Two Hours in a Wet February

Dear friend,

The riddle first, as the tradition demands. Back in the old town a man bought his tools of a maker he could walk to, and if the handle worked loose he knocked at that door and found the same man at the same bench. In your own century the maker is far off and may not be a man at all, which is a thing worth learning early.

Two electricians in your own century, the same vans and the same certificates to issue.
Both wanted the same program for writing them, and both had a free fortnight to try it.
One read the list of what it would do, found nothing wanting, and paid on the Tuesday.
One spent that fortnight sending the makers one awkward question and waiting with his watch out.
Same trade, same fortnight, same money a month.
In the wet February that followed, which of the two was issuing certificates again by the Tuesday?

Hold your answer. Story time.

The Story

Two electricians in your own century, each alone with a van, each certifying domestic work in the same two valleys. Both had tired of writing certificates out by hand of an evening, and both wanted the same sort of program to do it, the kind that holds the test results and sends the finished thing to the householder and to the scheme. Two firms sold one, each offering a free fortnight before any money changed hands.

The first man did what a careful person does. He set the two lists side by side at the kitchen table on a Sunday night and went down every line with the attention he would give a consumer unit. One carried four things the other did not and was two pounds a month dearer for them. He signed up on the Tuesday. Nothing in that was foolish: the list was accurate and all four things worked.

The second man spent his fortnight differently, and to a looker-on he spent it idly. He opened both trials, put a day's real work through each, then on the Wednesday at ten past nine in the evening sent both firms one question. He had issued a certificate with the wrong date on it, he said, and it had gone already to the householder and to the scheme. How was that put right, did she get the corrected copy of her own accord, and what would the record show. Then he noted the hour and went to bed.

One firm answered at twenty to eight next morning. A person with a name at the foot of it, who had plainly understood the question, gave the two ways it could be done, said which the scheme preferred, and warned that the corrected copy did not go out of itself. The other acknowledged him inside the minute, which is to say a machine did, and sent him to an article on how to issue a certificate, which was not what he had asked. On the fourth day a person replied, with an answer the first firm had already shown to be wrong. He bought the firm that had answered, whose program was the plainer.

That February an alteration at the scheme's end changed the numbering of one class of certificate, and for four days neither program would send that class at all. The first man found out on a Monday with two jobs tested and a householder waiting on her paperwork to complete a sale. He wrote to the makers, had the machine's receipt inside the minute, and then four quiet days, looking each evening at four things that all worked and none of which was any use to him. The second wrote to the firm that answered and had a reply in two hours: it was broken, it was being mended, and here was how to issue that class by hand meanwhile. He wrote three by hand on the Monday and two on the Tuesday.

The Lesson, Since You've Earned It

Before you pay for a tool, send its makers the question you will have on your worst day, and buy the answer rather than the program. A list of what a thing does describes the good days, on which almost any of these programs will serve. It says nothing about the day the work stops.

Consider what you are contracting for. A program paid for by the month is not a purchase but a standing arrangement with a firm, and the firm cannot be inspected as the program can, only sampled. A question sent before you pay is the sample, and it costs you ten minutes.

Now the part that makes the test worth the trouble. The answer you get before you pay is the best answer that firm will ever give you. They want your money and have not got it, and nothing in the arrangement will ever tilt so far your way again. So a reply that comes slowly, or from a machine alone, or from somebody who has not read what you asked, is not a bad day at their end. It is the ceiling, and it will not rise in the third year when you are one more small account paying eleven pounds a month. The argument cuts the other way too: a good answer before the money is no guarantee of one after, since firms are sold and support desks thinned to mend a margin. The test is a screen and not a promise. It cannot tell you who is excellent. It tells you cheaply, before you are committed, who is not. It is the same habit as the halt you press once before you need it: what you lean on in an emergency must be exercised when there is none.

Two cautions. Speed is not the measure. Grade the replies by the clock and you will choose the firm with the quickest machine, when a reply inside the minute answering a different question is worth less than one at eight next morning that answers yours. Hold nothing against a firm for a machine acknowledging you first, provided a person came after. The second caution is to ask the right question. Not how to do the main thing, which is in every manual and will fetch you a link, but the correction, the undoing, the getting of your own records back out again.

Three questions, for the week in front of you:

1. What is the question I will have on my worst day? Write it out as you would at nine at night, tired and wanting an answer. It will serve for every tool you are sold.
2. Who answered me last time, and how long did it take? Run the test today on what you already pay for. If you cannot remember reaching a person you have no arrangement but a direct debit, and a payment you renew by saying nothing is worth one awkward question a year.
3. What would I do for three days without it? Write the by-hand method on one sheet and keep it where you will find it. The second man's two hours served him because he knew what to do with them.

The Answer to the Riddle

In the wet February that followed, which of the two was issuing certificates again by the Tuesday?

The second, and he had the plainer program and the shorter list. Nothing was wrong with the first man's software. Every feature he compared so carefully worked throughout the four days he lost. What he never compared was whether anybody was there. A tool you cannot ask a question about is not a tool you own but a wager renewed on the first of every month.

The verse:

The list tells you true what a program will do
on the mornings when nothing goes wrong.
Ask them the hard thing before you have paid,
for that is their best, and their best is not long.

Until next time. Go and send your most awkward question to the makers of the one program you could not trade a week without, and note the hour you sent it.

P.S. The same test is being run on you, oftener than you suppose. The awkward little enquiry that comes before anybody has committed to anything, asking what happens if the date must move or the thing comes back faulty, is somebody doing to you what the second electrician did. Your answer is his only evidence of how you behave when something goes wrong.

Also in the archive: Still Setting It Up at Christmas, The Thing He Bought Twice and The Fortnight in Pencil.