What to do about the phone books

BUSINESS VIEWPOINT REX MALIK

For the last ten years, the effects of the com- puter have been largely felt by the public at second hand. But we are now within measurable distance of the time when the relationship be- tween western man and his computer will be much like this current relationship with his car: symbiotic. The key is to be found in the tech- niques which are enabling people and com- puters to hold something much resembling a conversation.

How? Television-type sets with special types of linked pointers called light pens, which can provide an instant draftsman capability to the most unskilled. Teletype machines with type- writer keyboards on which one types queries and on which the computer replies, and com- municators of various types into which one speaks or otherwise puts information; all these will be connected to computers via com- munications channels of many kinds. And one of the most obvious communications channels of all is the telephone network.

I am not here discussing the possibilities of science fiction or looking at a possible future as envisaged by Herman Kahn. These systems are already in existence in considerable numbers. There are at least a dozen in Britain already, and I have been using one from home on and off for a year now. Anyone who flies BEA and books his ticket say at West London Air Terminal will probably see a BEA employee use another.

And what, you may well ask, has all this to do with the row over the Post Office London Telephone Directories? As yet little, but it could have a lot. But first for those who are not quite sure where that row stands, the new Postmaster-General, John Stonehouse, MI', this week called a meeting of everyone from the CBI to the Townswomen's Guilds and a repre- sentative sample of users to discuss—the phrase is the oPo's—'how the user can best be served.'

Before that, the state of play was that the GPO was still going ahead with its plans to create thirty-six local directories to replace the present four London directories. The only concession they had made to meet the public outcry was that anyone who wanted all of them could have them. Now the PMG says that he will 'approach the problem with an open mind and will arrive at his decision after careful consideration of the opinion expressed, and any suggestions made at the meeting.' How- ever, the present four-directory system is also no solution to the problem that the GPO and the user both face: how to cope with all this information? As it happen:. the current directories come a long way in standard below what we should be aiming at. The cataloguing of names, particularly when lot of them are the same, is almost haphaiard the °Po indeed does not make as much use of the information it obtains from its users as it could. In many cases users can only be sufficiently distinguished from each other by address which is not much use when you come to the more than 7,500 Smiths in the existing directory (more than 500 of them J. Smiths). And if sou should happen to obtain a new phone just after your directory has gone to print. you will have to wait a ‘'ery long time before }on see your name in print in the book. As for phoning directory inquiries, most of the time it doesn't have street directories to call on. Yet with all this, the GPO faces the prospect that the tele- phone subscriber population may double in size by 1975, and the directories also double in size if nothing else is done.

At the height of the original controversy, a colleague and I produced a paper—which we sent to the GPO and anyone else who would listen—stating quite simply that computer technology is now advanced enough to cope with the directory, produce a service in- comparably better for the user than the present one, and what is more provide a long-term solution; one which would not need to be changed again as the service grew. My partner in this venture was an American, Alan Marshall, ex-whiz-kid (the American Navy's- senior computer man at the age of thirty-three, effective rank of vice-admiral, making him at that time the most senior man for his age in the American government).

Our plan was to do away with the directories

almost entirely. All subscribers would receive a small printed directory—which left plenty of space for writing in their own numbers, but would otherwise contain basic local and capital information. That would be it. For all other numbers, they would use the telephone. The directories would be organised across a set of eight numbers—preferably on a new exchange, though that is not essential—two each to cor- respond with the four-directory structure.

Once the information is in a computer system, there are two ways of getting at it. One, it is possible to devise a system so that the user dials in the information he has. Two, he speaks it. The technology in both areas is now advanced enough for this to be a realistic proposition. Our system would use a mixture of the two methods. On easily recognisable names the system would be able to interrupt the inquirer as soon as it had enough informa- tion, sometimes the name alone might be sufficient.

Where it is more difficult, the system would be able to hold a checking conversation with the user until it had enough information to be able to satisfy him. And should this not be possible for any reason, our system would allow him to be switched over to a human director inquiries operator who can then deal with the more esoteric queries.

How would it run?

Take an example. Our inquirer is searching for a J. Smith who lives in a particular street in the King's Cross area. That is all he has.

He dials the S to Z directory; incomplete ) information. Directory answers.

'Director Inquiries, S to Z, incomplete information. Please state surname and then initials of user whose number you require. If you do not know surname, dial one once.' Inquirer: 'The name is Smith, J.'

Answer: 'I understand the name to be Smith, J. Acknowledge if this is correct by dialling one once. If incorrect, dial one twice.'

Inquirer: Dials one once.

Answer: 'Address information please. Road or street number, postal district number or area reference.'

Inquirer: `Beezelbub Road, number unknown. Near King's Cross.'

Answer: 'Smith, J.'s number is 333 9999.'

The above inquiry would with our system have been dealt with—allowing for the maxi- mum of fumbling by the inquirer—in less than ninety seconds. The system is provided with constant check-back facilities, and observant readers will have noticed that we have thought up a system which coped on the basis of information inadequate for a successful out- come with the present systems. In the case of the printed directory it could take quite a long time; in the case of the present directory in- quiry service, it is almost impossible.

What would it cost? Well with the best (or worst) will in the world, we fail to see how the GPO could run the cost to more than half a million pounds a year—which is well below present-day costs. With urgency, it could be created and brought into operation within three years, which is about half the time that the oPo has allowed before it means to change the system anyway.

Above all, it gives us a chance to put some already existing know-how to use, and that in advance of everybody, and we haven't seen that for some time. Do we have to go on play- ing follow my leader for evermore? User convenience apart—and I hope I have demonstrated that—what's the GPO for?