david@ms.uky.csnet (David Herron, Resident E-mail Hack) (04/27/87)
The question for the day, gentle people, is "How to do software distributions"? The situation. We have a small-but-growing pile of locally-written software which we feel would be useful to the world at large. We could release these things one-by-one to mod.sources and let the network take care of it, but we're interested in doing it in a different way. We're thinking that, at some point in our future, it would be useful to be able to say that x places in the world are using software of our manufacture. Producing software and distributing it (even through mod.sources) does bring some sort of visibility to our little department and our university as a whole. We should be able to reap some benefits from causing this visibility. With that as justification, let me ask again. How do *you* handle the grubby details of doing software distributions? Like, advertising the existance of your software, making the tapes, handling verification, organization of the sources, or any other useful ideas. I would also be interested in hearing from professional software houses, not just other Universities. On a related point. I think it would be useful if there were a central place which collected descriptions of such things and kept track of ordering information. (& etc). But I know of no such central thing. What I have in mind is something like the List of Lists. I'll post a summary of responses if there is interest. I might even be interested in running a Distribution of Distributions service. hmmm... :-) advTHANKSance -- ----- David Herron, cbosgd!ukma!david, david@UKMA.BITNET, david@ms.uky.csnet ----- (also "postmaster", "news", and the Usenet map maintainer for Kentucky.) ----- "Doodle, doodle, dee; Wubba, wubba, wubba" /*
mark@mimsy.UUCP (Mark Weiser) (04/27/87)
In article <6396@ukma.ms.uky.csnet> david@ms.uky.csnet (David Herron, Resident E-mail Hack) writes: >The question for the day, gentle people, is "How to do software distributions"? At Maryland we do the following: 1. Collect 99% of our useful software into a package called the Maryland Software Distribution (in addition to occasional net.sources postings). 2. Advertise extremely infrequently on appropriate net groups, and in no other way. (Advertising consists of, when a question arises in a newsgroup for which our software may be the answer, posting a message which says: "mail to this_person@maryland for more info on our tape, if interested".) 3. Getting a tape requires (a) $250 (in cash, no invoices), (b) that you sign our "no commercial use or redistribution, this is good-for-nothing research software" license. (On occasion we have had derivative software on our tape, and have required copies of bsd licenses. But we try to avoid this.) 4. Tape copying and administration is handled part-time, during her "own" time, by a secretary who is paid a straight 10% of the tape fee per tape handled. This 10% is intended to handle any future administrative issues that may also arrive from a given tape. 5. Tape dollars go for: (a) buying blank tapes and other administrative overhead, (b) the secretary above, and, mostly, (c) sending worthy graduate students who have contributed software to the tape to conferences. With regard to (c), if your software is on the tape, and you are still a Maryland student, you have an undisputed right to have most expenses paid to the conference of your choice once a year. This covers only about 6 students at the moment. 6. Tape dollars are placed in an account, not of the University, but of the University of Maryland Foundation, to avoid University red tape. First aside: almost all our tape sales have been for one of our two big packages, Maryland Lisp (a variation on old Franz), and the Maryland Window System. We throw in lots of other stuff, but I only know of two or three places who bought the tape for the other things. We have sold about 300 tapes over the years. Second aside: we have scrupulously avoided giving any of the money directly to anyone (other than the secretary), to avoid any problems with anything (the nature of which I am not sure of, but I don't want to know). Third aside: we are now phasing out the Maryland Software Tape, because the professor in charge (me) is leaving Maryland. Most of the remaining dollars are going to establish a "Maryland Outstanding Programmer" award (also known as "Hacker of the Year"), to be paid from the yearly interest. -mark -- Spoken: Mark Weiser ARPA: mark@mimsy.umd.edu Phone: +1-301-454-7817 After May 15, 1987: weiser@parcvax.xerox.com
henry@utzoo.UUCP (Henry Spencer) (05/03/87)
> ... "How to do software distributions"? > > ... Producing software and distributing it (even through > mod.sources) does bring some sort of visibility to our little department > and our university as a whole. We should be able to reap some benefits > from causing this visibility. Mark Weiser's reply to this one strikes me as a good starting point. Some other thoughts: 1. Do not, repeat not, assume that making and sending distribution tapes is a trivial chore that somebody can do in their spare moments. One practice I particularly disrecommend is adding it to the responsibilities of your most junior programmer. (Speaking as someone who was once such a M.J.P., it's a time-consuming clerical chore that the M.J.P. will resent and perform poorly, especially if he's already got a full workload.) Mark's scheme of making it a *paid* piece-work (paid by the tape, not by the hour) job is rather better. The investment of time and effort is not trivial. 2. In the same vein, plan to charge a fee to recover that time and effort. Doing it free gets tiresome very quickly. 3. Even if you license the stuff on a per-site basis, look *very* hard at doing all distribution electronically. The AT&T Software Toolchest people consider this a major win, and refuse to supply magtapes even though they are a bit cheaper in some situations. 4. If you must use magtapes, invest money and effort in getting good tapes and keeping your tape drive tuned up. Unreadable distributions are painful. Supplying the tapes yourself is an order of magnitude less hassle than trying to use tapes that customers send you. -- "If you want PL/I, you know Henry Spencer @ U of Toronto Zoology where to find it." -- DMR {allegra,ihnp4,decvax,pyramid}!utzoo!henry
emrozek@ecsvax.UUCP (Ed Mrozek) (05/05/87)
I do not remember seeing the original article. If I have missed something, please forgive and forget. Software distribution is a costly venture. As the summary says, do it right; or don't do it at all. ----------------------------------------- In article <7993@utzoo.UUCP>, henry@utzoo.UUCP (Henry Spencer) writes: > > ... "How to do software distributions"? > > Mark Weiser's reply to this one strikes me as a good starting point. Some > other thoughts: > > 1. Do not, repeat not, assume that making and sending distribution tapes is > a trivial chore that somebody can do in their spare moments. One practice > I particularly disrecommend is adding it to the responsibilities of your > most junior programmer. (Speaking as someone who was once such a M.J.P., > it's a time-consuming clerical chore that the M.J.P. will resent and perform ^^^^^^^^ It *is* a clerical chore! Nothing more and nothing less. Q/A is another issue. It is another responsibility for another type of person. > poorly, especially if he's already got a full workload.) Mark's scheme of > making it a *paid* piece-work (paid by the tape, not by the hour) job is > rather better. The investment of time and effort is not trivial. If distribution gets to be a large operation, there are devices that can be purchased to automate duplication. Here, we supply software and data on floppies to 160 field offices. When we got started, copies were made by a clerk and anyone she could find to help. Soon after that first release, we bought a disk duplication system. What took approximatley 8 man weeks was reduced to a one- or two-day affair. I suspect that the device will pay for itself in about two to three years. It may be sooner because we are now renting it out as a sevice. This appears to be paying the maintenance contract. By the time I am ready, I suspect I will have found (and purchased) a device for duplicating tapes. Also, along these same lines, there are commercial operations that will duplicate, package, etc. for a fee. In some situations, these services are reasonable alternatives. > > 2. In the same vein, plan to charge a fee to recover that time and effort. > Doing it free gets tiresome very quickly. I will be a little stronger on this. I don't believe that any academic department can afford to distribute free. Also, I think that most people are willing to pay reasonable duplication and distribution fees. If available, I would also pay for some support for installation and configuration. > > 3. Even if you license the stuff on a per-site basis, look *very* hard at > doing all distribution electronically. The AT&T Software Toolchest people > consider this a major win, and refuse to supply magtapes even though they > are a bit cheaper in some situations. Yes. However, you may limit yourself. (please ignore if I've missed something in previous discussion about this). > > 4. If you must use magtapes, invest money and effort in getting good tapes > and keeping your tape drive tuned up. Unreadable distributions are painful. > Supplying the tapes yourself is an order of magnitude less hassle than > trying to use tapes that customers send you. > -- I would change this to ten orders of magnitude less hassle. In fact, if you are serious about distributing software make sure that you supply the media and use a cost formula that will allow recovery. P.S. If someone does figure out how to do it right, please let me know.
jbn@glacier.STANFORD.EDU (John B. Nagle) (05/13/87)
There's a well-developed industry, at least here in Silicon Valley, for copying 5.25" floppies. A disk, formatted, copied onto, labeled, and shrink-wrapped should cost you under $1 in quantities of a hundred or more. See the yellow pages for vendors in your area; don't query me, please. John Nagle