Aug 91 Letters
System 7.0 and the Developer
By Kirk Chase, Editor
System 7.0 - Boost or Burden
Gary E. Crandall
Co-Founder/VP, R & D
DataPak Software, Inc.
Might Makes Right is often an unfortunate reality in the advertising and P/R world for almost any industry, at least in the current culture in which we live.
Perhaps massively large, persistent media campaigns breed familiarity for an otherwise undeciding public; perhaps triple-page glossies seen over and over and over produce post-hypnotic commands on an unsuspecting subscriber.
Then again maybe its the old saying, ... I guess the product is good since they keep on advertising.
But for whatever reason, the winner in a competitive marketplace -- at least in this United States -- is far too often the firm who can successfully bombard the public into submission with an endless parade of glitzy presentations.
This can be frustrating to a programmer or small developer who, while creating excellent work, is barely heard through the loud cries of the Big Guns.
If you are one of these innovative but frustrated types, I have some good news for you: superior quality can and often does win in the end.
It may not seem that way from a casual glance, but take it from a Macintosh veteran developer that has seen this fact merge true on a number of historic occasions.
Witness the infamous flop of Jazz for the Macintosh in 1985 by Lotus Development: during an almost unprecedented pre-release media blitz (reported to have cost at least $10 million), Jazz made the cover story of the very best magazines and was promised to be the answer to everyones prayers.
Needless to say, the product soon became a virtual flop shortly after its release.
All of us have heard other stories about software companies armed to the teeth with fantastic venture capital attempting to buy their way into public acceptance only to find themselves bankrupt after the public discovers the products werent as pretty as the magazine glossies.
The point is, quality can and does win over well-financed lessor quality when put to the test of time.
System 7.0: Quality or Glitz?
In order to ask the question of whether System 7 is real quality that will withstand the onslaught of ages, or whether it is merely another well-financed buy-in that will wither away to insignificance, we must first define what is meant by quality.
As a programmer, I tend to think of my artistic skills, clever features and exotic coding as quality, but I dont believe that is the correct definition.
QUALITY (OF SOFTWARE) IS THE VALUE AND USEFULNESS PERCEIVED BY THE USER.
That is all it is, period. In fact it is almost arbitrary and abstract to this degree. It does not matter how many millions are poured into it, it does not matter what you and I think is wonderful or powerful, it doesnt matter what all the reviewers say in all the magazines between the Seven Seas.
It doesnt even matter what the Opinion Leader nerd at the local MUG group thinks -- quality is determined purely and only as the valuableness and the usefulness of the product perceived in actual practice by the end user.
In light of this definition, before we go off the deep end holding on the coattails of System 7, we should ask some very important questions:
1. How will the average end user perceive System 7?
Many developers, including myself, get so caught up in our own Mac IIfxs with full-page monitors and other peripherals that we completely loose sight of typical Joe User.
In fact, we even forget about that Mac Plus we have sitting at home (which is all we could afford since we couldnt buy it with company funds).
Remember, when push comes to shove in a competitive market, it doesnt matter one iota what the magazines say or even what Apple thinks, what only matters is the usefulness of your product to the average user.
If youre developing software for Hollywood agencies who wouldnt be caught dead with anything less than a IIci with two high res monitors, 8 megs of RAM and CD readers, then by all means System 7 is an absolute must for your inevitable success.
On the other hand, if your audience is the single user who deals with real estate on the side when he comes home from his 9-to-5, you should think twice about depending on System 7 since he probably wont even be able to run it!
THE RULE: DONT DEPEND ON SYSTEM 7 UNLESS YOUR AUDIENCE IS GUARANTEED TO US IT, OTHERWISE IT WILL BE PERCEIVED AS USELESS.
2. Will System 7 make my product more useful or will it make it more complicated?
This is a touchy issue because each and every programmer on this planet, myself included, thinks his program is the easiest thing to use in the universe.
But I can remember on all too many occasions where the user was completely lost with features that I thought were the essence of simplicity.
For instance, to this day I have observed users who simply cannot navigate a Standard File dialog successfully or who really dont understand the Desktop concept, and yet here we are adding yet more power features and in fact a whole new System! Does the users even know what a System is?
This is a most critical issue to address in any product: in terms of genuine simplicity, how will the user perceive your package?
Yes, put in useful features, but always step back and get the viewpoint of how it looks to the purely untechnical.
I remember an enlightening experience dealing with the original Apple IIGS in 1986. Our company released a simple, almost clunky product for simple writing an publishing and it sold like mad. Of course, we had to release an upgrade with some nice new features, but we did so and the product only sold 1/2 the quantity of the original.
But since we knew best, we released yet another upgrade only this time with enough features to rival Page Maker.
Needless to say, the new version practically did not sell at all.
Then one day I got a funny little letter from a user saying, I liked the original the best.
We finally got the SRA (Severe Reality Adjustment) that less features can sell more copies. As strange as it may seem, there is a painful truism to this idea!
RULE: TAKE ADVANTAGE OF SYSTEM 7 TO MAKE YOUR PROGRAM SMALLER AND SIMPLER.
3. Will System 7 cause better software to be produced?
To answer this question intelligently, one must first decide what is meant by better.
If better means we are now all going to produce better code, the answer is a resounding NO.
No operating system in the world will cause us to write better programs. Easier, perhaps, but better, I dont see how.
For years I have listened to excuses for poor code blamed on the slow CPU or the lack of RAM, yet in the pioneer days long before Apple I saw, with my very eyes, a multi-terminal database system holding 350,000 random-accessible records using nothing more than 64K of RAM with archaic hard drives and 8080 microprocessors!
Bad code is bad code, regardless of the System. Sometimes the superior hardware will mask the poor quality by its shear processing speed, but move that software to a slower machine and youve had it.
RULE: LET SYSTEM 7 DO THINGS FOR YOU BUT DONT USE IT TO MASK POOR QUALITY.
In order to produce quality software, the first rule is to understand your audience, acquire the ability to see through a mass media buy-in and to duplicate the end users viewpoint.
The second rule, after you have successfully applied the first rule above, is to gear all your new features -- System 7 included -- in that direction.
The third an last rule is recognize that nothing can replace responsible, quality code and that System 7 is a helpful means to that end, not the end itself.
Some New Products
Some interesting press releases for developers and engineers have been crossing my desk. Here are a few of them.
Some FORTRAN Help
2940 Union Avenue, Suite C
San Jose, CA 95124
Cobalt Blue has recently announced two MPW tools aimed at cleaning up old, messy Fortran code. PRO_STRUCT is used for small chunks of code that are typically under 500 lines per function. FOR_STRUCT is the second product aimed at larger code blocks. PRO_STRUCT will structure to MISPEC and VAX FORTRAN-77. Now you can "re-write" those FORTRAN-IV code blocks into maintainable and easier to read FORTRAN-77 code. It can also structure to FORTRAN-90 statements, and it has lint capabilities. Contact Cobalt Blue for more information.
Famous Engineer Brand Software
4855 Finlay St.
Richmond, VA 23231
digiMatic is a desk accessory that turns graphics into numbers. With digiMatic, you can scan charts, plots, etc. and then turn those digitized images into tab-delimited text or PICT2 formats for use in other applications. digiMatic can also determine screen pixel coordinates of window resources and the relative coordinates of regions. Price is $229.
XVT-Design and Developer Conference
XVT Software, Inc.
1800 30th Street, Box 18750
Boulder, CO 80308
XVT has announced its developer's conference for August 26-30. By now, the conference is on a space-available basis. This might be interesting to attend since the IEEE is using XVT as the base document to draft the standard for a Layered Application Program Interface for graphical user interfaces. XVT-Design is a design tool and application generator for its toolkit. This product is due out in the end of July and will help developers design GUI resources.