Martin Banks, Personal Computer World 09/81 - checked

September 1981

Software, if you'll pardon the pun, is a soft touch for anyone with a yen for making barbed comments about the computer industry. It is, in fact the moron that skulks inside the snazzy boxes of hardware that look so terribly clever.

See, its easy to have a go at software, especially in the microcomputer business. It rarely is what most people seem to think it might be, or hoped it should have been given luck and God's good grace. Software producers certainly over-step the mark at times, broadly alluding to the wondrous power of the XYZ suite of integrated packages in their advertising. Systems manufacturers often more than broadly allude to the infinite capabilities of their ticky-tacky boxes. . . 'You could run all of GEC's accounts on this wristwatch-styled gizmo, costing just 27 pee'.

Yes, the poor bemused punter is often oversold on the joys and wonders of the microcomputer. Sometimes, they are even oversold - dare I say it - deliberately. Well, a sucker is a sucker is a sucker, I suppose.

Many times, however, it would seem that the punters oversell themselves on the whole idea. Often, it is only with great difficulty that they are dragged, kicking and screaming, back to the world of reality, and other boring things like file sizes, field lengths and suchlike. This self-abusive overselling is probably more prevalent in software than hardware. Since the heady days of the late seventies, when that nasty thing called the microprocessor came to national prominence, there has been a slow but steady awakening in people's perceptions about hardware. They now yawn - through knowledge rather than disinterest - at the old chestnut about 'this box' holding the equivalent power to a computer that used to fill a house.

But the same cannot be said of software. To many people, and certainly to many potential users of microcomputers, the subject of software tends to elicit one of three different and distinct responses. The first is the blank stare, occasionally accompanied by a querulous `eh?' The second is the dubious enquiry as to whether it is illegal or should be mentioned in polite company. The third response follows the pattern: 'Ah yes, I know about that. I've seen adverts for complete accounts packages for £40.'

Now I know that there are many, many, happy and contented users of microcomputers running much good software and achieving exactly what they set out to do. I also know that there are many individuals writing that software, both for themselves and other people. But at the same time, I also know that there are many, many, discontented users who feel they have bought a pig in a poke, and many, many, more who haven't bought anything yet but run the risk of one hell of a surprise when they do.

Now, as I have also said (back at the top of the page, remember?) it is easy for someone to sit complacently at a typewriter and have a good whinge and whine about software. Therefore, from this point on I shall try to avoid that. Instead, I will make a suggestion.

It is a suggestion that has been borne out of a story that concerned a user, a particular computer system, and a software supplier. In this context the names don't really matter.

Now the user purchased the computer, because he liked it and then shopped around for some software that would run the applications he had in mind (mainly accounts and job-costing work). This, he now freely admits, was a mistake. He acknowledges that he should have found the applications packages he needed, and then found a machine that would run them. That, unfortunately, is the way that many first time buyers are liable to start.

Now, he found some software that could do the job, at least on paper, but in practice it didn't seem to work out that way. The implementation of the software involved work, correction and bug-hunting, accompanied by much burning of the midnight oil by all concerned. 'It doesn't work,' cried the user, 'and it can't be got to work.'

From that user's point of view that statement was true, and very self-evident, but there was a but (isn't there always?).

The `but' in question came from the supplier of the software package, who said, in effect, that he was hardly surprised it didn't work. This was not to say there was anything wrong with the package. In theory it would meet the application nicely. On a bigger machine it would have coped well. 'You see,' said the software man, 'that application, the volume of work to be processed, really needed a minicomputer, though some of the biggest micros could have handled it.'

Therein lies the rub. The user had a good idea - to use a micro computer to run the business accounts - but its implementation was a different matter. The simple (at least in theory) task of gaining a rough quantitative guide to the job, that could be gauged against the known capabilities of a range of machines and software packages, seems to have been missing.

In old style computing (mainframes etc), this exercise is sometimes called systems analysis, and is a job for the creme de la creme of the brotherhood. Needless to say, that makes it very expensive and takes it beyond the financial pale for the majority of potential micro buyers.

But there is scope for cheaper variations on that theme and, while not so accurate (and even in the mainframe business that is a very relative word), user requirements could be rapidly quantified at the dealer level and matched to software and hardware.

It occurs to me that there ought to be some way of grading both hardware and software, and also the user's application. Software, in particular would benefit from this, for it could help both the potential user and dealer. Both would have a common frame of reference around which to discuss an application and its implications.

Hardware would be relatively easy to grade, for it could simply be based on a system's capabilities. For example, on a scale of one to ten, in ascending order of 'power', you could have a Sinclair ZX81 without add-on memory as grade one; grade five might be an Apple/Tandy/Commodore machine running two mini-floppies and a printer then, at the top (purely for the sake of argument), machines like the Onyx 8000 running over 20 MBytes of hard disk.

Software would be more difficult to grade, but not impossible. Simple games and educational programs would be grade one, while a fully integrated accounting package with bells and whistles would be grade ten. It would of course, have to be usable on grade ten hardware. Intermediate gradings of software would be difficult: there would no doubt be a need for further subdivisions into different applications areas, such as engineering, accounting medical, etc. The gradings should, however, be matched to what the packages can actually do. For example, there could be a very fine general ledger package that does not integrate with any other accounting package. This would have a lower grade than a similar, text integrating package (unless, of course, that integrating package was a bummer).

By the same token, it should be possible for a dealer to have some form of checklist for the user. This could be a series of questions that would allow the dealer to quantify the application, and thus give it a grading that is relative to those for hardware and software, ie, a grade ten application will need a grade ten computer running grade ten software. As a side benefit, such an exercise would help the user gain a better understanding of just what he is trying to achieve, and why. He might end up realising that he doesn't need a computer at all.

I really don't know if such a system of grading is workable, or whether it is economically viable to establish. Ideally, all this would be done by some independent body with no vested interests to protect. The chances of that are remote, to say the least.

But there is scope for enterprising companies, or groups of companies, maybe even the Computer Retailers Association, to have a go. Even if the grading structure was not of the best, it could well be better than nothing at all. I am reminded of the CP/M operating system, which is certainly not the best there is in operating systems, but has succeeded primarily by being there.

A poor grading system could succeed by the same token. A poor grading system that at least gives some quantitative comparison for dealers and users to work on does, however, seem infinitely better than the current state of the oversold wing and a prayer.

end