Column 4, 8000 Plus 4, January 1987
DINOSAURS AND DASTARDS
History has a terrible grip on computer users. These little plastic boxes may be the latest in electronic niftiness, but we're still all stuck with the inefficient QWERTY keyboard ... originally designed to slow people down and avoid stripping the cogs of creaky Victorian typing engines.
There are word-processing equivalents of the QWERTY heritage, the most notorious being the dread WordStar. When computer buffs sneer at LocoScript, they generally drop a smug word about `real, industry-standard word processors like WordStar'. It's a lumbering ten-ton crane of a program, haphazardly designed and hailing from the days before arrow keys ... thus the long-term WordStar fanatic has a deep-seated belief that its traditional Alt-E, Alt-S, Alt-D and Alt-X are good, memorable key combinations for -- respectively -- Up, Left, Right and Down.
Want to stop editing in WordStar? All you have to do is remember that the `save' and `exit' functions are memorably located on the `Block Moves' menu, which is memorably reached by typing Alt-K. I don't recall just now whether K stands for Save, Exit or Block, but it sure is memorable....
Arthur Naiman, who wrote the best book about this program (Introduction to WordStar, published by Sybex), and made sure his contract with the publisher specified that he needn't use WordStar to write the book, said this: `WordStar is one of the most poorly designed word processing programs ever written -- a huge, elaborate farrago of kludgy patches, sort of like a Rube Goldberg machine gone berserk. All kinds of basic functions require disk access, thereby making the program fantastically slow....'
(`Rube Goldberg' is American for `Heath Robinson'.)
I've tried to use WordStar and I agree. Yet it lingers, because it's a `classic', available everywhere, and millions of people have learned its weird ways (new word processors are occasionally slagged off by computer-press hacks for failing to support the `universally accepted' cursor controls Alt-E, Alt-S ...). Just like QWERTY.
With `industry standard' opposition like this, don't be ashamed of LocoScript ... even if tempted to take the Locomotive Software boys aside and persuade them with a rubber hose to answer the Ultimate Question of Life, the Universe and Everything. Which is: `The advantage of a self-loading program which doesn't run under CP/M is that by use of direct hardware and firmware access it can theoretically run with blazing, mindboggling speed. So why...?'
Largely because LocoScript's huge character set requires graphics control of each dot on the screen ... lots more time-consuming than just using the 256 off-the-peg CP/M characters. Thus much of its slothfulness comes from the provision of things like Cyrillic and Greek which few people use.
LocoScript could itself become a QWERTY-style dinosaur, because everyone traditionally swears by their first word processor. I have a dark suspicion that IBM versions of LocoScript are being bolted together for those who plan to swap machines and would prefer a familiar program to a more powerful one. [I was right, too.]
My own first `real' word processor (not counting cassette-loaded Scripsit on a Tandy which wouldn't display lower-case, and EasyScript on a CBM64 with the famously naff 40-column screen) was SuperWriter. This is an efficient though slightly outdated program; out of nostalgic fondness I investigated and reviewed the PCW version released by Sorcim.
Something of my disappointment came across in the ensuing article for Another (Official) Magazine ... but not a lot, since the rudest bits were ruthlessly subjected to the CUT key. `Consummate stupidity ... naffness ... clowns ... wally practices' ... these were among the tactful reproofs which didn't make it into print.
This version of SuperWriter (since withdrawn, and no wonder) is a classic example of an OK program being made almost useless by lousy configuration for the PCW. Instead of adjusting it to the attractive 32x90 screen size, the makers require you to find and run the SET24X80 program before using SuperWriter. The `status line' information is supposed to be distinguished by reverse video display, but isn't. Likewise the print controls: instead of (+UL) and (-UL), SuperWriter uses a lower-case u, distinguished from a normal text u by special highlighting ... which got left out of the PCW version! You literally can't tell print controls from real letters unless able to `patch' the SuperWriter program and substitute distinctive symbols (that's what I did).
SuperWriter appears to have been adapted for the PCW by doing exactly two things: copying a generic CP/M version to 3" disk, and sticking a matrix-printed label on the front of the standard IBM manual to explain that this is the Amstrad version.
It's a pity: I still rather like SuperWriter. Unfortunately the above soul-chilling tale is nothing special in the sleazy world of software. Magazine letter columns are full of wailing about programs which require the 24x80 screen but don't bother to send the two character codes (ESC x) which set up this format; or manuals larded with references to appendices which have been ripped out because `they don't apply to the Amstrad'.
Next issue, the terrors of communications links. Armed with rod, gun and soldering iron, will our hero achieve brilliant success or second-degree burns? Aha.
Column 5, 8000 Plus 5, February 1987
C.E.T.I.
As every red-blooded SF or space fan knows, CETI stands for Contact with Extraterrestrial Intelligence. It's a good phrase for the awesome concept of using your PCW to signal across the illimitable void of incompatibility -- to communicate via feeble pulses of electrons with such hostile entities as Intergalactic Bug-eyed Monsters (another acronym explained) -- to boldly go....
Sorry, the SF writer in me gets a bit uncontrollable at this phase of the moon.
As threatened, I've been looking into ways of squirting text between the PCW and other computers. Suppose you've written a book in LocoScript and your publisher asks for an ASCII file on an IBM 5 1/4 " disk? Or suppose you could (like me) save retyping stuff if you could transfer it directly from 3 1/2 " to 3" disks? Assuming you own or can borrow a computer with the `alien' disk type, this is what you need:
First, your PCW requires the add-on serial interface which lets it talk to the outer world. Amstrad's own isn't too expensive, and can be installed in a few seconds without voiding your guarantee or hauling out the arc-welding gear: it fits on that wafer of circuit board sticking out of the back of the PCW. Two screws hold it there. Technical requirements: one Phillips screwdriver and the ability to rotate same in a clockwise direction.
The interface box announces its presence when you load CP/M, and doesn't otherwise get in your way.
Next, you must physically connect your computer to whatever else is waiting out there. The cables can be quite pricy, and I needed two (we'll come to that), so I decided to go stark mad, buy the bits and solder up my own. This proved much easier than expected, enabling me to save vast sums of beer money. If you can face the thought of using a soldering iron, here's the recipe.
Take a few metres of 3- or 4-core screened cable (only 3 cores will be used) and two of those arcane objects known to the Illuminati as `25-way D connectors', Consult your local electronics shop, or the ubiquitous Maplin catalogue. The PCW end of the cable needs a socket-type (`female') connector; at the far end, IBM-compatibles like the Amstrad PCW 1512 demand the identical connector, while other computers like Apricots may want the plug-type (`male') version. This is why I wanted two cables.
The actual wiring isn't too agonizing, as the pin numbers appear in tiny bas-relief on the connectors. Pin 2 at the PCW end should go via the cable to Pin 3 at the other end. Likewise, Pin 3 to Pin 2. The third cable wire links Pin 7 at each end. Finally wire together pins 4, 5, 6 and 8 at the PCW end, and then do the same at the far end. `Crossing' the links between pins 2 and 3 produces a cable suitable for talking to other computers rather than mere `peripherals' like printers.
After connecting this cable between the Amstrad add-on box and the `serial port' of the alien computer, there's the question of software to move stuff to and fro. No problem with the PCW itself: the famous MAIL232 program is lurking on your master disks and can be commanded by loading CP/M, putting in the LocoScript disk and typing MAIL232. Further details are in the booklet that comes with the interface.
My first experiment was with an Apricot and its ASYNC communications program, which like MAIL232 was thrown in free. Nothing much seemed to happen in the first few trials, until I tried a lower `baud rate'. This is a measure of how fast the information is moving down the wire, in bits per second. MAIL232's normal setting is 9600 baud, moving text at 1200 characters per second (one character equals one byte equals eight bits). Perhaps my rotten little wire couldn't cope with the traffic, perhaps the computers just weren't quite compatible at that speed. The problem vanished when I braked to 1200 baud, and text files starting moving across with gratifying ease.
(Many computers don't in fact transmit and receive at the `standard' baud rates claimed in manuals. For example, at the PCW's alleged 7200-baud setting, Amstrads will talk to other Amstrads but not to different computers since Alan Sugar's version of 7200 baud is somewhat idiosyncratic, at an actual figure of 7352.94!)
Even if the two computers talked the same language, I'd be nervy of moving program files between them: MAIL232 is a bog-standard communications program without fripperies like error checking, and sure enough I did lose odd characters from some transferred files. In text, such tiny errors merely get you mocked for leaving a `g' out of `mortgage' as I did in one column. Similar gaps in programs are liable to crash the computer, write rude graffiti all over your disk directories or initiate World War III.
Moreover, both programs and LocoScript files contain weird characters which simple-minded MAIL232 can't handle. You should use the `Make ASCII File' option in LocoScript, to get something which can be transmitted to distant galaxies.
Next it was time to try linking with a PC1512. This was harder, since Amstrad don't include any comms software with their IBM clone. A friend came to the rescue by bringing around his copy of Sagesoft's ever so upmarket program ChitChat for me to test: we shifted a few files in each direction and all seemed fine. For us cheapskates, though, I think public-domain `comms' programs are copiously available for the PC1512. Enquiries are under way. Our mighty radio telescopes are scanning the heavens. Beam me up, Scotty.
The Apricot is now obsolete and the PC1512 superseded; IBM serial ports often use a 9-pin and not a 25-pin connection (some models have both), and new comms systems have appeared -- such as LocoLink, which needs no interface box but does apparently require you to be running LocoScript on both computers. Adding a 3 1/2 " disk drive to the older PCW seems the best plan: you can then read from and write to IBM 3 1/2 " disks. See column 67.
Column 6, 8000 Plus 6, March 1987
THROUGH THE BARRIER
I staggered into the pub and cringed to see my usual drinking pals -- the directors of famously obscure computer firm Pangolin Systems Ltd -- holding pints. `We'll have to stop this,' I wailed. `There's a letter in 8000 Plus complaining that my references to pints in pubs are sexist because women are by implication excluded.'
`What a shockingly sexist assumption!' cried the female half of Pangolin, nudging her empty beer-mug significantly towards me. `Mine's another pint.'
The relative fewness of female computer hobbyists may be partly due to the deep-dyed chauvinism of this column ... but `hobbyist' is the key word. Flipping through other specialist magazines (model-making, yachting, motoring, gaming, fishing, and unnervingly many about blowing holes in things with guns) suggests that obsession with any hardware-oriented hobby is a largely male disease. Sociologists can take it from there. My one-time collaborator Charles Platt says the answer is simply that women are more sensible; I'm not sure whether this is an offensively sexist theory.
Unfortunately, without a touch of obsessiveness it can be hard to penetrate the barriers which surround the legendary Promised Land of Trouble-Free Home Computing. Magazines help a bit, except when the useful tip you need today has gone to ground in your heap of back issues covering the past eighteen months. When I first wrote a helpful piece for an Amstrad PCW magazine, I discovered there are people who want to reinforce the barriers from the inside....
`I saw your article on how to make a boot-up disk,' said this person who shall be nameless. (`Boot-up disk' is the insiders' way of saying `start of day' disk -- anything which automatically loads a program and gets going when inserted after first switching on.) `Good grief, they paid you for that? What a ruddy rip-off merchant you are. Everyone knows how to do that!'
He went on like this for a while, until I invited him to take a poll of everyone in the pub and find what percentage actually could prepare a start-up disk. (My deep apologies for being in a pub again.) The reply: `Oh well, that's different.' When he said everyone he didn't mean just anyone ... only real people, computer people, the sort of people who were used to working things out from a badly-written manual the way he'd had to. What he really meant was, why should all those parvenus with their nasty little PCWs have everything made easy for them?
Meanwhile, it's hard to imagine anyone working to strengthen the barriers from outside, to hamper their own journey to understanding -- but it happens. Those insiders who try to be helpful and give people a leg-up can often be defeated (and sometimes permanently soured) by that perverse animal, the outsider who wants to get inside without actually polluting his or her precious brain cells by understanding anything.
If this sounds peculiar and paradoxical, think about a first driving lesson which goes like this --
Instructor: `Now this pedal here is called the brake.'
Learner: `Stop that, STOP THAT, don't try to pull the wool over my eyes with all your technical jargon and gobbledegook, I don't want to turn into one of you motoring fanatics, just tell me how to drive the flaming car!'
People who take this attitude to computers usually end up with precisely what they've asked for: lots of long and (to them) meaningless sequences of keystrokes for doing this, that and the other task. One forgotten or miskeyed letter can leave them stranded and helpless, because they refuse to think about the underlying logic of what they're doing. `Stop explaining the page header menu to me, just say what I press to do this header!' Imagine learning a poem letter by letter, without reference to the words, metre or rhyme.
You can spot these users by the way they cross themselves and make signs to ward off the evil eye when confronted by the terrible A> prompt in CP/M. All right, I know, many struggling self-improvers do the same.
(An Instructive Aside For Newcomers: A> is the CP/M operating system's way of saying `Please type an operating system command or the short name of a program -- that is, the first name of any disk file whose surname is .COM -- followed by Return. Unless otherwise instructed I will apply the command to disk drive A and will expect to find any specified program on the disk in that drive. If you enter something I can't understand I will say it back to you, followed by a question mark to indicate my hurt bafflement.' If you have changed to the B> prompt by typing B: and Return, the message is exactly the same with B substituted for A. But CP/M doesn't actually display a long message like this because, first, it was written for Real Programmers Who Already Know, like my pal in the pub; second, CP/M is designed to work with a desperately small amount of memory and can't afford long helpful messages; third, all that screen-clutter of information would get very irritating after the first 2 1/2 appearances. Now you know.)
I now have to nip off and prepare a report for published on the new novel by a popular fantasy author. The typescript is printed in pale grey draft-quality dot-matrix, right-justified, on flimsy, unseparated continuous paper, without headers or page numbers ... all the things I've warned prospective writers against. The moral is that when you become a popular author you can get away with a lot. But not before.
The Controversial Bit
American computer magazines are nervy about offending advertisers: you learn to interpret their subtle codes like `... excellent lightweight word processor (translation: can't handle files of more than 1024 words). British reviews are blunter: for example, Dave Oborne in Another (Official) Magazine was quite acerbic about what seem to be worrying design flaws in WordStar 1512. Imagine my surprise on finding -- in the same issue -- an official user club newsletter which blusteringly attacks the review for being `subjective'. Translation: `We are going to call the flaws imaginary because we want to sell you the program and make money.' Business is business, but it's a shade ... unsubtle, isn't it? I certainly wish I had the privilege of angrily responding in the same issue when my own software gets slagged by an `independent' magazine.
Column 7, 8000 Plus 7, April 1987
WHY SOFTWARE ISN'T CHEAP
As an author, I always get depressed when some bore gets me in a corner and moans on about the price of books. The bore may own a VCR and several hundred tapes, a vast hi-fi system with thousands of costly LPs, cassettes and compact discs, a games console with endless cartridges at £10-£50 apiece ... but he or she is adamant that the modest price of a hardback book is Far Too Much. I even know people who've started small publishing outfits, confident of being able to clean up by undercutting the evil book barons. After exposure to the horrible facts of economics, these little companies either vanish rapidly or find themselves charging even more than the vile commercial publishers they distrusted....
As part of a tiny software house, I get the same depression from numberless magazine editorials about wicked software overpricing. Obviously it would be nice if no software cost more than £5. But I do gag a bit on reading an editorial saying sternly that £20 for a software package is a scandal on the order of Sodom and Gomorrah -- when simultaneously the advertising department writes to announce that one teensy quarter-page in which to tell people about said software will cost £200, £300 or more, depending on the magazine's pretensions.
How does that bloated, capitalistic £20 break down? [I updated all the following figures to January 1993 ... many still apply in 1997, but postage has risen significantly.] The first dent comes from VAT, which must be joyfully passed on to HM Customs and Excise (i.e. I have to act for them as an unpaid tax collector). Mumble mumble, count on fingers, the VAT-free amount is £20 times 100 over 117.5 ... that gives £2.98 VAT and £17.02 left in the kitty for the makers' corrupt personal gain.
Next, software has to be put on a disk. This may sound obvious, but I've met people who didn't twig for ages that the disk had to go in the drive. CF2 disks are no longer sought-after rarities, thank goodness, but there isn't such a glut as to allow massive discounting as with 5 1/4 " and 3 1/2 " floppies -- which can be bought in bulk for less than 50p each. Our little outfit could do slightly better by buying thousands of CF2s at a go, but we can't afford it. Right now they cost us £1.75 each. The kitty stands at £15.27.
A manual is essential, and people will be justifiably unhappy if this consists of one scrappy page of cryptic illiteracy. (Often the case. I have this theory that the Federation Against Software Theft encourages unreadable manuals, to deter bootlegging.) There are plenty of decisions here. The more manuals you print, the bigger the investment, the lower the unit cost, and the longer before you can revise the thing to cover software improvements. Cheap ring-binders may cost more to post and be harder to pack than pricier paperback-style binding. I reckon a decent manual costs at least £2.50 to produce. In the kitty: £12.77.
Postage and packing, that boring pair, come to around £1.20 despite frantic cost-cutting: envelope, address label, sellotape, heavy card stiffeners to protect the manual and disk, and those hideously overpriced bits of paper sold by the Post Office. I still haven't worked out the innovative fifth-generation technology which makes the perforations so much stronger than the stamps. The kitty: £11.57.
(Thanks to this emphasis on mail order, you're spared the pity and terror of discounting: dealers expect to cream off 35% to 50% of the gross, so feel free to recalculate my sums with a reduced initial kitty of £13 or £10.)
Estimating publicity costs is where this starts to get difficult. No publicity means no sales. Massive mega-publicity in all 5,271,009 Amstrad magazines means swift bankruptcy before you get any response. Let's take a modest advertising budget of £400 a month (i.e. quarter-pages in a couple of magazines.) These ads mostly produce requests for persuasive, lavishly produced brochures. An empirical statistic -- meaning I just looked it up in the records -- is that you get about one order for every four brochures sent out. A good brochure, with postage, costs at least 60p. The effective kitty, if you think about it, just went down to £9.17.
So here's our happy software firm, raking in a totally unjustified £9.17 per order -- after the first 44 sales each month, which merely pay for the ads. I think we'd better cut that advertising budget in half now! Hidden costs also remain. Credit card orders lose a percentage to Access or Visa; there are capital investments like company copiers, fax and answering machines, the manual-binding apparatus.... (See column 59 for where to get this.) The phone bill is hideously swollen with long business calls; the bank's promise `no charges if you stay in credit' doesn't apply to a business account no matter how low your turnover; and so on, forever.
The hardest thing of all to assess is time. Developing, maintaining, and copying the software; writing, rewriting and producing manuals; stuffing disks in envelopes; writing up the books and tax and VAT; answering endless written and telephoned queries; bit by bit it tots up to a 27-hour day. Most customers are friendly and patient, I hasten to add. Unfortunately a few belligerent or thick callers can spoil others' support by leaving the guy who answers the phone ragged and exhausted.
(We will not soon forget the man who twice rang us in a towering rage because our package hadn't reached his desk within two days, and ditto the replacement we apologetically rushed to him. Later, his secretary returned the extra copies of the software and we were interested to find on each a RECEIVED date-stamp showing that it had arrived the day after despatch. The [many expletives deleted] customer had been in too much of an angrily urgent hurry to bother checking his in-tray.)
Ultimately, for my own two-man operation, the question of pricing boils down to: what should a tiny software house charge for a task which somehow overflows into every second of free time? One of our very big suppliers justified its outrageous carriage by telling us the salary of the lowly employee who stuffed boxes into jiffy-bags. It was a damn sight more than Ansible Information's gross income.
This column has been a momentary aberration. Bear with me.
How to Madden a Software Company
• Ring before 9am with a long and complex technical query. • When ordering by credit card (which makes them groan, since they get less profit), always change a couple of digits round -- keep them on their toes! • Ring during prime-time TV viewing hours. Small outfits work from private homes and like to relax in the evenings, har har. • Never open the manual. Ring and complain first. • When you write, avoid quoting names, dates or reference numbers. `Some time ago I bought some of your software' will really make them hunt through the files. • Ring at lunchtime to demand help with the program that a pirate pal copied to you for nothing. • Get very abusive when the software company can't help with a program written by someone else altogether. • Cultivate an air of injured innocence as you complain, `It doesn't say anywhere in your so-called instructions that you have to switch the computer on....'
Column 8, 8000 Plus 8, May 1987
HOW TO ANNOY SOFTWARE CUSTOMERS
Last issue I tried to be provocative and have duly received 5,271,009 outraged letters from those who felt insulted -- i.e. all software dealers, all software users, and most small-press publishers. I didn't mean to offend the latter: the point was that very small publishers can't afford economies of scale and have to charge more than one might expect. Little did your columnist realize that one of the Old Barn hacks was also part of the tiny but classy SF publishing outfit Kerosina Books. Ben Taylor has gently remonstrated with me: I expect to be out of hospital any week now.
After explaining last issue how to irritate the people who flog software, it's only fair to tell them how to annoy you right back. Here, then, are ten tried and tested ways. Mind you, I have this terrible suspicion that the big software dealers know them all already....
• With PCW programs, always send the manual for the IBM version. A good ploy is to include a note saying, approximately, `Because you bought a miserable cheapo computer, you cannot use the triffic features described in chapters 7-15 inclusive.' A sprinkling of references to PC-DOS (the IBM operating system) commands will complete the process of demoralization.
• When someone rings your Technical Support department, always put them on hold for at least twenty minutes and play horrible tinkly music to them. Since few people can afford Muzak at peak telephone rates, this weeds out a lot of time-wasting queries.
• Always put at least two files on the disk which aren't mentioned in the manual (and omit one that is). Keep the customer worried and off-balance.
• Never call a bug a bug. Good alternatives are: `disk drive fault', `probable user error' and `quirk of CP/M'. When cornered, fall back on `undocumented feature'. This last term arises from the well-known fact that as soon as a bug is documented in the manual, it becomes a feature. Thus: `A convenient feature of the Grottyscript word processor is that you can reset the computer to power-on status at any time by simply pressing the space bar.'
• Establish dominance by making it clear to the customer that he or she is very ignorant. `You mean you've been using the PCW for three whole weeks and still don't know about RS-232 interface disk compiler mode I/O synchronicity overlay protocol debug incompatibility? We do have to assume some elementary knowledge, you know....'
• Be suspicious! This fellow who's rung with all the awkward questions obviously can't be a bona-fide user of your software if he's unable to quote (from memory) the full 64-character registration reference included with each package. Of course it hardly needs to be mentioned that this reference should be on a small, separate, easily mislaid piece of tissue paper, about the size of a bus ticket.
• Point out that although your advertisements do indeed promise full telephone support and advice, available 24 hours a day, the technical support team lives at the head office. `Just dial this Los Angeles number....' It's equally useful to insist that free program upgrades (i.e. to improved versions where the bugs have been cured or at least moved around a bit) are available only on sending the original disk and manual, in the original massive box, by registered air mail to Cincinnati.
• A refinement of the third point about undocumented files is to omit a vital step from the manual (`It is essential to press Shift-Exit-Paste twice and give a Masonic handshake in order to exit the Smartarse information menu') and hide it in a disk file. Most people have sussed READ.ME files, so call it something like PRGMAN1.MSG and don't use LocoScript format. The standard format for such files is WordStar's, since the ASCII text will then be full of funny characters which make it unreadable to non-WordStar users, har har.
• Why waste time looking for bugs in your programs when thousands of customers are eager to do it for you? If the accounts package doesn't round up the VAT properly, try it on the public anyway: there's always the chance that no-one will notice, in which case you needn't correct the program. At least, not until the VAT inspector has hauled a few customers off to jail: but there are plenty more where they came from.
• The Amstrad manuals tell people how to run the SET24X80 utility, load the specialist keyboard needed by your software, and use the CP/M program PIP to copy stuff to the M: disk as also required by your software. People with quite poor degrees in computer science have often been able to master the manual's descriptions of these processes in as little as six months. So when producing your own instructions, there's no need to explain how any of this is done. Paper costs money. Your ideal is a manual which will fit on one side of A4. Indeed, `Run the program' should be enough for anyone: after all, your software is self-explanatory, with lots of helpful messages like ERROR TYPE B34F DISK CALL BDOS ??? FILE ????????.$$$ Aborted.
These ten points are easy to master, and when you've done so you'll be well on the way to being a real, professional software manufacturer. Of course there are many further subtleties, like the importance of advertising your new product and getting in lots of cash orders before you start writing the program: but such techniques of `Advanced Cliveism' are beyond the scope of an elementary article.
Fair's fair. I've demonstrated how to annoy both dealers and customers. Next month, we'll discuss methods of annoying computer magazine editors. [Oh no we won't -- Ed.]
Column 9, 8000 Plus 9, June 1987
THE JOY OF MAILBAGS
My ego has just been expanded (unnecessarily, my friends will tell you) by a bag of fan-mail passed on from The Old Barn. Somewhere out there, a tiny but select group reads this column and each month manages to smile wanly at both my jokes....
One spicy bit was actually a typo, and cowardice may stop me ever again trying to type `leg-up'. Les Millgate pounces: `Please ask Mr Langford to forward to me the telephone numbers of those acquaintances of his who "try to be helpful and give people a leg-over".' For fear of sexism I'd better omit his PS restricting the request to `those who have lumpy things in their sweaters'.
But that's what this magazine is all about. To take computer virgins by the hand and gently lead them towards happy consummation -- first dispelling those fears of risky interfacing implanted by folklore, then advising on an ideal choice of software partner, and finally blossoming into steamy but oh so tasteful examples of the Joy of Computing. I only hope the innocent fun won't be affected by that recent Government warning campaign: is it an ominous sign that some firms already advertise transparent plastic protectives to be worn on your Amstrad's keyboard?
Onward, hastily, to one of those pesky first-time problems related by Paul Delderfield, who's having trouble with headers and footers. These recurring snags are torment for magazine editors. If they're covered every few issues, regular readers complain of repetition; if they aren't, newcomers feel let down. A quick look:
Problem one concerns the exact meaning of `header'. When it's just a bit of text which you want at the top of just one page, there's no point in using the special features for automatic insertion on every page, or every left-hand page, etc. Just enter the text as usual, in the appropriate place.
Problem two: setting up automatic headers is tortuous. First you press f7 `Modes' and Enter, giving a new screen where you can move between headers or footers and edit them freely, just as with your main text on the main screen -- e.g. right-justify with f5 `Lines'. From this screen hit f7 `Options' followed by f8 `Pagination' to set twiddly bits like initial page number. Successive use of Exit gets you back to the main screen. Eventually.
Problem three: headers and footers usually demand page numbers, inserted as noted on page 4 or with [+]PN These need special layout instructions, a bad lapse in LocoScript. Immediately after the (Page No) command you must add something like === to centre the number in a space three characters wide ... or, for example, >>>> or <<<< to right- or left-align it in a space four characters wide.
Fortunately, once this is set up to your liking, you can save a stripped-down document containing only your standard covering sheet (if any), headers and layout commands, as TEMPLATE.STD. Then it'll be popped in automatically whenever you create a new document in that group. (Obviously it makes sense to have different and suitable TEMPLATE.STD files in the different groups you have called LETTERS, PLAYS, NOVELS, and so on.)
Ken Hughes takes issue with my self-pitying moans about the problems of small software houses. `Instead of spending hours trying to invent new anti-copy devices, leave the program so it can be copied by anyone.' (Oi, what's this? I've never copy-protected a program in my life, as I believe that preventing people from making backups is a Bad Thing.) `Encourage users to give (yes, give) copies to friends, and post copies on Bulletin Boards. The manual should be included in the disk as a text file.'
The idea is `shareware', whereby you trust people who like your program to send you money for registration, upgrades and support. It's a lovely Utopian dream, and does seem to work for some (chiefly IBM) outfits in the States ... where lots more money and lots more computers are to be found. Three points bother me, though.
• Most PCW users I've dealt with seem pretty isolated, struggling in a vacuum to make sense of the system. They read occasional magazines. They certainly aren't tooled up for bulletin boards. To reach them, the poor old software firm still has to pay for those expensive ads.
• The system works against perfectionism in programming. It strikes me that the more user-friendly and bug-free you make your software, the less incentive there is for `shareware' recipients to send in that registration fee and claim support.
• Can shareware turn a fair profit in broke little Britain? The most famous IBM shareware program, PC-WRITE, suddenly started being marketed like ordinary software when it came over here ... I wonder why, or rather I don't. By way of research we visited a pal who reckons shareware is wonderful, and listened as he extolled his favourite packages. `Gosh,' we said, `how much did the amazingly cheap registration fees add up to?' He turned bright red! `Er, well, this one's American and sending dollars is a bother, and I don't use these two much, and I've only just got hold of that and, er, I'm still evaluating that other one....'
But I'm prepared to be converted -- the moment 8000 Plus gives up all that expensive bookshop distribution and merely puts its full text on a bulletin board, urging us all to copy it freely and send in the editor's salary if we really like it.
Another Bright Idea
Remember my complaint of a too-bright monitor? Hero reader Stewart McCall searched inside and found a built-in control. `There's a line going down from the tube to a black block. Between this block and the circuit board's mains lead connection is a variable resistor. Turning this will change the brightness range available. It's a case of trial and error: after each adjustment, turn on and start LocoScript.' No soldering iron is needed and no irreversible alteration is made ... but take care. I haven't studied the video circuitry, but there's a chance of high stored voltages even when the machine's not `on'. Cautious experimenters should make this adjustment a long-term project, altering the resistor setting only after the computer has been turned off overnight.
Column 10, 8000 Plus 10, July 1987
COPYRIGHTS & WRONGS
Suppose some advanced hacker trained a high-tech electromagnetic snooper on our PCW and stole the text of your precious best-seller as fast as you could type it in. Suppose -- switching to something that's actually happened to a friend of mine -- your disks were nicked before you could print out your epoch-making novel Son of War and Peace Has Risen from the Grave. What defence do you have against anyone who, so to speak, takes the words right out of your mouth and flogs them illicitly?
Few writers seem terribly clear about copyright law, especially when computers are involved. It takes the Writers' and Artists' Yearbook most of a page just to list amendments to the 1956 Copyright Act (star-studded successor to the Act of 1911). I've met people who produce private newsletters or science fiction fanzines and reluctantly send six copies of each issue to the British Museum Library or the Agent for the Copyright Libraries since `otherwise it isn't copyright'. Happily, they've got it wrong.
The `copyright libraries' are a red herring. They've been granted the right to demand freebie copies of everything commercially published in Britain, but failure to cough up doesn't affect copyright protection -- only your bank account, as the fines for non-compliance mount up. Amateur publishers have a loophole: the Act says the gratis copies must be in the same condition as those offered for sale. If you don't sell your publications commercially, you can thumb your nose at the libraries.
Copyright in printed stuff is fairly straightforward. You have the full protection of British and European copyright law the moment the story (or drawing, or limerick) is on paper. The work needn't be published or even shown to anyone else. If some low hack from the computer press sneaks a photocopy of your manuscript and snivellingly publishes it under his or her own name, the prison gates will loom -- if ,of course, you can prove it was originally your nicked epic. And as far as I can make out, US copyright protection is thrown in the moment you scrawl `© David Langford 1987' or its equivalent on the print-out. Apparently it has to be the real © sign: the (c) approximation cuts no ice in the USA.
`But,' I hear you wail, `I haven't printed out my novel!' Of course you haven't. No sense in wasting all that paper until you've got the hideous sexual perversion scenes just right, and checked the spelling of `formication'. Don't worry: any possible legal gap seems to have been plugged by the Copyright (Computer Software) Amendment Act of 1985. This essentially lays it down that copyright in software and thus other things normally kept on disk is identical to copyright in books. Once your golden prose or program is keyed into the new machine, it's theoretically protected against pirate publishers ... though not against your failing to save the file before you switch off, so watch it.
British copyright covers arrangements of words (or notes, or lines) but not ideas. If tomorrow some other hack publishes an article strangely like this one, my chances of persuading a judge to don the black cap would depend on how many actual phrases could be traced back to this column. Merely pinching the general idea isn't enough.
I was glad of this when years ago I wrote occasional pieces for Computer and Video Games, at the urgent request of my bank manager. My brief was to demonstrate how science-fictional ideas could inspire simple programs. Inspiration soon ran low, since I don't remember any SF novel which could credibly have been a source for the program called Attack of the Galactic Camels.
This was written to annoy my wife, who at the time was keen on camels and had a collection of stuffed ones, fortunately not life size. It was the work of mere days to set another little laser-armed phosphor blot jerking around the screen, zapping rogue camels at the player's command. (I was no as sensible then as I am now.) You could have knocked me over with a three-inch disk when the anguished letter of complaint arrived.
It wasn't the RSPCA who objected, but a computer outfit I'd never heard of, called Llamasoft. They were irate about evil Langford swiping the `camels idea', which was their very own, their own idea which was theirs. Their game was called -- with rather squalid sensationalism, I remember thinking -- Attack of the Mutant Camels. A friend cheered me up by libellously implying that said firm might be touchy about plagiarism because of this very program. In it, giant camels vaguely resembling landwalkers from The Empire Strikes Back lurched about the screen, as opposed to the giant landwalkers vaguely resembling camels which starred in the official Empire Strikes Back video game....
Armed with the Copyright Act and the Oxford English Dictionary, I hit back with the irrefutable fact that the first British emergence of what they called `the camels idea' would appear to be some time before either of our programs, in the Anglo-Saxon Lindisfarne Gospels circa 950AD.
After which, my next stunningly trivial C&VG program being all about falling down holes, I stayed up biting my nails in fear of a midnight knock on the door from the estate of Lewis Carroll.
All the above copyright © David Langford, 1987. Fantastically lucrative offers for film, TV, mineral or fishing rights should enclose stamped addressed envelope. Any attempt to show this column to someone who hasn't paid for a copy of 8000 Plus will automatically cause enormous thugs to break down the door and wave industrial-strength magnets all over your disks. Have a nice day.
For more information on copyright law, ask at your local library or write to The Copyright Office, The British Library, 2 Sheraton Street, London, W1V 4BH.
Just Like A Book
My favourite software copyright licence comes from Borland International of Turbo Pascal fame. They don't muck around with copy protection (I refuse to buy software I can't back up), and merely ask that you treat the package `like a book'. A book can be read by only one person at a time. So long as a Borland product is run only on one computer at a time you can move it between machines, make backups to your heart's content, even loan or sell the program to someone else ... all with Borland's blessing. This might sound too trusting: but Borland software is so good that serious users who `test drive' it are irresistibly compelled to get their own official copy with the fat and friendly manual(s). Being easy-going can be good business practice, it seems.
Column 11, 8000 Plus 11, August 1987
ARGUMENTS AGAINST
Several reasons why I held out for years against the lure of the word processor came back to me recently, as I sifted through old issues of SFWA Forum. This is a desperately secret gossip-magazine for members (only) of the Science Fiction Writers of America -- secret for good cause, as it would do these famous and unfamous authors no good at all if the public learned how boringly they can write....
The boredom would regularly escalate into actual pain when the subject discussed was the joy of word processing. `Look,' crowed writer X, `I was able to create this lengthy letter in only five minutes thanks to the wonders of technology. `Gosh wow,' writer Y would confirm, `I don't have to type two drafts of my book any more, and my output has trebled!' `It's fantastic,' writer Z would go on, `I can actually change every occurrence of a character's name by issuing one simple search-and-replace command....'
This kind of enthusiasm was usually accompanied by awestruck displays of everything the printer could do. You know -- three different kinds of italics, underlined boldfaced upside-down Greek letters, whole paragraphs printed right-aligned with a ragged left margin to demonstrate the equipment's limitless powers of naffery. Forum kept costs down by photo-offsetting straight from the original letters. As a result, my general impression was that word-processed output looked a right mess.
Then, when they gave examples of wonderfulness these always seemed so terrible. Writer X had prepared his letter in three seconds and printed out five copies with the merest gesture of his littlest finger. It was still a bloody awful boring letter, and you wished it had been harder for him to inflict it on the world.
Writer Y was able to produce more and fatter novels (I have a theory that the average book has been getting steadily longer as word processors continue to lessen the effort of redrafting). Unfortunately you couldn't help noticing that her fiction was visibly deteriorating as she presumably became more adept at tarting up a first draft just enough to make it publishable.
And writer Z, hooked on the major, sweeping changes you can make to a story once it's on disk, developed a weird stylistic jerkiness.... Well, think about it. Being able to change every occurrence of `Fred' to `Alfred' sounds great in principle. In practice, a simple Exchange has pitfalls. Unless you take care, other characters called Frederick and Freda will become Alfrederick and Alfreda, both of which have a nice exotic ring but may not quite fit. [Long after I wrote this came the famous story of the author who did a `David' to `Jeff' change and only just avoided going into print with a cultured reference to `Michelangelo's Jeff'.]
Aha, say the pundits, you should do the exchange not on just `Fred' but on `Fred' followed by a space -- which is fine except for the myriad cases where `Fred' is followed by a full stop, comma, colon, etc. Then there's the scene where the Voice of God says PREPARE TO MEET THY DOOM, FRED in block capitals.... And those are just the mechanical problems.
Other reasons why this often-quoted example is such a rotten one are more subtle and stylistic. Exchanging every Fred for an Alfred means that the rhythm of countless sentences will alter. Obviously a sonnet with the line Shall I compare thee to a summer's Fred? wouldn't scan if the chap were suddenly metamorphosed into Alfred. Similarly, a plain prose sentence that leads up to the crashing monosyllable `Fred' may sound all wrong when, like a cuckoo in the nest, Alfred substitutes himself. There again, `Alfred' -- unfashionable name of a long-gone King of England -- has all sorts of different literary associations from the matey, downmarket `Fred', and may well need to be written about in a different way.
The same criticisms apply to Writer Z's boasted ability to swap around vast chunks of text via the deeply wonderful Cut & Paste options. Careless minor alterations can damage the `microstructure' of prose, the rhythms which carry the reader from sentence to sentence. Major block-moved threaten the `macrostructure', the broader flow of paragraphs and pages which is generally a matter of logical rather than rhythmic development, ideas rather than words.
Sorry if all that sounds fearfully pretentious. No apologies at all if you think it's an attack on word processing. Where all those SFWA computer converts were going astray was in stressing the wrong things, and not pointing out how the technology can make you a better writer.
Yes, you can churn the stuff out quickly and correct it quickly: but there's no compulsion to print it straight away. Working with older equipment, a writer might bog down in exhaustion after two or three drafts; working with the PCW, you can do fifty or a hundred if every little titivation is counted as redrafting.
Still want to change a character's name throughout? Fine -- but use Find instead of Exchange, look at every sentence in its context before changing it, and rewrite with Alfred rather than Fred clearly in mind. Mutter the revised paragraphs under your breath until they sound right. Read and re-read the stuff in case your brilliant phrase `lickspittle running dogs of the repressive Thatcherite junta', on page 112, is upstaged by your having used it twice on page 111....
In short: word processors were hyped as offering the power to make gross, crude changed to what you've written. What the hype artists didn't stress is the boring fact that routine fine-polishing is also easier than ever before. If six drafts later it still doesn't feel right, you're not forced tom give up merely because the page is too full of marginal notes and scribbled-in corrections. Which is why I succumbed to the joy of word processors, and why, thinking of my erstwhile SFWA comrades' garbled enthusiasm, I remember a couple of lines from a Lewis Carroll pastiche: Although they wrote it all by rote / They did not write it right.
Column 12, 8000 Plus 12, September 1987
DIRTY WORDS
I started to write about computer jargon, but the words I was processing turned from green to blue when for no reason at all the disk in drive B started returning cheeky error messages. `Track 1, sector 1 missing address mark, ho ho' ... followed by that alarming choice `Retry, Ignore or Cancel?'
Shrewd PCW users will deduce that I don't use LocoScript for these columns (because when I did, all my italic markers got lost as your editor converted to ASCII format for typesetting). LocoScript gives different messages and won't accept a duff disk: you can go into an endless cycle of `Disk address mark missing', `Disk data error' and -- with the program now lying through its teeth -- `Disk has been changed'. Less `sophisticated' programs that run under CP/M may actually be more accommodating: what tends to go wonky is the beginning of the directory, and by patient pressing of I -- `Ignore the error and continue' -- it was possible to skip the naff sectors and load my current column file, which fortunately came some way down the directory. I optimistically typed R for Retry a couple of times first, in case the problem was just a shifting speck of dust on the drive head which might obligingly go away; I avoided C for Cancel since I didn't want to be thrown out of the program and back into CP/M....
The day was saved, but PCW newcomers might well blench at the jargon needed to relate even this simple tale. In two paragraphs I've smacked the neophyte round the head with `ASCII' and `CP/M', and much more confusingly have used many English words which have taken on new, computerish meanings: address, directory, drive, file, head, sector, track. Some people might even think they understand what's being said, yet be hugely or subtly wrong. (Wittgenstein wrote a lot about this, and he didn't even have a computer -- probably because LocoSpell would have complained bitterly about his title Tractatus Logico-Philosophicus.) At least we English can tell a computer program from a TV or theatre programme, though benighted Americans use the same word for both: I'm fond of similarly distinguishing a computer disk from its audio namesake, by spelling it with a K, but Locomotive and your editor think otherwise. [In this collection I get my own way at last.]
New gobbledegook keeps emerging: IBM have just announced that their new mainframe operating system features over 2000 new acronyms for users to learn. Other jargon goes way back. Bootstraps are little leather flaps used to help you get a grip when pulling on boots. Lifting yourself by your own bootstraps is a proverbially difficult athletic feat. Early computers, programmed by loading punched tape, faced an equally paradoxical problem: to read any tape required a tape interpreter program which, it seemed, would itself have to be loaded from tape.... The solution was called the `bootstrap loader' program, which in early days would be `toggled in' via switches on the front of the computer. You'd `boot up' a computer by loading this program which enabled it to load other programs. This is why, despite Amstrad's efforts to introduce the term `start of day disk', the big bad world of computing still refers to any disk which starts up the computer as a `boot disk'. It isn't very logical, since nowadays the equivalent of the bootstrap loader is kept on permanent ROM inside the machine, not on disk at all. In the computer industry they still talk about `punching' keys because all programs and data used to be punched on cards or tape, and ASCII character number 7 is called BEL since it used to ring the little bell on a teletype. (It produces a beep these days.) There's nothing like a dynamic new industry for producing hidebound traditions.
You have to decide how much jargon you can cope with. One friend of mine who came to computers late in life is determined not to let a single new word on board, and strenuously refers to disks as `tapes'. Another has soaked up neologisms even faster than he soaks up booze: offered a gin and tonic in the pub, he'll say, `No, I'm in beer mode tonight.' When asked `Do you really want another? There's a full pint in front of you,' he replies: `Ah, I'm double-buffering.' And his favourite joke is to croak `Pieces of seven, pieces of seven! -- Sorry, that was a parity error.' Please don't ask me to translate.
A Word from Alan Sugar
My pal John Grant (whose message to you all is: `I still don't know how to relabel worn-out keyboards, except with Indian ink followed by a coat of varnish.') recently finished collaborating with me on a new book. A successor to Earthdoom!, our spoof of disaster novels, it's called Guts! and sends up the horror genre. Since one subplot involves a teensy nuclear device, and since every chapter begins with unlikely but true quotations, we realized at once that hero entrepreneur Alan Sugar's famous remark was a must. You know, about how he'd merrily flog tactical nuclear weapons if there were a market for them.
Alas, my big mistake was to be courteous and ask permission. Fearless, hard-hitting Mr Sugar had no hesitation in telling his secretary to `respectfully request that you do not quote him in your book.' Another Amstrad publicity opportunity lost....
Column 13, 8000 Plus 13, October 1987
ESCAPE PLANS
As I type this, I'm preparing to get away from it all to the World SF Convention in Brighton -- which like August Bank Holiday will be but a lavender-scented memory by the time you read this. Computers will certainly have played their part in the event: Amstrads and many others were being ferried in by the carload to help run the mighty organization. Despicable authors like D.Langford will have prowled the hotel bars, cadging drinks from unwary 8000 Plus readers. And inevitably, the dark future of computers in SF will have been mentioned in countless convention talks and panels.
The science-fictional image of computers has changed in recent years. Admittedly, Isaac Asimov's robots still creak and clank and find new loopholes in the Three Laws of Robotics (his last book featured the addition of a Zeroth Law, which goes roughly `Stuff the other three laws -- the end justifies the means.') Arthur C.Clarke still churns out 2001 sequels about boring old HAL 9000. Authors whom I won't embarrass by naming are still writing versions of that old Fredric Brown story in which the ultimate computer is turned on and asked the ultimate question, and replies `Yes, now there is a God!' But the real action today is summed up in the newish word `cyberpunk'.
The master of cyberpunk is William Gibson, whose high-energy novels Neuromancer and Count Zero are recommended. In them, computers are much more personal and intrusive things than those we know: you plug right into them and bypass all those fussy CP/M commands or LocoScript menus. The typical Gibson hero is a sleazy, high-tech hacker who sits at the console with his brain jacked into the unreal world of `cyberspace', a hallucinatory realm of information transfers and software security in the world-wide data network. Down these mean computer banks a man must go....
There are deadly dangers there, Intrusion Countermeasures Electronics or ICE, which can feed back to burn out your brain if you try to hack into the wrong places -- rather more worrying than a `Wrong password' message from Prestel. It's all a sort of streetwise and nearly credible version of that uneven film Tron.
Cyberpunk SF is a very American product: the nearest thing to a British version is Gwyneth Jones's novel Escape Plans, which is fairly heavy going to begin with (lots of jargon and horrible acronyms) but opens out into a nastily persuasive vision of a future world where computer systems have been so absorbed into our environment that they virtually are the whole environment.
So much for the SF visions, in which we just think at our Amstrads and watch the exquisite sentences taking shape on the screen. In the real world, communications are so dodgy that not only can't the machines understand mere users, but most of us are left foxed by large chunks of the manuals produced by so-called experts in communication. The title of Gwyneth Jones's book reminds me that newcomers always seem particularly foxed by the words (don't all scream at once, now) `escape sequence'.
It's like this. In a feeble attempt at making it possible for all computers to talk to each other, text characters are stored in the machine as standardized numbers. This is the dreaded ASCII code, the American Standard Code for Information Interchange. On just about any micro, you can be pretty confident that a space will be coded as 32, the upper-case letters A-Z as 65 to 90, and the lower-case alphabet as 97 to 122. See the `Complete Character Set' table in your CP/M manual.
Characters with ASCII codes over 127 are a bit dodgy and vary from computer to computer. Those with codes from zero to 31 have special meanings: character 9 is a Tab, for example, and character 13 a Return (though not in Locoscript, which has its own perverse coding). And the tradition is that when you want to send a special control message to the screen (such as `clear screen') or the printer (such as `start printing in italics') it's done by pretending to display or print a sequence of two or more characters, starting with ASCII code number 27, alias ESC or `Escape'.
For example, the CP/M manual says that to print in italics, you need to print ESC 4, meaning character 27 followed by a 4 (which in the ASCII table is actually character 52!). From BASIC this could be done by LPRINT CHR$(27);"4"; ...the Escape won't be printed and neither will the 4, but the pair will be taken as a command to use italics for whatever's printed next.
When Protext asks for information for a new `printer driver' to run a different printer, you're expected to look up hordes of these boring escape sequences from the (usually impenetrable) printer manual, and type them in so that your word processor will know what codes to send when it's asked to do italics, underlining, elite or pica type, etc. Once the information is correctly entered, everything should be automatic: the word processor is then `configured' for your non-standard printer. It can be a long trial-and-error process.
If we lived in the world of cyberpunk, we'd just plug into the system, scan the manual and think the information into the computer. Come on, Amstrad, there's a whole new market waiting here.
Another Word from Alan Sugar
My partner in the tiny computer firm Ansible Information had a phone call recently: `Hello. We want free copies of all your software.'
The answer, not unreasonably, was: `No. Go away.'
`Ah,' said the voice on the phone, `but we're Amstrad!' The tone conveyed that all listeners should fall to their knees and worship.
Reasonably enough, my pal said: `Well, Alan Sugar can afford to pay retail price for a disk if he wants one.'
`Alan Sugar didn't get where he is by paying for software, sunshine,' said the far-off voice....
Column 14, 8000 Plus 14, November 1987
SPEAKING IN TONGUES
Once, being able to program computers was an awesome (if dreary) accomplishment. Friends would regard you with mingled amazement at your abilities and fear that you might bore them to death by talking in binary. Later things changed, and programming's public image became more like that of rock music: a field where teenagers made fortunes by writing games called Manic Space Goat Attack which did little for the human condition....
Since nowadays we all have computers, it's tempting to dip a toe into the water and play around with a program or two. The good news is that you can use Locoscript or your favoured word processor to write programs, provided you save the result in ASCII (plain text, no frills) format. A program is just a list of text instructions, after all.
This is handy, because one tradition of programming languages seems to be that editing program text is a horrible business. Most versions of BASIC offer `line editing' only -- that is, to edit line 100 you type EDIT 100 and then enter special and esoteric codes to make the actual changes. Mallard BASIC is unusually luxurious in that it actually lets you use arrow keys to move back and forth along the line. MicroSoft's `industry standard' BASIC won't let you move the cursor leftwards -- a hangover from the days of teletypes, when a line was displayed once and for all, and if you wanted to edit to the left you had to finish and start all over again with EDIT 100.
(Of the languages I've used on micros, only Borland's Turbo Pascal has really good, built-in, full-screen editing facilities.)
It's worth thinking long and nervously before playing around with programs, for two reasons. One is that programming's addictive: a simple exercise can swell to thousands of lines, by which time, if you've chosen an unclear language, you'll no longer understand large chunks of it. The second reason is Christopher Priest's Law: `You Get Used To What You've Got.' Hell hath no fury like the user who's still word-processing postcard-sized documents with one finger on an old Sinclair ZX81, and is told that the PCW is much better. He insists it isn't better. He's addicted to what he's got. Don't get addicted to a mediocre language.
What you've got is of course BASIC, some version of which comes free with most micros. It's easy to start with -- to add two and two and print the result you can enter PRINT 2+2, as opposed to the dozens of lines this might take in a long-winded `professional' language like COBOL, or in assembler. The trouble with BASIC is that it's also easy to lose track. You have to use numbered line references instead of meaningful labels when moving around the program, and the same applies to subroutines (bits of program designed to be used several times in different contexts, the way your built-in `Please Go Away' subroutine is equally useful for Jehovah's Witnesses or double-glazing salesmen). It's a real drag remembering whether it's line 10000 or 21334 that has the `print lewd limerick' routine. In a sensible language, couldn't we give the subroutine the memorable name LIMERICK and invoke it by name?
Oddly enough, this works in the Assembler supplied with the PCW (though never explained in the manuals). It's odd, because Assembler is primitive indeed: to use it at all, it's safest to be a computer fanatic with vast experience. You also need books explaining how Assembler programs can call CP/M functions, without which you can't even show the result of adding 2 and 2! I haven't space for an annotated Assembler program which could add and display two numbers. It's fast, it makes the best possible use of computer memory, and it drives you bananas.
There are languages which are interesting and useful in the world of big bad computers, but would be a bit eccentric for PCW use. FORTRAN is mainly for scientific number-crunching: you can give meaningful names to subroutines but not to individual lines, and printing the result of 2+2 would need three lines: one to add 2+2, one to print the result, and one to specify the format in which it's printed. FORTH is great fun if you like `reverse Polish notation' and know what a `stack' is: our 2+2 example in FORTH would go 2 2 + . (the lone dot in FORTH means `print number at top of stack')....
The serious contenders for the title of most popular, powerful and lovable small-computer language are Pascal and C. Personally I find C inscrutable: a pal uses it at work, and loves to explain how one line of C was recently passed round a roomful of professional programmers at ICL, not one of whom could decide what the line actually did.
So Pascal's my choice -- Borland `Turbo Pascal' for preference, this being available for so many micros that you can transfer your programs anywhere. When you've defined a `procedure' (alias subroutine) called Limerick, you can run it from anywhere in the program by just entering Limerick ... no line numbers to remember. Variables can all be given long memorable names, too. Pascal was originally designed as a teaching language which made it hard not to program clearly, and modern versions have the ease demanded by beginners together with the power needed by experts. Oh: in case you wondered, the Pascal command would be Write(2+2); (you can tell it's a classy language; nearly every line ends with a posh semicolon).
But nothing's more contentious than computer languages, and if I'm not here next month it may mean that a C devotee has stabbed me from behind with a sharpened pointer variable....
Genealogy Corner
All programming languages are ways of converting something vaguely intelligible into the horrible mass of numbers called machine code, which make sense to the computer. (Assembly language is just machine code slightly prettied up.)
FORTRAN and Algol are the great original languages. BASIC owes a lot to FORTRAN. COBOL is an entrenched but inefficient language used in business, where repairing duff COBOL programs is a never-failing gold mine. Pascal is based on Algol and has spawned a souped-up version called Modula-2. C is based on languages no one has ever heard of, called BCPL and B.
FORTH is one sidelong step away from Assembler. LOGO is unusual in being based on graphics (usually a non-standard `extra'). The trendy languages LISP and PROLOG are much loved by artificial intelligence buffs. The US Department of Defence wants everyone to use their new language ADA ... and there are far too many more.
Column 15, 8000 Plus 15, December 1987
READING FOR PROFIT
Most writers are compulsive readers, hopelessly addicted to the solitary pleasures of the printed word. (I keep waiting for our dear government to realize the perils, and plaster the country with posters saying FICTION REALLY SCREWS YOU UP, or warning of the terrible diseases you might get from sharing paperbacks.) Most writers, sooner or later, have a glorious moment of revelation when they find that one can lounge around reading books and get paid for it....
One of the reasons for my being a bit dogmatic about manuscript presentation -- see several previous columns -- is that when not at the keyboard, I intermittently suffer through all too many grotty manuscripts. If you want to lounge around earning ridiculously tiny sums of money, try the humble calling of `publisher's reader'.
The background is like this. Each year, far too many books appear. Those which are published are the mere tip of the iceberg, the thinnest possible skim of cream atop the vast churning unpublishable torrents which pour with terrible fluency from tens of thousands of Amstrad PCWs. Editors haven't time to read all the unsolicited stuff from unknown authors: they reject some at a glance for being handwritten, typed single-spaced without margins on both sides of translucent paper, or sabotaged by an inept covering letter. (`This is a totally new Sci Fi idea, its all about a huge Meteor weighing tons of light year's which is going to smash right into Earth's orbit ... OR IS IT??!!') The odds are that, while the full-time editor gets down to the serious work of copyediting some new Jeffrey Archer coprolite into readable shape, the brilliant novel by unknown you will be farmed out for a freelance reader's report.
The lowly reader is thus subjected to the real dregs. These haggard beings gather sometimes in pubs (where, in deference to the complaints of alcohol-hating 8000 Plus subscribers, they only ever drink slimline tonic water) and swap anecdotes about legendary grot. One well-known fantasy author, for example, apparently wrote a book which has never got past the publisher's-reader stage, being called Mercycle and dealing with the exploits of mermaids on bicycles. [This long-rejected title eventually appeared. I hope it was revised a bit.]
Before you all burst into tears at the thought of my sufferings, I'll admit I'm lucky enough to report mainly on writers who are publishable -- usually the book's been sold in the USA, and a British outfit wants an opinion. You see some funny things:
A high-tech author whose name is synonymous with glittering computerized SF still bashes it out on an old manual typewriter, the typebars so out of alignment that you'd think the writer was using a pneumatic drill with the other hand.
Locoscript may have its limitations, but (after the embarrassing early bugs of Loco 1) there's never been any trouble with page numbers. It was an author whose word-processing software alone cost more than a PCW who turned in a script with un-numbered pages....
A hefty book printed out on a swish laser printer (text quality as good as or better than this page, with real italics and boldface) suffered from a little software flaw: the author loved to show off with long passages of italics, and whenever one of these went from one page to the next, they somehow slipped back to ordinary type. (A problem for the copyeditor rather than the reader, but I maliciously noted all the times this happened in my report.)
Anticipating the paperless office, one author sent in a disk rather than a printout. When the postman bends an ordinary MS, legibility is rarely harmed; when he tries to bend a 3" PCW disk, it usually puts up a successful fight; unfortunately this disk was one of the limp 5 1/4 " monsters favoured by IBM and the PC1512. Through brilliant computer skills I eventually recovered the file with the novel, only slightly creased....
So much for anecdotes -- although my favourite computer-cum-publishing story is too good to omit despite having nothing to do with the toils of readers. Famous author X had the bright idea of arranging for the little printing firm up the road to typeset straight from his disks, thus saving the publishers staggering sums of money! Presumably the little typesetter wasn't frightfully efficient, since the unamused publishers later worked out that the book had cost them more than boring old conventional typesetting would have. This was also the author who made his alien speech authentic by cunning use of Exchange: he would write `rabbit', say, throughout the text, and when the story was finished the word processor would change every mention of rabbits to the more science-fictional sm'eerp. Please do not all imitate this technique.
What the poor sod of a publisher's reader hopes for is legibility (new ribbon, high quality print, and don't use 17-pitch), literacy (which lies between you, your conscience, your dictionary and Fowler's Modern English Usage) and liftability (a typescript weighing six kilograms must be separable into bite-size chunks for actual reading). The read-through by a conscientious editor or publisher's reader is the one time you can rely on the undivided attention of a professional at whom you are not actually pointing a gun. If the reader has had a retina detached by the attempt to follow faded text, and is also worrying about blood poisoning thanks to the jagged gash torn in one hand by your amateur job of stapling, he or she may not be totally impartial when reporting on your masterpiece.
On the other hand, clipping £100 in used fivers to page 94 (which Brian Aldiss told me would definitely help with the Booker Prize, at least while he was a judge) doesn't necessarily work either....
Unsolicited plug: you can learn lots about the pitfalls of novel-writing from Christopher Derrick's Reader's Report (Gollancz, 1969) -- wise advice from a publisher's reader who's seen it all. As the date indicates, computers do not feature; the warnings are still horribly true. Try the library.