5 June 1973
The ARPA Network seems to have developed the proclivity of dragging compulsive collectors and organizers out of the woodwork and placing them in the forefront to annoy everybody.
Recent occurrences have been:
- A set of charts on characteristics of the hosts. The orientation seems to have been: If you can come up with names for the horizontal and vertical nodes and if it has to do with the hosts, make a chart out of it. This collection of charts goes under the euphemism "ARPA Network handbook". Information on a host is scattered over all the pages which is a questionable organizing scheme. Additionally, since the charts contain much of what is already in the Resource Notebook, we now have the delightful task of maintaining two documents when changes are necessary.
- A telephone call asking for hourly loads on the TX-2 computer for every hour of the months April and May. One can easily imagine all this information being keypunched in some computer (on the network, of course) and then lovely bar graphs, curves, plots, etc., being generated. Probably in triplicate.
- A mailbox message about a "central software repository" and a personnel file. (Copy of the message is attached). This was just too much and is the immediate precursor of this RFC.
My first reaction to the "central software repository" was that this has got to be some kind of prank. But when the second message (identical to the first) arrived an hour later and when I learned that others had also received it, I reluctantly accepted its legitimacy. Actually, sending the message in duplicate fits in very nicely with the general bureaucratic syndrome evidenced by the contents of the message.
This RFC addresses itself merely to the idea of listings of every program. That does not mean that I think that the rest of the request is better, just that I don't have the time to write a treatise on the general subject. It should be noted (if not obvious) that what follows is being written with almost unbearable restraint.
Listings of every program available to network users? Has anybody calculated how much paper would be generated? How many trees would have to be cut down for this paper? How many filing cabinets are going to be needed? How is this massive amount of information in its totality going to be of use anyone? Is there going to be an answering service which will answer such questions as to what is on the third line of page 5 of the listings of the editor at a given host? Will one be "required" to send a new listing in order to change a program?
This material has not been reviewed for public release and is intended only for use with the ARPA network. It should not be quoted or cited in any publication not related to the ARPA network.
plf-2256 From the point of view of a site such as TX-2, the questions become even more intriguing. Many of our programs are written in assembly language. Should we, therefore, also send along a copy of our (incomplete) assembly language manual? Or should we drop everything else and complete the manual? What about listings of our operating system since the programs make calls upon the system for input- output, file management, etc.? (I could go on and on, but the readers should get the idea by now.) Much of this applies to any host, but for a host which has a one-and-only computer,the problems are more acute.
Once again, may I repeat my plea from RFC 459. There are small research sites on the network. TX-2 is one of them. Please, network community, don't drown us in a sea of make-work. We might get nothing done just keeping up with it. Or is that no longer important?
In particular, the network community ought to be glad that in the mid 1960's we at TX-2 weren't bombarded with tons of make-work and were able to get something done. What I have in mind is the initial experimentation with a small-scale network prototype with SDC which demonstrated the feasibility of networks and led to the ARPA Network. (Please see reference.) Who knows what we, or some other site, will come up with if given the chance?
Some people have suggested that I not write this RFC reasoning that if I just ignore it, the problem will go away. But the problem is not going away. If anything, it seems to be getting worse. Silence becomes in effect tacit consent. I do not intend to sit by and sacrifice useful work to satisfy bureaucratic compulsiveness.
It says someplace that one should end on a positive note. OK, here goes: May I respectfully suggest that the next potential perpetrator of network (make) work for someone else think very hard about the justification for it. Also, think about how much time it will take, remembering that not everybody is as fast or brilliant as you are. If you would like positive responses from others, you should consider why someone else should feel motivated to do the work you request. Then put all this down on paper as the introduction to the "work order." Then think some more. Try it on some colleagues. If it has still survived then maybe you have something. Just maybe.
T. Marill and L. Roberts, "Toward a Cooperative Network of Time- Shared Computers" in AFIPS Conference Proceedings, November 1966.
NET MAIL FROM SITE USC-ISI RCVD AT 30-MAY-73 10:59:44 ------- DATE 30-MAY-73 0740-PDT FROM RML AT USC-ISI RE: RML CENTRAL SOFTWARE REPOSITORY - - - -
RML CENTRAL SOFTWARE REPOSITORY
RML IN THE CAPACITY OF ARPANET MANAGER IS INTERESTED IN ESTABLISHING AT RML A CENTRAL REPOSITORY OF PROGRAMS ADVERTISED IN THE NETWORK RESOURCES NOTEBOOK BY THE HOST SERVER SITES AS AVAILABLE FOR USE BY NETWORK MEMBERS. IT IS ALSO DESIRED THAT PROGRAMS GENERALLY AVAILABLE FOR USE BY NETWORK MEMBERS BUT NOT LISTED IN THE RESOURCE NOTEBOOK ALSO BE INCLUDED. AVAILABLE DOCUMENTATION ON THE PROGRAMS IS ALSO REQUIRED. THE TYPE OF PROGRAM DOCUMENTATION DESIRED INCLUDES BUT IS NOT LIMITED TO - 1. PROGRAM DESCRIPTION 2. LISTINGS 3. RUNNING INSTRUCTIONS A. OPERATION INSTRUCTIONS B. INSTRUCTIONS FOR THE PREPARATION OF THE DATA TO BE PROCESSED 4. PROGRAM LIMITATIONS 5. ANY OTHER AVAILABLE DOCUMENTATION IN THE ABSENCE OF THE ABOVE. YOUR COOPERATION IS THEREFORE SOLICITED IN PROVIDING COPIES OF THOSE PROGRAMS WITH THE ASSOCIATED DOCUMENTATION ADVERTISED BY YOUR SITE AS AVAILABLE FOR USE BY NETWORK. IF THERE IS A CHARGE FOR THE MATERIAL PLEASE PROVIDE THAT INFORMATION BEFORE INITIATING ANY ACTION. IN THOSE CASES WHERE THE PROGRAM RESIDES AT A GIVEN HOST SITE AND THE DOCUMENTATION IS LOCATED ELSEWHERE SIMPLY PROVIDE THE LOCATION INFORMATION. RML IS ALSO ESTABLISHING A FILE OF HOST SITE PERSONNEL OR STAFF INTERESTED IN OR POSSESSING PARTICULAR TECHNICAL TALENTS OR CAPABILITIES IN ANY SCIENTIFIC FIELDS. THE PERSONS NAME, CREDENTIALS AND A SHORT SUMMARY OF PARTICULARS IS DESIRED AND WILL BE APPRECIATED. PLEASE MAIL COPIES OF THE PROGRAMS, DOCUMENTATION AND PERSONNEL INFORMATION TO - LT. COL. E.P. SCHELONKA RANGE MEASUREMENTS LABORATORY ENLD PATRICK AIR FORCE BASE, FLORIDA 32925 PLEASE SEND ACKNOWLEDGEMENT OF THIS MESSAGE INDICATING IF YOUR SITE WILL PROVIDE THE REQUESTED MATERIAL AND INFORMATION. SEND REPLY TO RML->ISI ATTENTION G. CLARKE -------