From: IN%"cmkrnl.com!jeh@rutgers.UUCP" "Jamie Hanrahan, Kernel Mode Systems" 10-MAR-1993 04:06:32.59 To: IN%"TERRY@spcvxa.spc.edu" CC: Subj: RE: Need some Quote's Return-path: Received: from VN-GATEWAY by spcvxa.spc.edu (PMDF V4.2-7 #042B2) id <01GVMOD94OYO8WW0IS@spcvxa.spc.edu>; Wed, 10 Mar 1993 04:04:06 EST Received: from mimsy.UUCP by rutgers.edu (5.59/SMI4.0/RU1.5/3.08) with UUCP id AA24802; Wed, 10 Mar 93 03:00:41 EST Received: by mimsy.UMD.EDU (smail2.5) id AA15114; 10 Mar 93 03:00:21 EST (Wed) Received: from uunet.uu.net (via LOCALHOST.UU.NET) by relay2.UU.NET with SMTP (5.61/UUNET-internet-primary) id AA24518; Wed, 10 Mar 93 02:45:14 -0500 Received: from cmkrnl.UUCP by uunet.uu.net with UUCP/RMAIL (queueing-rmail) id 024454.6639; Wed, 10 Mar 1993 02:44:54 EST Received: by cmkrnl.com (DECUS UUCP /2.0/2.0/2.0/); Tue, 9 Mar 93 23:38:31 PST Date: Tue, 09 Mar 1993 23:38:31 -0800 (PST) From: "Jamie Hanrahan, Kernel Mode Systems" Subject: RE: Need some Quote's To: TERRY@spcvxa.spc.edu Message-id: <0096945C6888DB60.000012DB@cmkrnl.com> Content-transfer-encoding: 7BIT X-Vms-Mail-To: UUCP%"TERRY@spcvxa.spc.edu" X-Vms-Mail-Cc: UU%"_hanrahan@eisner.decus.org",JEH > Do you have the one called "A Fable Concerning Product Design and > Implementation"? Starts with: Attached. I'll post it on DECUServe also. Might prompt a whole series of similar things. > Nope - I'd appreciate it. Also, note my post said "can't do general dis- > tribution". That means if one of my friends asks (like you), I can make a > tape with the whole mess. Just ask. I can do 1600/3200/6250 tapes (free), > TK50's (you supply the tapes), DC6250's in Unix "tar" format (likewise), > amd 4mm (90m, compressed). How about TK50's - how many should I send? --- Jamie Hanrahan, Kernel Mode Systems, San Diego CA Internet: jeh@cmkrnl.com Uucp: uunet!cmkrnl!jeh CIS: 74140,2055 ------------------------------------------------------------------------------ A FABLE CONCERNING PRODUCT DESIGN AND IMPLEMENTATION Consider a custom automobile production shop which is run as its analog in programming might be. Let us take an order and follow it through to completion. This particular order will be for a special high-performance racing car, a low-profile, 180 mile-per-hour job for oval track racing. The customer has given the description to the shop. A fixed-price bid has been submitted and accepted. Three automotive engineers are available. Two of them decide that the engine will be the trickiest portion, so they begin debating ways to operate the valves. In talking about the system, they draw diagrams on a chalkboard. These diagrams are of two kinds: Fuel flow diagrams and cam timing diagrams. These are the only design tools available, so they use them all the time. The third engineer is a specialist in springs. He goes out to the shop and, with a welding torch, starts piecing together a suspension system. He develops a combined leaf and coil system with trailing torque arms and transverse torsion bars that is the best ever produced. He then moves on to the problem of propulsion and begins toying with schemes for a spring-driven, wind-up engine. He has already started building this when the others, coming out to the shop to start building the intake manifold and valve assemblies, tell him that it is impractical. He sits down and just watches. Shortly the two senior engineers decide that they need a distributor. They carefully specify its size and shape and the material it is to be made of, as well as its maximum rotational speed and voltage and current tolerances. The idle engineer then isolates himself in his office to develop it. There he notes how elegant it is to divide a circle into six arcs using a compass and settles on a six-cylinder pattern. The engine is built from front to rear. When the eighth in-line cylinder is reached, the two engineers start designing the attached floor and body. This is not easy with fuel flow diagrams, so they simply start building without a design. Halfway through they find that their floor plan is too narrow for the elaborate rear suspension. Since it is already built, they are hesitant to throw it away; they angle the body outward so that it is wide enough at the rear. Noting that they will need a transaxle, they build that assembly integral to the body pan. The distributor has been completed and is fitted into place. A debate then ensues about whether the engine was supposed to have six or eight cylinders. A compromise is reached: The engine will have seven active cylinders; the eighth cylinder will be a dummy, and cylinders one and seven will fire together. The electrical system is modified to accommodate the extra surge. Later, another engineer suggests the ingenious idea of using the eighth cylinder as a pump to power hydraulic wipers and windows. As an added feature, the doors are made to operate hydraulically too. On delivery to the customer, attention is drawn to the unusual wedge shape and to the convenience of hydraulically operated doors. He is told how clever the hookup of the eighth cylinder is and that it will only be a minor inconvenience to have to leave either the engine running or the doors open. Later he finds that strict adherence has been made to his original specifications; the car only turns left (for counter-clockwise ovals). When an attempt is made to change this, it is learned that the steering gear and motor housings are a single piece. Since there is no access door, the engine has to be cut open, at considerable cost. But aside from the fact that the two halves of the steering wheel are of slightly different diameters and that the wipers will not operate with the windows up (due to a glitch in the hydraulic network), the customer is reasonably satisfied. The shop, meanwhile, has lost its collective shirt, largely because of the time and effort spent experimenting with various wheel shapes before finally settling on the optimal -- round. (end)