-formed UUCP batches in an organized manner.

B.  Software that can talk UUCP to the Unix/Usenet end in an automated
fashion (i.e. when the rest of the Fido transfers are taking place.)

C.  Some idea of the administrative issues that may occur.  I want to
help both the Fido and Usenet communities, not tick them off.

Any assistance would be *greatly* appreciated.  I have directed
followups to news.sysadmin, but if people don't have access to that,
feel free to send me email.  I'd like to get this going relatively
quickly (Labor Day?).  I will post a summary of how I make things
work, *if* I manage to make things work.


                        ...Mike

--
 Michael J. O'Connor           |  Internet:  [e-mail]
 Ford Motor Company, OPEO      |  UUCP:      ...!{backbone}!fmsrl7!mjo
 20000 Rotunda, Bldg. 1-3001   |  Phone:     [phone]
 Dearborn, MI  48121           |  Fax:       [phone]

[[Response]]

From: [e-mail] (Leonard Erickson)
Subject: Re: Need to know about Fido-Usenet gatewaying!
Date: 2 Sep 92 21:41:15 GMT

[e-mail] (Mike O'Connor) writes:

>What I think I need:

>A.  Software at the MS-DOS end that mutates the Fido echoes we want
>into properly-formed UUCP batches in an organized manner.

>B.  Software that can talk UUCP to the Unix/Usenet end in an automated
>fashion (i.e. when the rest of the Fido transfers are taking place.)

Both of the above are handled by UFGATE on the PC.

>C.  Some idea of the administrative issues that may occur.  I want to
>help both the Fido and Usenet communities, not tick them off.

The Fido person will need to have a *long* talk with his NEC (Net Echo
Co-ordinator), and possible his REC and the ZEC (Regin and Zone ECs).
This is to avoid dupes on the Fido side. If the group is being gatewayed
elsewhere, he shouldn't gateway it again, he should get it from the existing
gateway *or* make *very* sure that no one he feeds it to feeds it to a
site that has a feed path connectivity with the other gateway!

Once he has it set up ok, he should put a GUUCP flag in his nodelist
entry. (G= gateway to)

For that matter, you should talk to the existing gateways. Look for
the GUUCP flags!

Similar issues (with different solutions) apply on the Usenet side.

Most of the trouble is due to Fido using hierarchial distribution,
and relying on on restricting crosslinks to suppress dupes, while
Usenet relies on message-IDs. Fido also has message-IDs, but they
have a format the you can't munge a Usenet ID into. And while most
gateway software preserves the Usenet Message-ID as part of the body
of the message, if it gets such a message from another Fido node,
it can't reverse the process (if it did, it'd make forgeries *way*
to easy).

--
Leonard Erickson              [e-mail]
CIS: [70465,203]             [e-mail]
FIDO:   1:105/51     [e-mail]
(The CIS & Fido addresses are preferred)

------------------------------

From: Linus Benedict Torvalds <[e-mail]>
Subject: Article 8--Linux 0.97pl2 Information
Date: 8/23/92

[[ED: Note changes in anonymous FTP sites]]


finger [e-mail]

                  Free UNIX for the 386

The current version of linux is a 0.97pl2, released 92.08.23.  There is
a 0.97.1 rootdisk that should be used with the new versions, but earlier
versions do work, although they may have some minor bugs.

0.97pl2 supports X11r5 and the new gcc-2.1 (and newer) libraries with
multiple shared libs - as well as any old binaries (except the 0.12
version of gdb which used the older ptrace() interface).  It also
contains support for debugging (core-dumping and attach/detach) as well
as profiling: use gcc-2.2.2d for full utilization of all these features.

Linux can be gotten by anonymous ftp from 'nic.funet.fi' (128.214.6.100)
in the directory '/pub/OS/Linux'.  This directory structure contains all
the linux OS- and library-sources, and enough binaries to get going.  To
install linux you still need to know something about unices: it's
relatively straightforward to install, but the documentation sucks raw
eggs, and people with no previous unix experience are going to get very
confused.

There are now a lot of other sites keeping linux archives. The main ones
(as well as the above-mentioned nic.funet.fi) are:
    tsx-11.mit.edu (18.172.1.2):
        directory /pub/linux
    sunsite.unc.edu    (152.2.22.81):
        directory /pub/Linus

(and many additional sites: there are now sites in the uk, japan etc
that carry linux, but I have lost count)

There is also a mailing list set up '[e-mail]'.
To join, mail a request to '[e-mail]'.
It's no use mailing me: I have no actual contact with the mailing-list
(other than being on it, naturally).

There is also a newsgroup that contain linux-related questions and
information: comp.os.linux.

Mail me for more info:

        Linus Torvalds ([e-mail])
        Pietarinkatu 2 A 2
        00140 Helsinki
        Finland

0.97.pl2 has these new features:
 - major mm rewrite: 3GB virtual process size
 - filesystem error reporting corrections
 - minor bugfixes

0.97 has these major new things relative to 0.96
 - select() through the VFS routines
 - easily installable IRQ's
 - bus-mouse driver
 - msdos filesystem (alpha)
 - extended filesystem (alpha)
 - serial line changes (faster, changeable irq's etc)
 - dynamic buffer-cache
 - new and improved SCSI drivers

------------------------------

From: [e-mail] (Bernd Onasch)
Subject: Article 9--Announcement: VMS NEWS v1.24 released...
Date: 3 Sep 92 00:14:50 GMT

Hello alltogether,

VMS NEWS v1.24 is now available to the net...

VMS NEWS is a VAX/VMS full screen orientated NEWSreader supporting the
following network (TCP/IP) implementations:
  * CMU/tek
  * EXOS              (never tested, no site with it available)
  * MultiNet
  * Process Software
  * UCX               (1.x and 2.0 [DEC TCP/IP])
  * Wollongong
  * DECnet object     (tested the one of ANU NEWS 6.0.6)

The client supports various display methods:
  * Numbered  to just show the articles in order they came in
  * Subject   to display the articles sorted by subject line
  * Threaded  to display the articles sorted by threads (e.g. references)
In all cases, VMS NEWS offers a window where you can scroll around to select
the requested newsgroup or article.

Supported callable editors:
  * TECO (g'd old one)
  * EDT
  * EVE (TPU section)
  * EVE (LSEDIT section)

The command handling is VMSlike (DCL style). The display uses SMG$ routines
and is written for terminal usage (no X-Window support - I think that this is
not needed). The POST command should check for all possible problems that might
occur when a new user starts experimentation with it :-). A special "noscreen"
mode for BATCHes is also included. This allows programming of handlers to
save automatically parts of newsgroups etc.

*Grin* Not to forget - VMS NEWS is VMSlike - not UNIXlike - so there are no
compatibility modes to rn or any special handlings for native unix users :-).
The only reminder to rn is a possible conversion of rn-style register files
to VMS NEWS register files (that use much less disk space).
To be really VMSlike, there are HELP, MAIL and PRINT facilities included.

VMS NEWS is available from:
  * MAILserver [e-mail]
    package NEWS_124 - VMSshare'd source
  * FTPserver iraun1.ira.uka.de (129.13.10.90)
    /pub/networks/news/news_1_24.com - VMSshare'd source
  * FTPserver info.rz.uni-ulm.de (134.60.1.125)
    /pub/VMS/communication/news124.zip - VMS zipped source

I offer ONLY the source (that is written entirely in VAX C) and NO executables.
I think sources are better for network software because they avoid trouble with
different libraries of different TCP/IP implementations. The pure execuables
might lead some sysop's to think of traps and hacks.

Greetings, Bernd Onasch

PS: VMS NEWS stands for "Very Mad Students NEWS"
    (no offense to DEC intended)

,-----------------------------------------------------------------------------.
|Bernd Onasch                       _,---._  One half of mine is human [Spock]|
|Informatik Rechnerabteilung [IRA] /       \__,--.I'll never understand humans|
|University of Karlsruhe     __,--/         \     \ No, I'm from Iowa - I only|
|Germany (FRG)              /    |  Uranus   |    / work in outer space [Kirk]|
|   [e-mail]       \     \     ____/----'  Hello computer ?!? [Scott]|
|   PSI%45050365300::ONASCH  `_____\---'   / There shall be no peace...as long|
|   [e-mail]        `-___-' as Kirk lives [Klingon Ambassador]|
`-----------------------------------------------------------------------------'

------------------------------

From: [e-mail] (Ken Germann)
Subject: Article 10--Cardinal 14.4 S/R Fax Modem
Date: 3 Sep 92 14:16:48 GMT

[[ED: I don't usually put in ads but this looks like a great deal!]]

ZEOS International, Ltd.

Product Release with Specs.

Cardinal 14,400 v32bis Internal Modem with v42 bis, MNP, Send & Receive Fax

- High Speed Hayes compatible modem - 57,600 max data throughput.

- Full duplex 14,400 bps to CCITT V32bis standard.

- V42, V42bis, & MNP 1-5 error correction & 4:1, 2:1 data compression.

- FAX: Group II Fax send and receive up to 14,400 data rates.

- Both Class 1 and Class 2 fax command set support.

- PC/XT/AT compatible half card. Assignable COM 1-4. IRQ 2-5.

- 2 telephone jacks.

- DSP Signal Processing

- 16550 UART

- Rockell Data Pump.

- Quicklink for Windows/DOS Fax and Data Communications.

***     195.00 U.S. Dollars.
--
Ken Germann                    ZEOS International, Ltd.
[e-mail]   INET                Technical Support
uunet!zeos!support UUCP                530 5th Ave NW
612-633-4607        FAX                St. Paul, MN 55112

------------------------------

From: [e-mail]
Subject: Article 11--Microsoft Windows NT Device Driver Developers Conference
Date: 31 Aug 92 21:28:04 GMT

Microsoft Windows NT Device Driver Developers Conference

Monday, October 26 to Wednesday, October 28, 1992
Anaheim, California

Location: Disneyland Hotel
Anaheim, California
Telephone: [phone]

Sponsor: Microsoft Corporation

The purpose of this conference is to provide information and understanding
on the model to support various devices, such as displays, printers,
network adapters, SCSI peripherals, audio products, and more, under
Windows NT. The presentations will discuss the entire structure and design
of Windows NT as it relates to developing drivers to support various
devices. The pre-liminary Windows NT Device Driver Kit is included with
registration of this event and includes sample source code for many
drivers, testing and debugging tools (the Win32 Software Development Kit
is required for actual device driver development and is not included with
registration)

At the Windows NT Device Driver Developers Conference, the developers and
designers of Windows NT will present details on various components of the
operating system like the Windows NT Kernel and the I/O Subsystem and how
they impact device drivers. Further, you'll have numerous opportunities to
discover technical "how-tos" and the latest development tips for a broad
range of devices including:

    * Printers            * Communication devices
    * Keyboards           * Mouse/Pointing devices
    * Display adapters    * Audio
    * Network adapters    * SCSI devices
    * 3270 adapters       * Fax
    * Scanners            * and other devices

This is a great opportunity to hear directly from key developers,
including David Cutler, Architect and Director of Windows NT Development.

And, so that you can be among the first to have your hardware take
advantage of this powerful operating system, we'll give you a free copy of
the preliminary Windows NT Device Driver Kit (DDK) and successive updates
include the final version, a $500 value.

Windows NT's architecture provides a 32-bit device driver model with great
functionality, security, and portability across different hardware
platforms - Intel, MIPS, and DEC Alpha. Imagine how users could benefit,
and you could profit, from your hardware products running under the
Windows NT operating system on not only the huge installed base of PCs but
on workstations using high performance RISC processors or multi-processor
systems. This hardware support will enable PCs to meet the increasing
end-user demands for varied functionality such as fast, high resolution
graphics and integrated networking.

Some of the speakers are the same as those who presented at the Win32 PDC
in San Francisco, but all of the talks at this conference will be focused
on hardware support and the particular needs of device driver writers. The
second and third day will be made up of more than 35 breakout and general
sessions containing completely new material focused on the specific issues
of the devices listed above.

To order a preliminary Win32 Software Development Kit (SDK) for Windows
NT, please contact [phone] for specific details. The preliminary
Win32 SDK has a retail price of $69 for a CD only option, and $399 for an
option including the CD and hard copy documentation. These prices include
follow on updates to this product including the final release.

-------------------------------
Preliminary Program

Monday, October 26, 1992
    Keynote: Microsoft Operating Systems - Cameron Myhrvold, Director,
        Developer Relations
    Windows Family Demo
    Windows NT User Mode Architecture
    Windows NT Graphics Engine Architecture
    Lunch
    Windows NT Executive
    Windows NT I/O Subsystem
    Device Driver Architecture
    Windows NT Software Development Kit
    Windows NT Device Driver Development Kit
End of Sessions - Day 1

Tuesday, October 27, 1992
    Keynote: Windows NT Kernel Architecture - David Cutler, Director,
        Windows NT Development
    Hardware Abstraction Layer
    Windows NT Registry
    Lunch
    Development and Debugging Tools
    Performance Tools
    Setup for Device Drivers
    Performance Counters in Device Drivers
    Q&A Panel - all speakers
    Breakout Sessions:
        Display and Printers - Device Driver Interface
        Networks - Networking Internals and Overview
        SCSI - SCSI Overview
        Devices - Q&A Panel
End of Sessions - Day 2

Wednesday, October 28, 1992
    Breakout Sessions - all day:
    Track 1
        Displays: Display specific Device Driver Interface and Miniport
            Architecture
        Printers: Printing Architecture
        Networks: Network Driver Interface Specification (NDIS) details
        SCSI: Miniport Architecture
        Devices: COM
    Track 2
        Displays: Miniport Architecture (continued)
        Printers: Driver details (example postscript)
        Networks: NDIS details (continued)
        SCSI: Class drivers
        Devices: Mouse drivers - base and Windows
    Lunch
    Track 3
        Display: Linear Frame Buffer driver details and VGA driver details
        Printers: Print Processors and Print Monitors
        Networks: Driver details (example Sonic)
        SCSI: Filter drivers and Testing
        Devices: Multimedia Audio
    Track 4
        Display: VGA driver details (continued) and S3 driver details
        Printers: Halftoning
        Networks: Driver details (continued)
        SCSI: Tape driver details and Tape format
        Devices: Keyboard drivers: base and Windows
    Track 5
        Display: S3 driver details (continued) and Q&A panel
        Printers: Minidriver and Testing
        Networks: Windows NT Streams
        SCSI: Tape and Floppy Tape drivers
        Devices: SNADIS - SNA adapter device support
End of Sessions and Conference

----------------------------------------------------------------------
Conference Registration Form
----------------------------------------------------------------------
Microsoft Windows NT  Device Driver Developers Conference
Disneyland Hotel * Anaheim * California * October 26-28, 1992

BY PHONE:  (800) MS SHOWS or [phone]
           (Monday through Friday, 6:30am to 5:30pm PST)
           International: [phone]

BY MAIL:   Return this completed form and your payment to:
           Microsoft Windows NT Device Driver Developers Conference
           One Microsoft Way
           Department 747
           Redmond, WA  98052-6393

BY FAX:    Return this completed form to:
           (206) 93MSFAX or [phone]
           Attention: Department 747

Conference Fees: $845 per person (postmarked on or before September 21, 1992)
                 $795 per person who attended the Win32 PDC in San Francisco
                     (again postmarked on or before September 21, 1992)
                 $895 late registration (postmarked after September 21, 1992)

Includes: a preliminary Microsoft Windows NT DDK, including updates
              (a $500 value)
          Conference presentations
          Conference guide
          Conference polo shirt
          Notepad and pen
          and more!

Cancellations: Cancellations made by September 21, 1992 will be subject
               to a $90 cancellation fee. No refund for cancellations
               after September 21 or for no-shows. Cancellations must
               be made in writing and you must recieve a cancellation
               number.

Please Check One:
____ Check  enclosed (payable to Microsoft Corporation)
____ Credit Card
       ___ Visa       ___ Mastercard       ___ American Express
Sorry, no purchase orders accepted


________________________________________________________________
(card number)                                  (expiration date)

________________________________________________________________
(authorized signature, must be same as card)

________________________________________________________________
(name)                                                   (title)

________________________________________________________________
(company)

________________________________________________________________
(address)

________________________________________________________________
(city)                                  (state)        (zip code)

________________________________________________________________
(phone)                                        (fax)

----------------------------------------------------------------------

To order a preliminary Win32 Software Development Kit (SDK) for Windows
NT, please contact [phone] for specific details. The preliminary
Win32 SDK has a retail price of $69 for a CD only option, and $399 for an
option including the CD and hard copy documentation. These prices include
follow on updates to this product including the final release.

------------------------------

**********************************
End of Art of Technology Digest #4