A reprint from August, 1969 Computers and automation Multi-Access Forum, Page 38 COMPUTER CENSUSES - DISCUSSION
I. From Walter A. Magee, Management Analyst ADP Management Staff Executive Office of the President Bureau of the Budget Washington, D. C. 20503
Your monthly census of computers of U.S. manufacture has proved useful on a number of occasions in comparing statistics on computers in the Federal Government to the market as a whole. In your February, 1969 issue, your census appeared to drop to less than 30,000 from more than 60,000 in previous issues. Needless to say, this radically changed some of my carefully prepared charts and tables. Letters reprinted in your April, 1969 issue show that others shared my reactions.
On piecing together several of your tables, however, it becomes obvious that the difference is not so great after all. The following table shows what I mean: | Category of U.S. Manufacturer | Count | As of | C & A Issue |
| Top ten, installed in U.S. | 29,271 | 9-30-68 | 2-69 |
| Top ten, installed overseas | 15,497 | 12-30-68 | 3-69 |
| Others, installed in U.S. | 1,038 | 12-30-68 | 5-69 |
| ToOTAL | 45,806 | - | - |
| Earlier Total, All U.S. Manu- facturers, U.S. and overseas | 62,800 | 6-15-68 | 7-68 |
I hope I have interpreted and summarized these figures correctly.
I sympathize with your intention to publish only figures that are based on a verified list of installations. At the same time I feel obliged to (gently) chide you and your source, DP Data Corp., for apparently excluding from your lists some of the computers of the largest user, the U.S. Government. The "Inventory of ADP Equipment in the U.S. Government" has been published annually beginning with the 1960 edition. The latest published figures, released through the Government Printing Office, February, 1969, show 4,232 computers on hand as of June 30,1968. (Later unpublished reports list 4,357 computers installed as of December 31, 1968.) A comparison between your published figures and the Federal inventory shows ten instances where the number of Federal computers equals or exceeds the total in your tables. The ten are:
| (1) | (2) | (3) | (4) |
| MFR. & System | C & A 9-68 | C & A 6-68 | FED. INV. 6-68 |
| CDC 3800 | 8 | * | 10 |
| GE 435 | 6 | * | 6 |
| IBM 7080 | 13 | * | 34 |
| IBM 7090 | 4 | * | 17 |
| IBM 7090 I | 10 | * | 47 |
| IBM 7090 II | 6 | * | 7 |
| NCR 390 | 81 | 660 | 131 |
| RCA 3301 | 24 | 75 | 24 |
| RCA 501 | 22 | 96 | 25 |
| UNIVAC 1107 | 8 | 33 | 9 |
*Not listed as a separate figure. (2) from Computers and Automation, Feb. 1969. (3) from Computers and Automation, July, 1968. (4) from "Inventory of ADP Equipment in the U.S. Government," Fiscal Year 1968. (For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, D.C. 20402; price $1.75.) (Count in (2) and (4) excludes overseas installations.)
I realize these are only ten out of well over a hundred systems. Even so, this discrepancy could indicate the omission of many others. Enclosed is a copy of the current Federal Inventory so that you can crank into your census whatever systems are not included. I hope this contains enough information on each system to suit your needs.
II. A dialogue between Jack Harvey, Vice Pres. National Data Systems, Inc. 210 Summit Ave. Montvale, N.J. 07645 - and the Editor of Computers and Automation
a) From Mr. Harvey
If the computer census published in your May, 1969 issue is an attempt to jar loose some hard data from the manufacturers, I wish you good luck. However, it is so patently ridiculous that it is useless. Since the census published in previous years was at least consistent with most other sources, it had real value to me, and was by far the most useful feature in Computers and Automation. Unless the census regains my confidence, I will not renew my subscription.
b) From the Editor
This computer census in the May issue covers a ground no other computer census that we have previously published covers: it reports information on computer installations outside of the United States. Was this clear to you from the description of it? Should we have included some repetitive remarks so as to emphasize this difference?
We are seeking intensely to obtain good computer census information. Even the Central Intelligence Agency office in Cambridge, Mass., telephoned us the other day, asking us for some information about the number of computers in the Soviet Union! In spite of several contacts that I have there, I have not found any good figures.
c) From Mr. Harvey
I must say that the special viewpoint of the May computer census was certainly not communicated to me. The change from the prior policy of a standard monthly census turns the feature into disconnected individual articles, each of which must receive individual evaluation to be useful. With a standardized monthly census, the reader can develop his own correction factors and have confidence in the result.
I would suggest at least a monthly directory of previous special censuses and (hopefully) a schedule of future issues. But continuity and standardization seems very desirable.
III From the Editor
In the Midyear Computer Directory issue of Computers and Automation, we have published a World Computer Census, covering four and a half pages, which represents a careful summary of all the computer census information available to us on June 15, 1969. This is in accordance with the helpful suggestion of Mr. Harvey.
Our present plans are to publish a computer census in every month, publishing either one of two parts alternately: Part 1, reporting on the computers of United States companies; Part 2, reporting on the computers of foreign companies. Each part will be updated every second month. The next computer census will be published in the September issue of Computers and Automation.