shaffer@operations.dccs.upenn.edu (Earl Shaffer) (06/09/87)
In article <4146@prls.UUCP> weaver@prls.UUCP (Michael Gordon Weaver) writes: >If you are writing a program as a commercial product, the time of availability >of the program may be critical to the financial success of the program, usually >because others will be trying to produce something to fill the same need. > >If in such a situation your program sells, you will be able to afford >to hire the additional manpower to do the extra work to make a good >program out of a quick hack. Conversly, if it does not sell, you won't >have to worry about maintenance. This is just the line of reasoning that has made software a "kludge" rather than a regular business product. AT&T does not "hack" a phone together so that they can beat the market. AT&T would be broke disgraced & out of business. Since learning UNIX I have seen these "hack" programs everywhere. People say "oh, well lpr isnt supported but its great..". No its not! Its a terrible print queue! It fails more often then any other thing I have ever seen. Many other utilities, rn, talk, mail, etc. can be made to fail with the slightest "push". Its just plain poor programming. I dont know how your company does business, but if programming in general is ever going to get a good rep, it will have to come from well tested and well designed products. Not things like DOS 3.3 and other "quick" marketing solutions. (said with the proper number of happy faces to foster harmony) ============================================================================== Earl Shaffer - University of Pennsylvania - Data Communications Department "Time was invented so that everything wouldn't happen at once." Steven Wright ==============================================================================
ekwok@mipos3.UUCP (Gibbons V. Ogden) (06/10/87)
In article <1320@super.upenn.edu.upenn.edu> shaffer@operations.dccs.upenn.edu.UUCP (Earl Shaffer) writes: >In article <4146@prls.UUCP> weaver@prls.UUCP (Michael Gordon Weaver) writes: >>If you are writing a program as a commercial product, the time of availability >>of the program may be critical to the financial success of the program, >usually >>because others will be trying to produce something to fill the same need. >> >>If in such a situation your program sells, you will be able to afford >>to hire the additional manpower to do the extra work to make a good >>program out of a quick hack. Conversly, if it does not sell, you won't >>have to worry about maintenance. > >This is just the line of reasoning that has made software a "kludge" >rather than a regular business product. AT&T does not "hack" a phone >together so that they can beat the market. AT&T would be broke >disgraced & out of business. > A few years ago, there was a race to get to the market first in the CAE area; looking back, most of the engineering workstation software of that era might have been products under pressure to get out. The performance on most stations were dismal (they blame it on the hardware) and the "integrated solution" they offered were ususally mish-meshes of whatnots of the past eras they happen to find from different sources. (You can disagree, of course; that's my personal view). Amazing enough, I didn't find among the Daisies, Mentors and Valids a clear winner and any dominance in the market because of early presence. To the users I have met, reliability and "kludginess" characterize the day. I wonder if software methodology has anything to do with this. -- Where is Rose Bird, Jerry Brown and Don Maclean? Are they on the net?