HALIFAX AREA
                   PERSONAL COMPUTER SOCIETY

NEWSLETTER                                   FEBRUARY 1995

Meets 4th Sunday of each month, 7:00 pm
     Nova Scotia Community College
      Institute of Technology Campus (NSIT)
     Leeds Street, Halifax, N.S.

26 FEBRUARY MEETING

Topics This Month:
     - Windows Feature - Bill Marchant
     - Question & Answer - Rob MacCara
     - Chebucto FreeNet Update - Clive Bagley and David Potter
  - 4DOS - Norman DeForest


IN THIS MONTH'S ISSUE:

Announcements

Planning and General Meeting Notes

What Might Have Been
Part two of a series on a look back in time

Introducing OS/2Warp: A Rant
A few comments from one member

Caches, Buffers, & Virtual Memory
A quick look at what these are

The Newsletter
A few notes about this document

Meeting Schedule 94/95
Dates for the rest of the year


GENERAL INFORMATION

This document is mailed to all paid up members and to anyone who
has attended a meeting within the past three months.  Yearly
membership dues are $15.00.  The text of the newsletter is
available on ATAB 435-0751, (8N1) and other local bulletin
boards.
Society Mailing Address -
  P.O. Box 29008, Halifax N.S., B3L 4T8
Executive
  Chairperson - David Potter
  Vice-Chair - Andy Cornwall - 861-4503 (h)
  Secretary/Newsletter Editor - Colin Stuart - 461-0370 (h)
  Treasurer - Garth Bennett - 492-3368 (h)
  Disk Librarian - Thayne MacLean - 461-0082 (h)
  and but not least Norman DeForest, Henry Hill, Arthur
  Layton, Rob MacCara, Bill Marchant, George Richards, and
  Diane Smith


ANNOUNCEMENTS

Andy Cornwall has written an article which directly follows this
with the notes from the last planning meeting.  The next planning
meeting will be held on the 6th of March, at 7:30 pm at the
Institute of Technology Campus (NSIT), Leeds Street in Halifax. 
Anyone wishing to have input into what goes into the next meeting
or has some comments on any other Society business is welcome to
attend.

Society Draw
The last draw, for the US Robotics fax/modem, was won by Cyril
Meagher.  The next draw will be for some Income Tax preparation
software.  Tickets will be  available at the next meeting with
the draw being held during the coffee break.  They are two
dollars each or three for five dollars.

Good luck!

Membership Expiry Dates
Please look at the upper right corner of your newsletter mailing
label for the expiry date and renew  if possible.  This is one
way to ensure that you receive a copy of each newsletter and know
what interesting meetings are coming up.

OS/2 Users Group
Meets the third Thursday of each month.  For more information
contact Carey Rolfe at 462-4551 or E-Mail crolfe@fox.nstn.ns.ca

WordPerfect Users Group Meeting
Meetings of the Atlantic Canada WordPerfect User Society are held
on the third Wednesday of the month at 6:30 pm.

For more information contact:

Therese Mackintosh, 435-5456.

Advertising and Want Ads
We don't charge for small individual want ads.  That is any
Society member or other interested person with some computer
related item that they wish to sell, trade, or give away can
contact the editor to place an ad similar to the ones above in
the newsletter.  We would expect that more commercially oriented
advertising provide the Society with some remuneration for
carrying the ad.

An ad will normally only appear once but let me know if you need
it repeated.  Ads can be given to me at meetings or give me a
call two weeks to ten days before the next general meeting
(newsletter deadline).


PLANNING AND GENERAL MEETING NOTES

by Andrew Cornwall

Attending the planning meeting on February 6 were Clive Bagley,
Rob MacCara, Henry Hill, Andrew Cornwall, Garth Bennett, Norman
DeForest, Colin Stuart, and Thayne MacLean.  Amidst the usual
lively discussion the following presentations were scheduled for
the General Meeting on February 26: Bill Marchant providing the
'Windows feature of the month', Norman DeForest discussing 4DOS
(an extended instruction set for DOS), David Potter and Clive
Bagley offering a Chebucto Freenet up-date, and Rob MacCara
leading a question and answer session.  Thayne MacLean will be
preparing a 4DOS library disk for members who are inspired to try
out this popular addition to the DOS command structure.

A meeting topic which has been proposed at planning sessions
quite frequently is computer programming.  In the early days of
microcomputers programming was a common skill among microcomputer
owners.  Now, a wide availability of 'canned programs' to do
almost anything has reduced the need for microcomputerists to be
able to write their own programs.  As David Potter noted several
meetings ago, however, aspects of programming are still commonly
practised in the form of macros and script languages built into
word processors, spreadsheets, communications packages, etc. 
According to the survey of Society member interests conducted in
the fall there is still a strong interest in learning about
traditional computer programming.

Previous planning meetings have grappled with how to structure a
general meeting topic on programming.  This is a difficult topic
to present because there are at least a dozen different computer
languages, programming environments are often obscure to use and
demonstrate, and, let's face it, the details of programming that
inspire programmers may not be a lot of interest to other
mortals.  A good approach to popularize any arcane topic,
however, is to Hold A Contest.  This way even people without a
clue about what's going on can emotionally participate in the joy
of victory and agony of defeat.

Therefore, in the March general meeting we are proposing holding
a Contest Of Computer Programs written by members and guests.
Contestants will be asked to bring their own programs to
demonstrate and a sample software listing to briefly show how it
was written in source code.  We will have a laptop and overhead
projector panel for contestants to run their programs and look at
a sample program listing (which can be an ascii file).  We will
be able run QBASIC and earlier MS-DOS Microsoft BASIC programs,
otherwise it is recommended that contestants bring their programs
in executable form and include any runtime modules or DLL files.


The common theme of all the programs to be judged is to play an
interactive game of Tic-Tac-Toe, where the machine plays against
a human opponent.  This common theme will allow the members to
see how different languages and techniques are used to achieve a
similar purpose.  Spiritually rewarding if not materially
enriching prizes will be awarded to programs in the following
categories: Smallest (i.e. shortest source code listing), Most
Elegant, Most Entertaining.


WHAT MIGHT HAVE BEEN - PART TWO

by Garth Bennett

Lets continue our discussions of some of the things over the past
10 years in computing with a continuation of our discussion of
the builders of the Osborne Executive.

The Osborne Computer Company (OCC) had a huge following from the
Osborne I and now with this faster and bigger machine the world
should be their oyster.  Problem!!  Adam decides to tell the
world that the Osborne III was on the way with full IBM
compatibility.  This was back in the days of Super Vapour
Product.  Announce first and then start the development.  The
problem was that everyone stopped buying and the @#&* hit the fan
at OCC.

Think of what might have happened to this company with a more
forethought.  The CPM80 machine with CPM 3.0 had lots of power
and a tremendous installed base.  The software was mature and
rich in features.  The Osborne Cooperation had a Users Group
which peaked out at 19,000.  These are people who are prepared to
pay 30-70 ($US) per year to learn more about their machine and
software and how they worked.  The Osborne, along with The Kaypro
and some excellent lesser lights, could have become major forces
in the 16 bit world (these machine were 8 bit).  The Kaypro
Company, which produced a similar product with bundled software
(including 2 word processors), lasted until just a year or two
ago but was never able to duplicate the success it had with its
original Kaypro2 and Kaypro10.  The IBM juggernaut was on the
loose and you had to move quickly to stay ahead.

Last year I used my Exec for the last time.  It's still ready and
able to work and I still have a few projects I will do on this
machine.  Over time I purchased a graphics packs that wowed them
with my monthly sales bulletins (how quickly we forget) and an
accounting program which did a great job running my home based
business for a couple of years.  Personal Pearl kept my expenses
on a monthly basis and MBasic served its purpose with many quick
and dirty programs along with a couple that I'm still proud of. 
Along with Grammatik II and SpellStar I was leading edge in a
trailing edge world.

The association with the Users group, called FOG (First Osborne
Group), was a tremendous experience as you got to read letters
and comments from people who are still writing excellent software
today.  People such as Irv Hoff and Ward Christensen (of Zmodem
fame).

After the dust settle the Osborne name was sold to a person in
Montreal who produced some clones and then the name settled into
oblivion.  Adam Osborne, the Osborne whose name we see on the
Osborne McGraw Hill computer Books, went on to found and run
Paperback Software.  He ended up being sued for copyright
infringement by Lotus for the similarity of the interface on his
main product, VP Planner. 

Next time out lets look at the bridge between the 8bit CPM crowd
and the overweight 16 bit machine. (Do I detect a bias??).


INTRODUCING OS/2 WARP: A RANT 

by John Howard Oxley 

IBM's announcement of OS/2 in October is old hat by now, but the
way in which IBM has chosen to market the product is still worth
some heated denunciation.  For those of us who are hard-core
computerists, marketing seems like a trivial part of computing --
what interests us is whether a product works and what it does. 
But unless a product is marketed successfully, it simply won't
survive, and that goes double for operating system software.

If any other HAPCS member was at the October product seminar,
they probably found it as depressing as I did.  The seminar was
rather raggedly organized, and the opening video featuring
Leonard Nimoy (I can just see the marketing types thinking "WARP,
Spock, wow, all those computer nerds will be really blown away by
this) was far too much talk and too little product display.  In
another video IBM seemed to think that a bunch of people in
Alaska all talking up the virtues of OS/2 would be engrossing --
again, I don't want to hear what people SAY about the product --
I want to see it in action.  The presenters were rather junior
types, who did not seem to have a comprehensive grasp of their
product (they were unable to answer my simple question as to how
well WARP supported virtual DOS machines with individualized
drivers, and one of the audience answered instead) or of their
setup system (they could not get one of their machines to work,
and the actual software demos on the other were a bit slow [the
reaction of the room when the presenters said that they were
using their backup machine, a 386 with 4Mb of memory had to be
experienced to be believed] when on later analysis all that was
wrong with the main machine was that a cable was not connected.

All of this left a impression of slapdash, second-rate activity,
which was not dispelled by the high-quality OS/2 pens and the
designer-produced "Post-It" notepad (featuring the OS/2 logo in
full colour) which was on each chair of the presentation.  Again
the lack of drawing power was evident by the number of empty
chairs -- if this had been a Microsoft presentation, the place
would have been packed, and the joint would be jumping.  Say what
you will about the ninjas from Redmond, they really know how to
put on a show, compared to which the OS/2 presentation was
flaccid, tepid, and not particularly compelling.  Let me be
clear: IBM wasn't doing this on the cheap, and all the technical
materials were first-class in terms of quality -- what was
lacking was CONTENT.

But if that was not enough (though one thing I am trying to
suggest is that in performance terms it wasn't just that IBM was
having a bad day with its presenters -- better personnel and
fully functioning equipment still couldn't save an essentially
bland production), IBM's pricing policy is even worse.  When I
bought OS/2 2.1 FOR WINDOWS [which I have never been able to make
work, by the way] it cost me about $75.00 -- the upgrade will
cost $105.00 (even when mitigated by a $25.00 off coupon, the
upgrade still costs more than the original).  Products which have
upgrade costs greater than their initial purchase price face a
lot of resistance from this consumer, and any rate (which is why
I have never upgraded to ACCESS 2.0).  When I pointed this out to
the presenter, their response was "look at what you are getting".

"What you are getting" is in my opinion a nearly brain-dead
marketing ploy, because your upgrade cost includes an IBM
productivity bundle" [remember these are the folks which made
bundling an awesome industrial strategy].  Instead of applets
like a calculator and a notepad, you get a full-fledged
OS/2-capable spreadsheet and word processor, of a quality and
power, I gather, equivalent to Microsoft Works.  This is supposed
to justify the additional cost of the operating system upgrade. 
But if you're like me, you ALREADY HAVE a spreadsheet and a word
processor that you have spend long, sweat-soaked weeks and months
learning how to use -- I certainly don't want to have to learn
how to use another, I want the operating system to be able to run
the one I already have.  There are (according to industry
estimates) about 60 million WINDOWS 3.1 users out there, 99.9% of
whom are foaming at the mouth for release from GPF errors and GDI
memory limitations [I may exaggerate this percentage a tad]. 
This is the market IBM has to grab if they are going to enjoy any
success with OS/2, and forcing bundled applications on these
people is not the way to win friends and influence.

But wait!  That's not all.  I asked the presenter (who probably
was ready and willing to shoot me by that point) whether OS/2 WFW
supported WINDOWS FOR WORKGROUPS peer-to-peer networking. Nope. 
It supports WiFWo on a stand-alone basis, but you can't use any
of the Microsoft product's networking tools.  So where does that
leave those of us who are running little home networks (O, give
me a home, where the EtherNet roams, and the disk drives are
humming all day...) as far as OS/2 WFW is concerned?  Just
nowhere -- peer-to-peer networking is supposed to be announced
in January 1995, although I have not heard anything yet, and that
too will cost extra.

As a final sting, for those of us who paid for OS/2 For Windows
2.1, there is no discount for upgrading (this was offered
initially, and was later rescinded).  This is the sort of
cheese-paring which simply turns prospective loyal users into
people more interested in Windows NT 3.5 -- the upgrade reduction
needn't be much, but not offering one is nearly insulting.

All of this moves me to outraged sorrow, because from the
technical point of view [even if I can't make it work] there is
just no question that OS/2 WARP FOR WINDOWS is simply a superb
operating system.  It is offering now all the functionality that
WINDOWS 95 is promising (sometime early in 1996, no doubt),
including a full InterNet toolset -- just dial your local service
provider and off you go.  It also offers a vastly superior
capacity to run separate virtual Windows 3.1 machines, so one
application failure won't wipe the whole workspace out, which
WINDOWS 95 emphatically does NOT offer, and enough configuration
options to keep you so busy you won't notice if there is a hockey
season or not.  But OS/2 is going to remain a niche product at
best, because with nearly a year's lead time in a well-defined
market, IBM has already failed at marketing the product, right
out of the gate.

No wonder Bill Gates ditched IBM years ago -- the first rats to
desert the sinking ship are always the smartest.  The fact that a
technically better computer product with great potential to place
real, controllable power in the user's hands so the hardware can
do all of which it is capable will fail over such a trivial issue
is enough to make you question progress.


Biblical Bits #17
-- Hard Disks? --
he wrote on stones a copy of law of Moses  JOSHUAH 8:32


CACHES, BUFFERS AND VIRTUAL MEMORY

by Bill Marchant

This month's subject comes about because some of the questions I
hear concerning these items seems to show a large measure of
confusion regarding each of them, and what its purpose is. 
Caches and Buffers have the overall purpose of helping to
overcome differences in speed between various parts of the
computer.  Virtual Memory is disk space used as temporary memory.

Let's start with caches.  Caches come in three kinds.  These are
internal, external and disk.  An internal cache is located inside
the CPU.  CPUs frequently operate at speeds of 100 Mcs.  This
means that a single clock cycle of the CPU will occupy one one-hundred 
millionth of a second, or one one-hundredth of a micro
second.  Not all commands are executed in a single clock cycle,
but even those which take several cycles are executing very
fast indeed.  Contrast this speed with the access time for the
fastest RAM; which is typically in the order of 70 or 80 nano
seconds.  The difference could be as much as five times as fast
which means that if no other steps were taken, the CPU would
spend most of its time waiting for memory accesses.  

Introduce the internal cache.  486 and Pentium CPUs have either 8
Kbytes or 16 Kbytes of special fast memory included as part of
the chip.  When a program is being executed, chunks  of  it are
brought into the internal cache all at once, and are then made
available to the chip at speeds nearer to the CPU clock speed. 
Of course, such memory is rather expensive, which helps to
account for the higher cost of the chip.  And at least part of
its speed is accounted for by the fact that it is so close to the
rest of the CPU, that time is not wasted moving the data through
the wires, as would occur if the cache was farther away.  The
program which controls this cache is a part of the code included
in the CPU itself.  It is not under any kind of user influence.

External caches, fill a similar role of equalizing speeds between
the I/O cycles of the CPU, and the RAM memory.  External caches
are outside the CPU, but are separate from normal RAM.  They are
special super fast memory, and are not counted as part of normal
RAM.  Thus a computer may be advertised as having 128 Kbytes of
external cache.  That means that when a program is running, the
computer loads a group of code into the external cache ready to
be fed to the chip (into the internal cache) to keep the CPU
running at its most effective speed.  The program which controls
the external cache is a part of the BIOS (basic input output
system) of the computer.  If the machine is built to take
advantage of  external cache, it is unlikely that the amount can
be changed unless the BIOS of the computer is also changed.  

A disk cache, is a section of normal RAM which is reserved for
special use as a cache.  In this case, the speed differential is
between the 70 or 80 nanosecond speed of normal RAM, and the 10
or 15 Millisecond speed of a fast hard disk.  Again, the
difference is something like four or five million times, so there
is much to be gained by reading chunks of data or program from
the disk, and making it available at RAM speeds.  SMARTDRV is a
typical disk cache program available with both DOS 6.0 and with
WINDOWS.  It is loaded as part of the CONFIG.SYS file, and is
thus very much under the control of the user.  The amount of RAM
dedicated to SMARTDRV can be specified, and if not specified,
will be between 1 and 2 Mbytes, by default depending on available
memory in the computer.  The RAM used for this cache becomes
unavailable for other uses while SMARTDRV is running, but most
users consider this to be a valid use for memory beyond 1 Mbyte.

One of the features of all these caches, is the ability to cache
data going both ways.  That is both when reading data from the
slower to the faster medium, and also when writing data from the
faster medium back to the slower medium.  Caching both ways is
efficient, and is recommended by all the experts, but it has its
down side.  Only in the case of the disk cache are we normally
given any control.  When SMARTDRV is in use, the write cache
feature can be disabled.  This may be done to prevent loss of
data in the event that a power failure catches you with a cache
full of data which has not yet been returned to the disk.  In
practice, the risk is minimal, since the delay is seldom more
than a second, and the benefits outweigh the risk, but the option
is yours.  The use of the command SMARTDRV at the DOS prompt will
reveal the status of the cache in your computer.  

Buffers are small caches which go all the way back to the first
versions of DOS.  They are 512 bytes in size, and the number used
can be specified in the CONFIG.SYS file.  Application programs
used to specify how many buffers they would require in order to
run  efficiently.  With the advent of SMARTDRV, and newer CPUs, 
buffers are less important than they used to be.  If the programs
you use do not specify the number of buffers required, you may
wish to use none.  Leaving the BUFFERS= line out of the
CONFIG.SYS file will not achieve this however because DOS will
use a default value depending on available memory.  To have no
buffers, include the line BUFFERS=0 in CONFIG.SYS.  If you
experience problems, try putting a few buffers back in.  Use
BUFFERS=2 or 4.  Too many buffers can work at cross purposes with
disk cache programs, and actually slow the system down.  The
problem with modern computers is that they are so fast you might
not even notice.  So if what you have now is working, it is
probably best to leave it alone.

Virtual Memory is a little different from buffers and caches, but
I have included it in this article partly to fill out the space,
but also because it has some kinship with use of RAM as disk
space.  That is, it is the opposite.  It is the use of disk space
a temporary RAM.  This is an old concept, and has its origins in
main frame computers from many years ago.  Computers with very
limited memory were able to run very large programs by swapping
data and program in and out from disk to memory and back again. 
Every program, thus had virtually unlimited memory (governed only
by the size of the disk system in use) i.e. Virtual Memory.

In this era of Windows and vast extended memories, it is easy to
forget that there are still systems in use which are restricted
to the 1 Mbyte limit of the 8088 CPUs.  Virtual Memory is the
answer.  In fact, Windows is itself an example of a program which
uses virtual memory.  The windows SWAP file is temporary storage
for windows applications which are too large to all be contained
in RAM at the same time.  Of course it is slower than RAM, but it
makes the otherwise impossible, possible.  I am aware of one
application besides windows, which allows the user to specify
Virtual Memory.  The shareware spreadsheet As-Easy-As permits
disk space to be designated as memory, thus allowing spreadsheets
of several megabytes in size to be run in computers with no more
than 640 Kbytes of RAM.


ABOUT THE NEWSLETTER

Newsletter Articles
We are almost always in need of good articles.  If anyone has
something that they feel would make a good article, an
interesting story to tell, or even a good meeting topic, please
don't hesitate to pass it on.

Articles can be submitted in almost any format, ASCII text, AMI
Pro, MS Word, Windows Write, and of course WordPerfect.  We look
forward to hearing from other members who have a good story to
tell or come across something that may interest others.

How does one get an article to me?  Various ways are available. 
You can bring it to a meeting or give me a call and upload it to
me.  If you are on the Chebucto FreeNet you can send something to
me via the internet (or have your own account) to where I work -
hlfxtrad.educ.stuartce@gov.ns.ca

It does work, but if you are sending a file attachment to your
message, it should be UUencoded and not a mime attachment.

Newsletter Production Notes
For those who may be interested, the newsletter was formatted
this month with WordPerfect for Windows 6.0a and 6.1.  Some of
the clipart is from Novell (formerly WordPerfect) Presentation
3.0.

The original was printed at 600 dots per inch resolution on a HP
Laserjet 4M.  If I don't have access to this printer then we
print it on an Okidata 850 at Bits and Bytes on Queen Street in
Dartmouth where they allow the Society to print the originals at
no charge.

The main body of the newsletter is set in 10 point Palatino with
the article headings being 14 point bold.  The title on the first
page is ITC Zapf Chancery Medium Italic 19.2 and 16 points.

There were about the same numbers made this month as last,
approximately 95 copies of which around 80 of these were mailed
out.  The extra copies from the previous few months issues will
be brought to the next meeting for those who are new to the group
or may not be in regular attendance.

I do have a complete set of all the previous newsletters and if
someone wanted to look through these, let me know and I can bring
them to the next meeting.




MEETING SCHEDULE - 94/95

The dates for the general meetings for this year are listed
below. 

26 February
26 March
23 April
28 May
25 June

The planning meetings are normally held on the second Monday (8
days) after the general meeting.  They are also located at NSIT,
normally in a small boardroom to the left as you come in.  Anyone
is welcome to assist in the planning of future meetings or
events.

Any changes to the scheduled dates will be announced where
possible at the regular monthly meetings and/or in this
newsletter.