Personal Computer News
5th January 1985Categories: News
Author: Bryan Skinner
Published in Personal Computer News #093
"Hardware Glamour" Is Simply Not Enough - Homefront
Why can't they get it right? Micro manufacturers and micro specifications, that is. This outburst follows the so-called release of the Enterprise, sixteen months old at birth.
It should be so simple. If Amstrad can cobble together a more than half-decent micro from off-the-shelf parts in under a year, companies with more experience should be able to turn out everyone's favourite micro with one hand tried behind their back.
But the Commodore 64 has lousy Basic and the Spectrum has a strangely mapped screen, as does the Amstrad. Only the Enterprise comes with a joystick built in - without a fire button - and only the Amstrad has a dedicated tape deck and monitor as part of the package. The C64's drives are slow, the disk drive from Oric and Amstrad won't support random access files and the BBC has a severely limited memory. You name it, there's at least one department where my micro beats yours.
What's the answer? Well, it's clear that you don't really need a 16-bit processor, so a Z80 running at 4MHz or so will do. Certainly the machine must have sprites, and more than the C64's miserly eight. Say 16, and here MSX gets it almost right with 32, except you can have only four on one line at once.
Micro graphics, like sound, need a dedicated processor, and at least 16K video RAM, preferably 32 so we can finally sort out all those attribute problems and have a decent colour range in high-resolution. And the screen handling must offer Boolean operators to allow for fore/background activities. The colour range is a tricky issue. The Dragon lost out largely because of its pathetic resolution and lack of colours, while Atari and Enterprise demonstrate that it's not impossible to offer a wide range of colour tones, without slowing down the cpu or demanding masses of screen RAM.
On a similar note, interrupts should be available from Basic and there should be some attempt at multi-tasking - even if it's restricted to three or so tasks.
One thing I'd really like to see is a built-in toolkit. The BBC may have an assembler, but what's needed is a full Basic toolkit, including such features as find and replace, variable dump, in-line assembling of machine code and so on. And on similar lines, modern machines shouldn't have only a cartridge port but, as with the Beeb, ROM bays too for applications software.
Any micro these days should include some built-in mass-storage device, and Amstrad almost got that right. But the tape deck isn't under micro control (which the microcassette of the Epson PX-8 lap portable is) so you must go through all that press-play-and-any-other-key to get anywhere.
This is surprising given the company's experience in hi-fi. The QL has its Microdrives, but reliability is still a major problem and capacity is low. So what's wrong with a single 3" drive or whatever, like the Macintosh?
Memory is still a thorny issue. Is 48K enough? It doesn't look like it, but 64K seems to be a sticking point. Is it really worth producing a 128K micro if the code for applications software must be so large you're left with only 15K work-space? No, 64K seems quite adequate.
But sound is a different matter. It's here that manufacturers have come closest to the ideal. Three or four voices, tone and volume envelopes, sound synchrony and so on all fit the bill.
Sound commands could, however, be simpler. Surely micros offering hardware defined (but user redefinable) characters could offer an analogous feature for sound at the same time?
Lastly, any modern micro should have a built-in modem - preferably with BT approval.
Expense should be no problem. I don't believe for one minute that what I've proposed need cost you more than £400.
It all boils down to 'hardware glamour'. Once, a micro's specifications were what sold it. I think and hope this is changing. Buyers are more discriminating and select a computer on the strength of what they can do with it, not what the ad blurb promises. Essentially, this means more bundled software and even more thought in design if any new micro is to make a name for itself. Surely someone out there can get it right?