Kay Nishi, the brains behind MSX, was in London changing planes ten days ago. He'd just come from seeing Philips in Eindhoven, and was batting off in the general direction of Anchorage/Seattle, presumably to see Boeing about MSX standard 747s, but his minders from Microsoft had allowed him an overnight stop, so he drew breath to give an interview.

It's difficult to overestimate Kay's importance to Japan Inc's microcomputer arm. If MSX was anybody's idea it was his, and if anyone can be said to be in control of how MSX is going to develop, then it's Kay Nishi.

The chip for the second generation MSX machines, for example, will have to be decided on fairly son, and the flower of Japan's electronic industry, plus upwards of ten fellow-travellers, will be going with it. Tens of millions of dollars will be riding on the choice and, says Kay. "I'll probably decide in January." You couldn't be sure he was joking, either...

The current generation of MSX machines doesn't look like sweeping the country this Christmas and it's doubtful whether they'll get anywhere near the manufacturer's projected market share. But Kay professes to be happy with the response they're getting from customers, and reckons on a three to four year period before the standard's established. By 'established' he means the machines being accepted in the machines being accepted in the home in the same way that phones and TVs are accepted. But it's his thoughts about the next generation, and developments f the current generation, that are most interesting.

The machines currently in the shops are pricey, but although we're not liable to see reductions this Christmas they could work out a lot cheaper by Christmas 1985. And one of the main engines that will drive down prices will be custom VLSI (Very Large Scale Integration) chips.

Today's MSX machines have more than 15 LSI (Large Scale Integration) chips inside them, and although the development costs are hoorendous, putting all these on one VLSI will in the long run bring costs way down. From the MSX companies' point of view it's a shrewd move, because it allows them to use their financial muscle to leapfrog smaller companies - Sinclair Research, for example.

Development of the first of the VLSIs is now complete, and we should be seeing them in MSX machines next autumn. Prices won't be down to the $100 dollar mark that's becoming the vogue in Japan, but it shouldn't be too difficult to cut them to around £150.

Custom chips will also play an important part in MSX 2. Kay won't talk about the next generation yet, but it's easy to see the way he's thinking. He defends MSX against the 'old technology' criticism by pointing at music - "Hey! Mozart wrote symphonies over a hundred years ago, and I can still understand them". He also offers the more relevant examples of Intel.

Intel's success, he says, has been based on the fact that its chips are compatible, whether 8088, 8086, 80186 or 80286. So the next MSX machines will run the Zilog Z8000? Definitely not, he says - he'll be looking at Intel or Motorola, and by the dreamy look that comes into his eye when he says 68000 you can guess what it's going to be.

But that's not the whole story. MSX 2 may be going 16-bit, or it may jump straight to 32, and to maintain compatibility with the current generation of machines it will need some sort of Z80 chip built into it. Kay suggests that a custom chip involving a merger between the Z80 and 68000 is a solution.

In the medium term MSX will be following the example of phone and TV networks by getting itself wired. We'll be seeing a Prestel link some time next year, but the big difference between this and current systems will be the fact that it introduces graphics mail. The idea of people being able to send electronic Christmas cards plainly excites him.

Part and parcel of this will be videotex and light-pen circuitry integrated in MSX graphics, so an enhanced video chip is on the cards. And next year we should also be seeing the product of his visit to Eindhoven.

Philips and Sony are currently working on a compack disk interface for MSX, scheduled for September/October, and giving 550Mb of storage.

Straight disks are also on the cards - but then they would be with Sony and Hitachi involved. These two companies will probably battle it out over the rival 3" and 3.5" format, but it would appear MSX will be available on practically all disk formats.

One interesting idea involves integrating home MSX machines into businesses by allowing you to swap data disks between your office and your home machines. You could work on Wordstar in your office, for example, then pluck your disk out and carry on using your MSXDOS version of Wordstar at home.

The fascinating thing about this is that the MSX machines, being 8-bit, will have to be upgradeable in the direction of CP/M, which is of course a product of Digital Research, Microsoft's arch-rival. If Microsoft doesn't want to wind up paying DR licence fees it's going to have to think hard about making MSX 2 MSDOS compatible. When you add to this a range of home control systems (arriving in about six months), Kay's belief that MSX will be selling to a market untouched by today's micro manufacturers starts to look plausible.

Eventually, he says, people will be buying MSX micros and they won't even know they're micros. Back to that 747 interface?