page break

Typometry.

Every printer has found th need of some system by which the length of a given line in any size or face of type might be ascertained without the necessity of composing and distributing trial-lines. No such plan has yet been devised or adopted. In all well-arranged offices some scheme has been tried with regard to the fonts in use. One method is to set a given line in every type in the office, arrange the lines according to length, print a proof, and post it up as a scale. Another plan is to set the alphabet of each font, and paste it on the case or in a scrap-book. As a general rule this plan answers fairly well with old stock-fonts, but when new styles are added no one takes the trouble to complete the series of proof lines. What is really required is a system that will apply not to one office or series of fonts but to all; that will enable not only the compositor with the types before him, but the overseer or reader at his desk to ascertain certainly and promptly, without any experiment, whether a given line, in a given font, will fit a particular measure. To do this, all that is required is to reduce the width or « set » measurement of every letter to one common equation. We tried a plan some years ago which we find gives absolutely accurate results, without any printing of special proof-lines or alphabets. It gives the printer a good deal of preliminary trouble, but it is trouble well-bestowed. Our object in writing is not so much to bring the plan before printers as to endeavor to save them the trouble altogether by securing its adoption by typefounders, to whose department it properly belongs. We have found them ready to adopt useful and practical suggestions, and that which we now offer, would, we feel sure, be of lasting benefit, saving a vast amount of unprofitable labor in every job-office, and reducing to one perfect system the entire stock of job and body fonts of all standards at present in existence. Unlike the point-system of bodies, its adoption involves no alteration in scales or manufacturers' methods. Nothing more is required than the simple measurement and registration of measurement of every alphabet in the specimen-book. Typefounders possess the requisite scales; a printer must make his own. Set [unclear: to] the alphabet of the font—the 26 letters A to Z only. Omit points, ligatures, ornaments, and sundries of every kind, as these are all variable quantities, and the introduction of any of them would derange the whole plan. Set a line of pica quads (point system). the length of the alphabet. Make up the difference between even ems pica with pica or nonpareil spaces. Then multiply the number of pica ems by 12, and add in the fractional points. Let us say that the alphabet equals 47½ ems pica+a nonpareil thick space. 47½ × 12 = 570+2 (= the nonpareil thick space) = 572. This is the number of typographical points, the printers' smallest unit of measurement. This unit we propose, as it is now the generally-recognized standard of bodies, and the scale of measurement is everywhere accessible. Next measure the lowercase alphapet in the same way; then the small-caps (if any); then also the extra-large initial alphabet, should the font possess one. Print the specimen-line or label—any line you please—for attaching to the case or specimen-book, and add the figures representing the number of points. Take for example the following, as a specimen-label:

—Here we have set out the name of the letter, and the measurement, reduced to points, of the two alphabets. From the common equation, it is easy to deduce the length, almost to a hair-apace, of any given line. Take for example, the font above-quoted. The average width of letter is obtained by dividing the measurement-number by 26. If decimals are preferred, the alphabet may be treated as 25 letters instead of 26, the font-number multiplied by 4, and discarding the last two figures we have the number of points to the letter. For a single letter the result is worthless, but it is sufficiently accurate for any display-line, no matter how short, required in actual use. Divided by 26 the measurement-numbers of the font above-shown give for the capitals 15 1/8 and for the small letters, 8½; or decimally (with a + error amounting to 1/25) 15.1 and 8.8. In casting-off, the space between the words should count as a single letter, and the points (which are very variable in job-founts) 1/3, ¼, or whatever seems nearest. The calculation required may appear complicated; but it can be worked out in less time than it takes to set a trial line. We will show how it answers in the case of the type we have given. Our measure is 20 ems. 20 x 12·= 240·. This gives for caps 15½•, and for small letters 28¼•, (of course infinitesimal fractions are discarded). Here is a line of caps, set from calculation, and not from experiment:

In the case of a line of lowercase the only difference is that the number of caps has to be considered, and the allowance for the small letters reduced accordingly. Let the question be for example, ؟Will the line « The Sheffield Type Foundry » come into 20 ems? The letters and spaces, including four capitals, are 26. Proportion of four caps to lower case letters: 151/8 × 4.=62÷8½=7+, the whole equalling 29 letters, and our figures show the average for a line to be 28¼. The surplus 3/4 letter, (about 6·, or a nonpareil em), would if all the letters were average width, allow 6·, about a middle space, between each word; rather close, but quite allowable spacing. But there are four thin letters in the line, admitting of a reasonable space between the words, and of a point at the end. So it proves when actually tested:

Both these lines were set from the calculations given, and proved correct. And the computations required are really much more simple and self-evident than they appear when set out in detail. With no other aid than the single measurement-number attached to each alphabet, and the required measure, the display-lines of a whole book of advertisements or a type specimen-book could be set out without the necessity of touching a single type, and as readily as with the actual letters, if not more so.

It may be said that as in practice it is often convenient to reduce the equation to that of a single letter, that this should be the sign of the font, and that the Charlemagne, for example, instead of bearing the numbers 378·/221• should be marked 8·8/15•1. There are many good reasons against this. The alphabet-figure affords much better means of comparison than the letter-figure. To be sufficiently accurate, the latter should go into two decimals; there would be as many figures after all, besides the decimal-point, and whole numbers are always preferable to fractions.

As a means of fixing proportions of fonts the method is invaluable. Take for instance the following two faces by Stephenson, Blake, & Co.

Comparison of figures will show that the proportion is as 50 to 363/4. That is to say, taking ordinary job measures, a line that in the condensed would occupy 18 ems, would fill 25 in the bold. (The small fractions do not equal the ordinary give-and-take allowance for spacing). We try the experiment, and find that the result comes out exact.

Apart from the great advantage of being able under this system to accurately sketch out a display-job with the aid of a specimen-book alone, and with the certainty of each line occupying its appointed space, (which is at present impossible), our plan is of immense service to the job-comp who does not trouble his head with calculations or with preliminary sketching at all. He has for example, tried to set a display-line, measure 22 ems, and finds that the type is just too large. He has to get in 27 letters, and only 26 come in. He must find a type which fulfils the double condition of being suitable, and of filing she line. We have known four or five trial-lines set or partly set and distributed before such a result was obtained. With the equation numbers before him, no elaborate calculation is required. He looks at the font-number and finds that it is 264·. He finds another suitable letter—it is 266·; and he knows at once that instead of being smaller it is infinitesimally larger. The next, say, is 232·—this would involve wide spacing. He fixes at length on one marked 250·, and this comes out right—he has no need to go into precise calculations and find out that he wants 254·, when there is most likely not a font to that exact scale in the office. We feel certain that a few weeks' practice with fonts thus marked would give a compositor a better grasp of the geometrical proportions of types, than a whole term of apprenticeship in the absence of such a system. Not only this, it would enable the overseer without hesitation or mistake to fill those gaps in proportionate length of chief display-lines, which are found to occur in the outfit of the best-furnished offices.

We have given the American typographic point as the basis because it is the infinitesimal unit in most common use, and is very close to standard measures—in fact in all ordinary work, it is, as compared with the inch, within the ordinary limit of error in setting composing sticks, and cutting rules and leads. Founders who do not accept the system would no doubt adopt their own pica as the basis —it would be near enough for ordinary purposes. In France and Germany the Didot point (about one-twelfth larger than the American) would perhaps he insisted on; though as the Germans now cast type to American standard it would simplify matters if, for this purpose, the one point were made international, and the same figures had precisely the same value everywhere. If not, an initial would be required, as « 213·d » to indicate the Didot scale and types would require to bear a double figure, thus:

But a single scale would obviously be the best for all concerned.

Our scheme has the merits of extreme simplicity and universal applicability. Once adopted, new advantages would readily appear. For instance, there has been great debate, especially in the United States, as to a method of charging composition which should allow in proper geometrical ratio for the difference between fat and lean types. The old American system was to take them as the basis; but after a time the comps alleged—with what amount of truth we do not know—that some of the masters overreached them by having thin fonts cast with the letter m disproportionately wide. A number of schemes were proposed, some of them scientifically correct, but all without exception too intricate for daily use. By our plan, the single figure giving the equation of the font would be the basis of calculation, and no trick of casting special sorts could disturb it.

Of course the typographical point is appropriate only up to about 72-point body. For large job-type, the pica would be a sufficient unit, and for poster-type the inch. The principle and the advantage in all cases would be the same.

We will be glad to hear the views of printers and typefounders generally on this suggestion. Meantime it is open to any printer, who will take the necessary trouble, to make a trial on his own account. If the typefounders approve the idea, let them put the figures on their next specimen-sheets.—Typo (N.Z.), April, 1891.

page break