Text content (OCR)
g== ADDIE
Orchara
The Premier Magazine for Apple Computer Users ™
VOLUME 3 NUMBER 4 SEPTEMBER - OCTOBER 1982 $3.25
ET A A TS a Se” — a. eS
| SF I Se | A Te a a
+f tfremaoe Apples fT
ae ae ({Graphics Methods __|
7 Apple /// thr Hunfans ||
a ee ae oe
APPLES and an ROS
FOR THE TEACHER
Allan L. Rogers
ow do you get the most effective
computing power for your
$50,000? That was the happy question
facing us last June when we received
notice from the Chula Vista City
School District that our application
fora Federal grant had been approved
and funded
Our story began in 1979 when
Tiffany Elementary School was di
rected by the Schoo! District to
developa plan fora “Magnet School”.
This would be part of Chula Vista's
plan to implement voluntary integra:
tion by attracting children of all back-
grounds to various special programs
in the community: fine arts, remedial
reading, special programs for gifted
children, bilingual programs, etc.
Tiffany School had been using com:
puters since we soldered together our
first S-100 kit in 1976. We had ac:
quired five Commodore Pets in 1978,
and had been teaching computer
literacy and programming to children
in grades 4 to 6 for some time.
Consequently, it seemed natural that
Tiffany should develop a plan for a
“Computer Magnet Schoo!"
Federal funds were available
through the Emergency School Aid
Act (ESAA) to help school districts in
their efforts to integrate schools, and
so we wrote an application fora grant.
Using the best strategies, we “shot for
the moon”. After re-writing and re-
submitting the proposal under two
separate categories in 1980, we were
hopeful that from our “Cadillac” pro
posal we might receive enough funds
to at least equip us with a “sub-
compact” computer center.
Finally, in June 1981, we learned
that we had received funding. As
expected, the grant request had been
pared down. But as we didn’t expect,
both categories were approved, and
we had our “Cadillac. Not only did
AVAL
we receive $50,000 for the computer
center, but my salary for the next year
as well, so that | could devote full time
to setting up and developing the
magnet school program. What a
happy hacker | was!
Program Outline
Where to begin? First was the se
lection of a computer. Tiffany Ele-
mentary School serves Kindergarten
through Sixth Grade. Whatever we
chose needed to be compatible with
that range of children. We needed a
durable, reliable computer, capable
of producing color and sound, with
plenty of readily available educational
software. We wanted one with
enough flexibility to easily add a va-
riety of peripheral equipment, such as
graphics tablet, light pens, joysticks,
printers, and eventually a hard disk
system. And we wanted a system
expandable to 64K of RAM and able
to run Pascal, Pilot, and the “new
computing language for children,
Logo. Finally, we wanted a computer
which would be compatible with our
older CP/M system. When we wrote
the specifications for competitive bid,
only the Apple met all of our criteria.
Since we wanted to create a com-
puter center which housed all 21
Apples together, we realized we had
more than one choice regarding disk
storage. Although we could have pur
chased a Disk Il for each computer,
there were several problems with that
approach. Expense, the need for mul-
tiple copies of disks, the security of
disks, the carelessness of young
children, and all those mechanical
devices waiting to break, were factors
which intimidated us. The available
local-network systems seemed very
attractive to us, and with the number
of satellites we had, we hoped they
would be cost-effective; what we pur
chased could cost no more than the
cost of 19 or 20 Disk II's, or about
$11,000.
Nestar and Corvus seemed to be
the only contendersin the Summer of
1981 as we were exploring our op-
tions. Nestar and the Corvus Omninet
were both too expensive for our bud:
get, but the Corvus Constellation
seemed a good bet. It allowed up to
64 satellite computers to share the
resources of the Corvus Winchester
disk which was integral to the system.
It promised the best of both worlds;
disk capability without the problems
we faced with individual disk drives.
However, a major flaw we saw with
such a system was its total depen-
dence on the single Winchester drive.
What would we do if it failed? Would
the system be useless until it was
repaired? How reliable were hard
disks? How fast was the turn-around
on repair? These were crucial ques:
tions we felt needed answers before
we could make such an expensive
commitment.
Consequently, when we ordered 20
Apple Il Plus computers in October
September - October 1982 17
1981, the issue of disk drives went
unresolved while we studied these
questions, | proceeded to scrounge
all of the old, crusty, cassette tape
recorders | could find around the
school district, and when our com:
puters arrived in November, | jumped
into teaching computer literacy be
ginning with Third Grade classes.
Although the children loved coming
to the computer center, we all grew
frustrated with the limitations im:
posed on us by the cassette players.
Obtaining disk capability rapidly be
came a high priority.
Meanwhile, a small article titled
The Diskless Apple” in the October
5 issue of InfoWorld had given us our
first clue to the existence of an inex-
pensive local network for the Apple. It
was described as a Icw-cost alterna:
tive to the sharing of disk drives by up
to 128 Apples. Eureka! Well, maybe.
The system was what we were looking
for, but what we couldn't find was the
manufacturer. The name ' Softworks’
was mentioned, along with a cryptic
reference to “ROS' but there was no
clue as to their location.
We were frustrated because we felt
that we couldn't make a decision until
we evaluated this new product. We
finally discovered that there was a
Softworks” in Phoenix; a call to602 -
555 - 1212 gave us a phone number,
and we made our first call regarding
the “Remote Operating System”
(aha!),
Bob Gabriel and Bob Benton of
Softworks helpfully answered my
questions and, at my request, sent a
18 Apple Orchard
system manual for ROS to help me in
my evaluation of the network. The
manual proved to be attractive, sim
ple, and informative. It even had an
index and a glossary of computer
terms in the back, presumably for
people who were new to the com
puting business.
The ROS System
ROS seemed to bea well-designed
concept. There are two kinds of inter-
face boards required for the network:
one “Central” board, and as many
Remote” boards as are needed for
the remaining Apple computers. The
Central board is installed in Slot 2 of
the “Central” Apple. This Apple acts
as the “host” computer, performing
all of the physical disk accesses and
arbitrating the demands for disk use
among all of the “remote” compu
ters. During periods of intense acti
vity, the Central unit would be a dedi-
cated host. At other times it can be
used for other purposes and even, if
needed, be disconnected from the
network.
A Remote card is installed in Slot 7
of each remote Apple. The remotes
are connected serially to one another
by 10-wire ribbon cables in a “daisy
chain", with only one connection
made to the central unit. ROS is
described as an active network; un:
like passive networks, ifthe power ata
remote is off or a cable is removed,
the physical link between central and
all remotes beyond the bad link will be
lost.
Installation and connections app:
eared simple and straightforward
The documentation clearly illustrates
the proper installation, and with the
ribbon cable connectors fixed to en-
sure proper orientation of pins and
sockets, the process looked simple
and foolproof.
Each of the interface boards uses a
simple design consisting of only
seven integrated circuits, one of
which is the bootstrap ROM. It
seemed that troubleshooting and re-
Pair would be simple. With several
identical boards and so few IC's, a
malfunctioning IC could be discov-
ered easily by swapping chips with a
good board. We felt that hardware
problems would be negligible.
An important consideration was
cost “Inexpensive” was the word
used in the magazine article. The
Central board with system software
lists at $309, while the remotes list at
$189 each. With the 5 per cent dis-
count Softworks offers for multiple
purchases, we were looking ata com-
plete local network for 21 computers
for less than $4000. ROS was de-
signed to use standard Apple Il disk
drives, so they could be replaced
easily in case of trouble. Further, ROS
was designed to be compatible with
several hard disk drives; additional
storage could be added later. Things
were looking better all the time.
Of course, one consideration notto
be overlooked was performance. The
manual we received described Soft:
works’ claims for ROS. But what did
they actually deliver? In January,
George Wright (Tiffany's Principal)
and I attended a conference at Ari:
zona State University, and experi:
mented with the ROS system. To
make a long story short, we placed an
order for a central and 20 remotes.
When we installed ROS in January
of this year, we found that installation
was indeed simple. The system oper-
ated almost perfectly from the begin-
ning. Of 21 cards, we found one
which failed to operate. A few minutes
of chip-swapping revealed an IC
which had one pin folded under. A
quick straightening and re-insertion
gave us a 100 per cent functioning
system. It has so far proven to be
reliable,
In February we added an XCOMP
ten-megabyte Winchester hard disk
to the system. Installation of the hard
disk consisted of plugging the con-
troller card into an empty slot, run-
ning a simple configuration program,
and loading some of our software
onto it When we fired up ROS, it
didn't mind in the least if we were
reading or writing to a floppy or the
hard disk. Of course, response time
with the hard disk is considerably
faster.
20 Apple Orchard
4
Ribbon cables can connect up to 127 Apples.
Advanced Software
Softworks has improved the opera
ting system; they now call it AROS,
the A for “advanced”. We've been
using AROS Version 2.02+ since
March, and are even happier with it
AROS permits each remote to ac-
cess the central disk with virtually
every standard DOS command, ei-
ther in immediate execution mode or
as a command imbedded in a pro-
gram. The only exception, for good
Teason, is INIT; the remotes neither
need to (or should!) initialize a disk.
We have had children running pro-
grams which use almost all of the
BOS commands, and each has func:
tioned perfectly. (That may not be a
valid test. as many adults have disco
vered after being amazed by skillful
youngsters —PCW).
Additional commands provided by
AROS include KAT, FEED and BYE.
KAT displays the most recent catalog,
which has been stored in memory,
and prevents duplicate disk access.
FEED permits the central system to
send a single program simultaneous-
ly to two or more remote computers.
BYE releases the remote from the
network, protecting users from un
authorized use of their accounts by
others.
Passwords, “tags”, and “drive con-
figurations” are used to allow the
system operator to maintain file se-
curity and assign various levels of
access to remote users. The pass-
word, which a remote user enters
when he boots into AROS, prevents
unauthorized use of the network. It
consists of up to eight characters and
is unique for each separate user. The
tag is a unique three-letter combina:
tion which is used to identify the
owner of user files or programs saved
on a disk. The tag is automatically
appended to a remote user's catalog
entry when the file is written to the
disk. When the remote user catalogs
a disk or volume, he can see, load or
modify only those files which have his
unique tag. Although he may be
Physically sharing a disk with an en-
tire class, the other users’ files will be
invisible to him. Of course, the central
unit can see and access every file.
The drive configuration tables
permit the central operator to control
access to the disk drives and hard
disk volumes. A remote user may be
granted or denied access to a variety
of floppy disk and hard disk volumes,
and any volumes can be protected
from remote access. The configura:
tions are maintained by the AROS
Password utility, and can include
name and reference number of each
user along with class, period or
section.
Operation
The central computer operates on
an interrupt basis. When a remote
accesses AROS, the central compu-
ter halts its processing and services
the request. If several concurrent re-
quests occur, it serves each in turn on
22 Apple Orchard
a “round robin” polled basis. If re-
mote disk requests during any given
time are not too frequent, the central
system can be used as a non-dedica-
ted terminal to perform other pro:
cessing tasks, such as print spooling,
word processing, and program entry,
When a request for service from a
temote arrives, processing halts at
the central to serve the remote. It is
also possible to disconnect the cen-
tral system from the network tempor-
arily so that certain jobs can be com:
pleted without danger of interruption
from a remote computer.
For those times when there is more
intensive disk access occurring,
AROS includes a utility called SPEED,
which buffers up to 96 sectors of data
in the Central memory. If a remote
requests a sector so buffered, the
request is answered from RAM, with-
out accessing a disk The screen
displays the number of requests an:
swered from RAM and from disk,
which helps to select the most fre-
quently-used sectors for RAM stor-
age. This feature demonstrated to us
the importance of such buffering ina
classroom environment where num:
bers of children may be working on
identical or similar activities. We have
often found that more than half of the
Tequests are handled from RAM, thus
reducing disk drive wearand tear, and
greatly improving data transfer time.
Under AROS, each user has access
to both “public” and his own private
files (those saved with the user's
unique three-letter tag). A public file
ey
(including program, text and binary
files) can be loaded and used by all
users sharing that disk or volume. In
order to protect everybody, the re-
mote user may not execute the DOS
commands DELETE, RENAME,
SAVE, BSAVE, LOCK or UNLOCK on
any public file. Public text files can be
written to, however. AROS has a lock-
out feature to prevent simultaneous
writing to the same text file; the first
user to write to a text file retains
exclusive read/write access until he
closes that file.
The FEED utility and TURNKEY
command are also useful. FEED
allows the operator to send the same
Program simultaneously to two or
more remotes. TURNKEY allows the
central operator to designate what will
happen automatically after a remote
logs in. A selected program may be
RUN or BRUN, a catalog may be sent,
or Integer or Applesoft BASIC may be
loaded automatically after the pass-
word is entered.
Although we have been pleased
with AROS and have found it to
function reliably and well, the rose is
not without a few thorns. As would be
the case with any network consisting
of 20 or more remotes sharing a
central disk system, highly disk-inten-
sive sessions can result in noticeable
delays at each remote. This has not
been a problem except with very
young children who do not have the
patience or understanding to wait
quietly while the central computer is
servicing all of the requests.
Perhaps the biggest problem at this
time is that because of their recent
appearance, networks have not yet
become a widely recognized com
modity in the software industry. ROS,
as well as the other networks, require
the standard Apple DOS format to
function. The increasingly common
copy-protected disks, which use non:
standard DOS, will not work on ROS
at this time. Softworks informs us,
however, that they are receiving in
quiries from software producers re
garding ROS, and are working with
some vendors toward compatibility.
As more and more institutions ac
quire networks as an alternative to
disk drive capacity, there will surely be
more licensing agreements for some
of the better or more popular soft
ware. (The survivors are likely to be the
vendors who are least greedy: one
objection to networks by software
pe fewer copies of a p
gram will be purchased for a network
than for a group of standalone
Ww)
In the several months during which
we have been using AROS, we have
had the opportunity to put it through
its paces. As a budget saving alterna
tive to individual disk drives, it effect
ively lives up to its claims. Of course.
as with every other decision in the
purchase of computers and peripher-
als, no single approach meets every
situation or need. Each decision in
volves tradeoffs of capabilities us.
limitations. Numerous factors need
to be considered in making the de
cision to use a local network as op-
posed to separate disk drives. How:
er, if you're considering a network,
certainly the Remote Operating Sys:
tem from Softworks deserves to be
among the systems which you eval
uate to meet your needs.
ople is th
systems. —
e