Martin Banks, Personal Computer World 09/84 - checked

September 1984

Once upon a long, dim, distant time there were two hairy sort of persons who stood facing each other, scowling. They stood like this for some time until one, enraged beyond containment, let forth a strange growling sound which scared the other so much that it ran off.

The essence of communication had been discovered, as had its value. From that point the human race has developed to the stage where it has produced a number of different ways of communicating to complement that first form, vocalisation. One of the most important of these has been the development of the written word and, from that, the development of the machines that help humans create the words more easily: stone tablets, pens, paper, pencils and the typewriter.

This last one brought with it one of those inventions that is, at one and the same time, both incredibly clever and a pain in the ... That invention is the keyboard. The qwerty keyboard is the bane of many people's lives, especially as it has been universally adopted as the standard form of input device for the computer. The technical reasons for doing this are quite sound and when the computer was a machine that was only used by trained personnel (either operators and programmers or typist-oriented key-entry staff), the fact that the keyboard was being used didn't matter too much.

While the keyboard was being used exclusively by those explicitly trained in such arts, intimidation did not matter. Now it's different. Personal computers are everywhere. Workshop foremen use them, children use them, senior company executives use them. The intimidatory value of the keyboard has therefore become rather more significant.

A way around the keyboard was needed and over the last year or so technology has come up with some answers. The mouse is probably the most famous so far and, as far as it goes, is an excellent tool for moving the cursor around and entering simple commands by pressing. (Sorry, but it still has keys.)

Another device that has been employed of late is the touch-sensitive screen. This is actually a misnomer, for the thing is light-sensitive not touch-sensitive but, despite such split hairs, it allows the user to point to locations on the screen with a finger or similar apparatus and identify tasks, functions, windows or whatever is required. Again this is fine as a means of imparting simple instructions to the computer quickly and in a form that the user can readily comprehend.

Now, however, technology has come up with that which has long been predicted - the form of communications for which humans are rightly famous. Yes folks, the gabby computer has arrived. There are, to be fair, several add-on units that can be bought for the most popular personal computers which offer some degree of speech recognition and synthesis capability. But one of the first to come from a major manufacturer, to my knowledge at least, is the latest variant of the Texas Instruments Professional Computer.

TI has been in the speech technology business for some time, having produced such famous toys as the Speak'n'Spell educational unit. It also produced a speech synthesis add-on for its now defunct TI99/4a. These, it must be said, are just kids' stuff to what is now available. TI has produced a £1250 add-on board for its hard disk variant of the machine which really does have some interesting possibilities, and which could become the next generation of executive status symbol.

Early versions of the speech synthesis system tended to work only with small amounts of verbosity, and the digitised data for this was normally held in PROM on the same board as the speech processor. To limit the capacity further, the actual spoken sentences were constructed from individual words and phrases rather than long word strings. This meant that the recorded voice used in the first place had to be flat and uninteresting due to any intonation inevitably making a constructed sentence sound odd.

The TI system can now record a voice with any necessary intonation directly onto disk. It can then be read back for synthesis. On a 320k floppy, for example, TI claims it can record 20 minutes of continuous speech which can, as is the way with synthesis systems, be speeded up or slowed down without pitch changes as required. That may seem like a novelty but it has some uses.

The speech recognition system can identify some 50 different words in up to nine different vocabularies (that is, different individual voices). TI has produced a routine that allows the user to construct a file of commands which simulate the command keystrokes of any application program. Therefore it becomes possible to have the computer recognise you saying an application program command, 'scroll down' for example, and execute that command.

For a large number of applications this capability will allow quite a reasonable measure of 'hands-free' computing. With a spreadsheet, for example, it will be possible to have all the key commands and numeric data entry 'keystrokes' as spoken commands. Imagine it - the executive's status machine. You will sit at your desk and blithely say something like: 'Cell A4. 47321 point 68 return. Calculate.' The computer, with the right programming, will not only do the requested job but could also obsequiously muster 'I hear and obey, oh Master.'

TI has introduced, at the same time as the speech system, a networking capability with all the usual bells and whistles including an electronic mail facility. The company confirms that there isn't such an official product yet, but it doesn't take too much thought to see that it should be possible to combine speech with electronic mail - after all, the digitised speech is just another disk file which can be squirted around the network.

Here is the ultimate executive's toy. Send someone a text document and append to it a speech file with myriad words of comfort, clarification, excuse, and so on. This could have some really interesting possibilities.

end