Voice
Communication in Business Volume 1
Essays on telecommunications,
1969-1980
Chapter 4
Some Correspondence
During the next few
years, I only wrote a few letters to the editor. Some of the editors
resisted, but here I combine even the unpublished notes into a
chapter for my hopefully enthusiastic public.
During our adventures
with the RFP, Telephony published a special issue on modern
PBXs. In several of the articles, the modern (Series 300) PBX
features were extolled, pointing out how much extra money they could
make for telcos renting such modern equipment. What was never made
clear was that (a) modern PBXs were and are very expensive, (b)
program controlled features can be added at very small incremental
cost, and (c) the presence of many features, some of them novel,
masks the high cost of the PBX and the absence of features formerly
standard. The following letter appeared in Telephony for
March 6, 1972.
***
After reading Telephony's special PABX
issue (Dec. 6, 1971), I must agree with Mr. (author of "The silent
revolution," p. 29) that "... the new generation equipment has been
designed with the needs of the telephone operating company very much
in mind." Unfortunately, however, the user obtains little benefit
except the opportunity to pay more for questionable "... features
that mean increased revenue" to the telephone company.
People in the telephone industry with their
customers' interests at heart might well ask themselves this
question when considering the new PBX equipment: Does this new PBX
provide the user with the services and features he now obtains?
Chances are, the answer is "no." The user
will still have to make extensive use of key systems, restrictors
and other separate black boxes. Why? Because the telephone industry
is notorious for subdividing itself into non-communicating segments
such as "transmission," "switching" and "station apparatus" and, as
far as I can tell, has never taken the trouble to get these separate
groups to look together at PBX requirements from the user's
viewpoint. It is small wonder that consultation hold, call waiting,
abbreviated dialing and other features newly discovered by PBX
designers are second-rate copies of features already supplied in
more useful form in such station apparatus as key telephone systems.
Further, key systems are still required, even with the most "modern"
PBX, because a secretary must pick up her boss' line every now and
then, consultation with non-PBX users is often necessary, etc. In
larger systems where tandem switching is needed, it is fascinating
to see the number of four-wire term sets required to make tie-lines,
all four-wire facilities, conform to the two-wire Procrustean bed;
and don't count the number of dial tones the innocent user may have
to deal with.
Until designers start considering users'
needs in addition to those of telephone companies, progress will
continue to be marked by "tail fins" such as streamlined consoles,
exotic hardware and other assorted bells and whistles of varying
degrees of frivolity. Messrs. Dittberner and Rojahn's article
("Dramatic outlook for new PABX capabilities in coming decades," p.
60) offers some hope for the next generation of equipment, but the
PBX customer can only wonder why the telephone industry never
considered his existing needs before trying to think up something
new.
***
While our efforts to
find economy in interconnected equipment continued, RCA sent me to
Washington to be on the FCC committee trying to figure' out how
interconnect could be accomplished without causing "harms" to the
public network. I had written RCA's response to the original
National Academy of Sciences report on possible "harms;" it was
described privately as perhaps not the longest but certainly the
funniest.
In principle, I was
representing the RCA Service Company, an interconnect vendor.
However, RCA Corporate was a major telephone customer, RCA Alascom
was an Independent Telephone Company, and Government and Commercial
Systems wanted to make PBXs and CO switches for the interconnect and
Independent telephone market. Defining overall company best
interests posed something of a challenge.
The IEEE, which seldom
concerns itself with business communications, ran an article on
interconnect at the end of 1972. By now considering myself an
authority on the subject, I took typewriter in hand to offer a few
comments. Here is the result:
***
Mr. McKenzie's article, "PABX: the Epitome of
'Interconnection,"' in the December, 1972, Spectrum was most
timely, particularly as a vehicle for demonstrating IEEE interest in
business communication systems. This fascinating topic has too long
been excluded from the main stream of electrical engineering, to say
nothing of college curricula. However, in view of the stakes
involved in the young interconnect business, one might have hoped
for a little more than the telephone company party line (if you'll
pardon the pun) and the iteration of the design philosophy of a
small group of circuit engineers whose contact with the needs of the
market place is, to put it kindly, remote.
To illustrate some of the different points of
view in this controversial area, I'd like to discuss briefly why
interconnect came into being in the first place; why many current
PBX design principles, accepted almost completely on faith, are open
to serious question; and how the barrier, the so-called Voice
Connecting Arrangement (VCA), is little more than a rear-guard
action to deter business users in their attempts to obtain the kind
of communication services they need.
One has only to read Aesop's fable about the
dog in the manger to understand the telephone industry's attitude
toward interconnection with privately owned equipment. For many
years they have simply refused to provide needed communication
services themselves and have refused to allow anybody else to
provide them either. Under the circumstances, the only remarkable
thing about the Hush-a-phone and Carterfone decisions is that they
were so long in coming. Other services and features which the
telephone industry has either denied or has grudgingly agreed to
provide in a suboptimal manner include automatic dialers, automatic
answering systems, intelligent calling-range restriction from
business phones and PBXs, traffic recording, cost allocation
information for multi-department customers, answer supervision to
PBXs from calls to the public network (See note 1, p.18), PBX
entrance to the public network at tandem rather than local switches,
4-wire tie-line switching, and perhaps most important, direct inward
dialing to PBX extensions (Note 2).
The right to interconnect one's own PBX or
key telephone system is now established (but again grudgingly) as
long as a VCA (Voice Connecting Arrangement) is used on each line at
the interface. The VCA restriction may soon be lifted, but probably
at the expense of cleverly rearranged tariffs to maintain a high
cost to the interconnect customer. In any event, every effort will
be made by the telephone industry to prevent interconnect suppliers
from providing, or providing in a more satisfactory way, services
which customers want and need but which do not fit in with
preconceived visions of what the telephone company thinks they ought
to have.
The myth about good telephone service being
related to uniform engineering standards needs some amplification.
While undoubtedly some standards do exist, the telephone industry
has, so far, been unable to produce them in written form for
inspection. Although this may be surprising to most engineers, it
has long been obvious to communication managers for interstate
enterprises. A major portion of a communication manager's time is
spent in the frustration of endless hours of meetings and
conferences with the telephone company in state A, trying to
convince them of the theoretical possibility of doing something that
is a standard service offering in state B. Perhaps one of the
greatest advantages the interconnect industry has to offer is an
overall concept of nationwide standards that do not evaporate at
arbitrary political boundaries. Once the state regulatory
commissions, primarily (and quite properly) concerned with
residential telephone service, become aware of the business users'
needs for standardization, it is likely that some encouragement may
be offered the telephone industry to follow the example which the
interconnect people, of necessity, must establish.
An advantage that the telephone industry will
reap from interconnection is the opportunity to find out what
business customers really need without huge expenditures on design
and development efforts centered around unwanted products. For
instance, Bell's relatively new (1967) 800A PBX has already been
rated manufacture-discontinued, presumably because enough people
have discovered the hard way some of its more serious design
limitations; further, there has been something of a customer revolt
against augmenting PBX automatic switching with consoles rather than
cord-type switchboards. Although it may be a bitter pill for the
telephone industry to swallow, customers may really know what they
want and why; if given an opportunity, they may vote with their
dollars in the best traditions of free enterprise, and expensive
decisions from on high about what they ought to want can be
eliminated. An added advantage of this approach is, of course, that
the enormous problems of raising capital, described so feelingly and
at such length in both the public and professional press by the
telephone industry, will be greatly reduced; this, in turn, will
allow scarce and hard-won capital to be devoted to improving the
public network (sorely needed in most metropolitan areas) and
providing better residential service.
Turning now to design principles, Mr.
McKenzie states twice in the course of his article that Step-by-Step
(SXS) systems do not have the ability to "back up," because they do
not have the advantages of common control. This, in turn, makes it
impossible to upgrade most existing PBXs to permit them to provide
new features. This dogma has been preached for years; its only
difficulty is its tenuous relation to fact. Most modern PBXs,
whether of Bell, Independent, or Interconnect supply, do NOT use
their common controls except in a peripheral way to provide the
service features they advertise. In general, most of the features
actually performed by PBXs are built into the trunk circuits which
meet the wires to the local central office, or into other external
equipment. Further, many modern SXS PBXs (such as the Automatic
Electric 301 and 311) provide almost the full range of "series 300"
features with SXS equipment. Indeed, with regard to station dial
transfer, consultation hold and three-way conference, the 301 beats
most common control systems hands down by making these features
available on outgoing calls and tie-line calls in addition to
incoming calls from the local central office. Tie-lines are, of
course, a sore point in Bell's common control PBX designs. The No.
101 ESS will not deal with tie-lines to other PBXs in any reasonable
manner, and one Bell company has bought several foreign-made
crossbar systems to handle tie-line switching for 101 customers. The
101 is an all electronic, common control, stored program system; its
"flexibility" is thus well illustrated. It should be clear from the
above that common control is neither necessary nor sufficient to
insure needed customer features.
Actually, most of the "features" needed by
business customers are provided by key telephone systems. The
ability to pick up more than one line, to hold one call while
answering or making another, etc., are handled in a thoroughly
satisfactory (although expensive) manner with multi-button telephone
sets. Indeed, an analysis of the telephone bill of most medium and
large business customers will show that the monthly expense for key
equipment is greater than the cost for the PBX and the regular
single-line phones combined. The "Series 300" features cost extra
and must be taken on all telephones on the PBX regardless of need;
because they do not replace the features provided by key equipment,
the customer still has to incur extra expense to obtain what by now
is "standard" service.
The ability to "back up" which can, with some
justice, be attributed to certain kinds of common control, is
primarily of interest in automatic alternate routing between very
large switching machines in the public network. Its application to
PBXs is questionable, and the assumption that desirable features can
only be made available in this way shows the depth of
misunderstanding that is rampant at the design level.
What can a common control do? Among other
things, it frees the switches from direct control by the telephone
instrument and allows designers much greater freedom to demonstrate
their virtuosity in manipulating state-of-the art components. While
this freedom is leading to some innovation beneficial to the user,
particularly in the hands of designers for the interconnect market,
too much is being dissipated in meaningless exercises in irrelevant
hardware. Increased power consumption and floor-loading that exceeds
the standards of modern office buildings are among the unfortunate
results.
Freeing switches from subset control allows
such innovations as Touch-tone calling to be introduced; indeed, SXS
systems and certain portions of key telephone systems require the
addition of expensive adaptors if they are to use Touch-tone type
sets. Since replacement of the dial with a tone-signaling pad
doubles the cost of the telephone instrument, use of this telephone
tailfin permits expansion of the telephone rate-base; there are
minor advantages in the reduced holding time of certain central
office equipment, but this equipment now costs more because of the
elaborate circuitry required to detect tone signaling. Tone
signaling can, of course, go end-to-end, making direct communication
with computers possible; I have been told that efforts are underway
to inhibit this feature except for customers who pay extra.
Finally, we come to the Voice Connecting
Arrangement, the toll-gate that extracts the telco tribute from the
customer who connects his own telephone system to the public
network. A brief study of the equipment usually supplied to
implement the service offering designated CDH shows that what would
ordinarily be a pair of wires for each CO trunk is expanded as if by
magic at the customer's location to five pairs (Note 3). This
five-to-one increase is effected by 8 transistors, 3 logic gates, 17
diodes, 7 relays, and a variety of resistors and capacitors. The
only actual protection comes from a repeat coil (transformer to
non-telephone types) and two varistors in the speech path. These
components attenuate longitudinal induction, block possible crosses
to foreign potentials, and limit excessive signal levels. They could
do their job just as well in the PBX trunk circuit; unfortunately,
the telephone industry seems to feel that only its technicians can
properly maintain these completely passive non-adjustable circuit
elements.
Transmission is, of course, somewhat degraded
by the repeat coil and varistors; dial pulsing is degraded in
passing through additional circuitry including a relay. In general,
any circuit containing that many components must have some adverse
effect on reliability, particularly when it is compared with a solid
pair of wires; the fact that it is usually powered by an ac-to-dc
converter, and drops any call in progress when a momentary power
failure occurs, is another of its interesting attributes.
But the protection required against signals
outside the voice band, incorrectly adjusted dials, data
transmission (you aren't allowed to send data through this gadget),
etc., etc., are provided by the words of the tariff, not the
hardware. It is easy to understand why the telephone company insists
on this device in every central office trunk from a customer-owned
PBX; its relatively high per-trunk cost often makes the difference
between savings and loss to the customer. The only block to
understanding lies in taking seriously the telco "protection"
rationale.
These, then, are some of the more interesting
facets of interconnection at the beginning of 1973. Although there
are many points of view, there is one point on which I'm sure all
parties will agree: the telephone and interconnect industries are
going to be in for some very lively times during the next few years.
Notes:
-
Answer supervision, while important to
all PBXs, is particularly vital in the hotel-motel business
where call charges must be added to bills on a daily (or more
frequent) basis. Answer supervision to operate hotel message
registers was standard in SXS central offices; few common
control offices offer this feature without extensive additional
hardware.
-
Direct Inward Dialing (DID) can only be
obtained as part of the Centrex package. The other part, totally
unrelated technically, is Identified Outward Dialing (IOD).
Although IOD has been marketed in the past as a sometimes useful
independent feature (QZ billing), users with WATS or tie-line
networks find it a very expensive way to divert calls to more
expensive facilities. (Centrex IOD does not, of course, work on
tie-lines, WATS lines, FX lines, etc.). Under New Jersey
tariffs, IOD is guaranteed neither to work nor to be accurate;
nevertheless, the user must take IOD to get DID. DID could be
implemented with negligible additional equipment at the PBX and
the central office; IOD, however, when carried out
automatically, requires extensive modifications and equipment
additions at both locations. IOD only works on toll calls; I
know one Centrex installation where the customer, in an effort
to stop escalating toll charges, is renting an expensive
restrictor circuit from the telephone company to prevent toll
calls from being placed over IOD trunks. The telephone company
seems to be happy in this imaginative inflation of its rate
base.
Until you've seen Centrex CO, however, "you ain't seen nothin'
yet!" Wires to the central office which does the PBX switching
increase by a factor of 10 and a very complex data link is
needed for the customer's switchboard attendants. Since a
console is provided, more switching equipment is needed than a
cord-board would require, and the customer's flexibility is
greatly reduced. One Centrex CO installation I know adds a plain
extension phone for the attendant so that she can do some of the
things that a console prevents. With Centrex CO, the customer
saves some floor-space but, due to the requirements of key
telephone systems, it is seldom as much as is advertised by the
telco
One of the principal advantages of Centrex CO is the way new CO
equipment and buildings can be justified as a replacement for
more satisfactory SXS equipment which has already been written
off.
-
Mr. McKenzie missed one pair, but not all
are used at every installation.
***
Naturally, the IEEE
declined to run my letter. However, because the issue seemed
important to me, I used the FCC committee mailing list as a private
citizen to send out the following:
***
Although the work of the Procedures and
Enforcement Subcommittee has been completed, there are certain
facets of the monopoly issue that might well be discussed at this
time. While the "monopoly is evil" argument, in general, is not
applicable to regulated public utilities, the monopoly of
information (particularly with regard to telephone switching) that
has fallen by default on the telephone industry is, in my opinion,
harmful to the development of the art. No publicly available
education, so far as I know, offers courses on telephone switch
design or system planning; as a result, almost everyone working in
the field, even when serving public utility commissions or the
interconnect industry, has been trained by the telephone industry
and is thus dependent on "basic principles" which may or may not be
true. Since these "principles" do, in fact, rationalize and defend
the equipment which forms much of the rate base of telephone
utilities, independent evaluation of existing or future designs is
difficult. As this letter will show, entering a dissenting opinion
into the professional literature is almost impossible; fortunately,
the profit-making trade press is more interested in giving "equal
time." Perhaps a lesson can be learned from this.
For the past several years, I have been
engaged in attempting to specify and buy customer-owned PBXs to
serve large industrial complexes. After a careful study of the
services and features actually paid for by the users in question, an
associate and I sent out a Request for Proposal to see if suitable
equipment was available. Our study showed several requirements that
were somewhat surprising; feeling that the industry might be
interested in these results, we prepared a technical paper with the
hope of presenting it at the International Conference on
Communications or the National Electronics Conference. The paper was
rejected, as shown in Attachment A. Even though the IEEE turned us
down, Communications News published the article about a year later
in the February, 1972, issue. Many users and people in the
interconnect industry have commented favorably; unfortunately,
indexes of the professional literature do not include Communications
News; young designers, carrying out a literature search, will never
find this input from a real, live customer.
As a second example, the IEEE Spectrum
published, in the December, 1972, issue, an article entitled "PABX,
the Epitome of 'Interconnection.” This article is available in
reprint form from the IEEE for $1.50, and should be studied
carefully by the interconnect industry. The author, an authority on
radio, did a fine job assembling and presenting information
available to him; unfortunately, he was offered only the telephone
company point of view. As an IEEE member of 25 years standing, a PE,
and a telephone engineer, I wrote a letter to Spectrum
suggesting some different points of view, particularly with regard
to design philosophy. A copy of my letter is included as Attachment
B. The IEEE acknowledged receipt of the letter, but after several
weeks, when I called to check on it, I was informed that the letter
has been "lost." I sent a second copy and, after several telephone
conversations over a period of two or three months, received
Attachment C rejecting my letter. Perhaps the overly-vivid writing
style I chose to use was a factor in the rejection; but once again
the "party line" has gone unchallenged.
Where does all this lead? It leads to
everyone, including members of regulatory commissions and the
interconnect industry, thinking one way—the way of the telephone
company. The true nature of the situation concerning the
interconnection of privately owned communications equipment is
obscured and the interconnect industry is in danger of copying
blindly telco designed equipment, even when that equipment is
completely unsuitable for user needs.
As long as technical documentation is in the
hands of the utilities, everyone involved can be conditioned to
ignore the many interesting and productive alternatives which
privately owned equipment can make possible. The interconnect
industry, if it is to survive, must not only provide equipment to
match telco performance, it must go beyond the limited interests of
the ivory-tower telco designers to create equipment of such
capability and scope that whole new approaches to business
communications can be developed and utilized. If this is to be done,
however, means must be found to insure public discussion of user
needs and design philosophy. The present professional press does not
seem to be a practical forum.
***
Attachment A was a
letter from J. H. Weber of Bell Labs, for the IEEE Communication
Technology Group, rejecting “Voice Communication in Business”
(Chapter 3 in this volume) for presentation at a national IEEE
meeting. The paper contained "many random thoughts which had not
been organized toward developing a central thesis."
Attachment B was my
Spectrum letter itself, and Attachment C was a rejection of that
letter. The IEEE's reviewers (guess who they must have been) felt I
had not covered certain problems in traffic and service, and that my
remarks were more emotional than technical. Well, you can't win 'em
all.
As the above interchange
indicates, it is difficult to explain a technical problem to people
whose knowledge of the subject matter tends to be limited. This
suggests a need for education in business communication. Where there
is a need, a response will sometimes appear. The following letter,
written to a client a few years later, after I had gone into private
practice, reviews a Master's program in Telecommunications at a
major university. I hope, somehow, my objections have been noted and
acted upon. Certainly, in recent years, other schools seem to be
getting interested in somewhat more realistic material.
***
Thanks for letting me look over the Master's
Program in Telecommunications. It was a most interesting experience.
In brief, I cannot imagine any way in which
the proposed program would benefit a member of your
telecommunications group. The course material and orientation seems
to be based more on knowledge available to an average engineering
school faculty than to any needs of the industry or the students.
The approach is theoretical and (although a math prof would scoff)
mathematical; there seems to be no practical information or even any
theory which might facilitate the acquiring of a practical
background on the job. In my opinion, you would be doing your fellow
ICA members a service if you actively opposed continued sponsorship
of this program.
To give you a few examples: Course 651 is
called "Principles of Communication Systems." It covers modulation
theory, information theory, noise theory, etc. None of this has
anything at all to do with systems in the sense that a communication
manager understands. He is not going to design a front-end for a
microwave receiver—he wants to select circuits and terminals in such
a way that he can move signals of a particular class from one point
to another at minimum cost.
Or consider "Probability and Statistics for
Engineers I." Here we hit sample spaces, moment generating
functions, and other academic goodies. No mention is even made of
queuing, or Poisson and Erlang tables. Apparently the student will
learn to derive formulas that have been in use for 75 years, but he
won't have any idea how to estimate operator work-load.
"Technology and Regulation" sounds more like
an orientation to AT&T than a guide to reading and using tariffs
effectively. "Computers and Communication" seems to be related only
to data—doubtless covers ARPA in great detail. The several
broadcasting courses seem totally out of place.
There are no courses on economic evaluation
of proposals, traffic handling systems, preparation of programs for
computer analysis of communication costs; the words "switching" and
"key systems" do not, of course, appear. I can find no indication
that specific equipment of any sort is discussed or analyzed.
To summarize, I find the entire program
totally irrelevant to any form of corporate telecommunications that
I have encountered over the last ten years. Further, I cannot
imagine a context, in design, operation or management, in which this
program might be of value. If you have any inside information on how
it was developed, I'd be very much interested in knowing what was
intended.
[
Top ] [ Next ] [
Table of Contents
] |
