=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= EDline Vol. 7, no. 9 (14 January 2002) Editorial mailing list (digest version) Published by the Electric Editors =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= Contents: Q & A [2re] Spaces for Plus-Or-Minus =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= ---[2]-- Q & A -------------------------------------------------- ** [2re] Spaces for Plus-Or-Minus Date: Wed, 12 Dec 2001 From: Lane Lester, llester@simpub.com What is the preferred spacing with the plus-or-minus symbol, as in 326±5? (I'm not sure that symbol will display correctly for all of you) Should there be no space on either side of the symbol, full space on both, thin space...? ---------------------- Date: Wed, 12 Dec 2001 From: Anna Beth McCormack, mccormack@goulburn.net.au You need space on both sides of the symbol. I prefer a full space, especially if there are other maths symbols nearby, but a narrow space may do. A 'thin space' is too thin. ---------------------- Date: Wed, 12 Dec 2001 From: Geoff Palmer, gdp@lineone.net I'd say whatever space would normally be used between the elements of a maths equation, which I think means a "thin space", but someone more skilled in typography might be able to offer a more precise answer. No space isn't a very good option (x+y=5?!). A normal inter-word space isn't an option either, because the association between "326" and "5" needs to be clearly visible. Some modern-day typesetters seem unable to set "thin spaces" of any kind. If you're faced with that situation, it might be better to close up. ---------------------- Date: Wed, 12 Dec 2001 From: Nancy Boston, boston.editorial@ntlworld.com It's usually a matter of house-style, I have clients who close up all mathematical expressions and others who insist they should be spaced either with a breaking or a non-breaking space. In my experience, if it's a space, it's a normal word space and not a thin space. If there's no house style to follow, choose your own preference (or the one that the author uses most often) but make it consistent throughout. ---------------------- Date: Wed, 12 Dec 2001 From: Susan Roberts, susan.l.roberts@saic.com In my experience I've always seen it with a space between the first number and the plus-or-minus symbol, as follows: 326 ±5 326º ±5º ---------------------- Date: Wed, 12 Dec 2001 From: Nancy Boston, boston.editorial@ntlworld.com That's fine if the ±5 represents a standard deviation or standard error. You could also have the ±5 in parenthesis to denote this and write 326 (±5) or 326(±5). In my previous reply I was thinking of an equation, which now I think of it again, is less likely (although not impossible) with a ± sign. ---------------------- Date: Wed, 12 Dec 2001 From: Eddie Kent, edlineek@aol.com Susan Roberts wrote: > 326 ±5 No. Won't do. + (etc.) is a binary operator and must be equally spaced. ---------------------- Date: Wed, 12 Dec 2001 From: Mark Hendy, markhendy1@compuserve.com Geoffrey Palmer wrote: > Some modern-day typesetters seem unable to set "thin spaces" of > any kind. Yes, and like to tell you they are state-of-the-art! I wonder if there is another name for such people. Round where I live, people whose professional agricultural practice is limited entirely to matters bovine, chemical and mechanical, to the neglect of soil quality, hedges, ditches, roadways, older buildings, standing timber, watercourses and so forth are said to be "not farmers, just cow-keepers"... Is there an equivalent term in typesetting, for the non-possessors of necessary so-called refinements? There certainly ought to be! ---------------------- Date: Wed, 12 Dec 2001 From: Eddie Kent, edlineek@aol.com Yes. Nerds. ---------------------- Date: Wed, 12 Dec 2001 From: David Ibbetson, isserlis@rogers.com I call them "typists" or "secretaries". ---------------------- Date: Wed, 12 Dec 2001 From: David Penfold, penfold@eps-edge.demon.co.uk In the old days of 'real' (Monotype) typesetting, each character such as ± carried its own space on either side, so that this was not a problem. I would say that, if you have the facility to create one (Word does not, unless later versions than Word97 have introduced the facility), then use a thin space. If not, then ideally use a non-breaking space. Note that if you use MathType (of which Word's equation facility is a subset), then you can indeed adjust the spacing as you wish. The problem with in-line equations, however, is then getting the spacing around the equation right. The moral of this is really that one needs to use a proper typesetting system, rather than a word processor if you want proper layouts! ---------------------- Date: Wed, 12 Dec 2001 From: Corinne Orde, c.orde@btinternet.com Geoff Palmer wrote: > Some modern-day typesetters seem unable to set "thin spaces" of > any kind. There's no excuse for this. It's laziness or lack of skill. ---------------------- Date: Wed, 12 Dec 2001 From: Susan Roberts, susan.l.roberts@saic.com BTW, I know about the non-breaking hyphens, and non-breaking space techniques, but am not sure about "thin spaces," unless it is changing the font size for one character. ---------------------- Date: Wed, 12 Dec 2001 From: Ian Kingston, i.kingston@ntlworld.com Typesetting software has a special 'thin space' character. As you've guessed, in word processors you have to fake this by setting a space in a smaller font. There are added complications, in that not all thin spaces are the same. In scientific work, something like 3 m s^-1 would require a thinner space between the 'm' and the 's' than between the '3' and the 'm'. ---------------------- Date: Wed, 12 Dec 2001 From: Lane Lester, llester@simpub.com David Penfold wrote: > In the old days of 'real' (Monotype) typesetting, each > character such as ± carried its own space on either side, so > that this was not a problem. Do you mean Linotype? That's the name I think of when I think of "real" type. Perhaps the Monotype company had something similar. > I would say that, if you have the facility to create one (Word > does not, unless later versions than Word97 have introduced the > facility), then use a thin space. If not, then ideally use a > non-breaking space. I use Corel Ventura which, like other good desktop publishing packages, provides control over character spacing and thin spaces, too. The manuscripts which come to me have already been edited and proofread, so I'm more concerned about layout than wording. > Note that if you use MathType (of which Word's equation > facility is a subset), then you can indeed adjust the spacing > as you wish. The problem with in-line equations, however, is > then getting the spacing around the equation right. I'm not a mathematician, but I don't think of the plus/minus (±) character as being part of equations. I see it most often in reports of measurements, indicating a range in which a value falls, e.g., average inches of rainfall: 46 ±5. I'm leaning toward this type of spacing because of this consideration. ---------------------- Date: Wed, 12 Dec 2001 From: David Ibbetson, isserlis@rogers.com David Penfold wrote: > In the old days of 'real' (Monotype) typesetting, each > character such as ± carried its own space on either side, so > that this was not a problem. Lane Lester wrote: > Do you mean Linotype? That's the name I think of when I think > of "real" type. Perhaps the Monotype company had something > similar. Side-bearings are up to the font-designer. ---------------------- Date: Wed, 12 Dec 2001 From: John Morris, johnjeff@meadowdance.org What is wrong with simply setting the kerning for the plus-or-minus character? This ensures that the line will not be broken between that character and the next and allows you to adjust the space to exactly what you want. In a "real" layout application, you would be able to adjust the kerning table so that every plus-or-minus character automatically received that extra kerning. I have not found this feature in Word. ------------------------ Date: Thurs, 13 Dec 2001 From: Eddie Kent, edlineek@aol.com David Penfold wrote: > The moral of this is really that one needs to use a proper > typesetting system, rather than a word processor if you want > proper layouts! Yes. In math, TeX ------------------------ Date: Thurs, 13 Dec 2001 From: Nick Hudson, hudson@c031.aone.net.au Geoffrey Palmer wrote: > Some modern-day typesetters seem unable to set "thin spaces" of > any kind. and Mark Hendy replied: > Yes, and like to tell you they are state-of-the-art! Ouch. Can I put in a word for maths typesetters. If they are starting from scratch (and charging the rate this involves) they have no excuse for bad spacing. There are relatively simple ways getting it right, e.g. the use of a program like Mathtype for the mathematical expressions. The spacing is correct and largely programmatic. This has become even easier since Mathtype was adopted as the maths component in MS Word, provided your layout program can handle the later versions of MSWord, which some do not. However, all too often typesettters are expected to tidy up a maths MS supplied on disk, all at a price which is little more than normal text. The problem is not that they don't have thin spaces or don't know how to activate them, but that it is a major job. Nobody has devised a macro which will do the job automatically. For what it is worth, here's what we do (and I supply this not FYI, but in the hope that somebody has a better answer): (1) Global search-and-change on + ÷ < > = (and any other characters used only as maths operators) and remove any space before or after. Then it's easy: global Find +, Change to thin space/+/thin space, etc. (Once for each operator.) (2) If the author has used the true multiplication sign, proceed as (1). But most authors use x or X, so these have to be found one by one and turned into ¥ (opt-Y). Then a one hit global to add the thin spaces and change the font to Symbol. (3) Minus signs are the real bore, as they can be hyphens, en-dashes or em-dashes, with every combination of space before and after. Find them all, and change them to an otherwise unused character, say #. Then global Find #, Change to thin space/en-dash/thin space. (4) Scroll through the whole text finding the long numbers and inserting a thin space every third position from the decimal point. This is an all manual operation. It is often quicker to reset the whole lot in Mathtype. The Indians will do this for less than US$5.00/page. I won't. ------------------------ Date: Thurs, 13 Dec 2001 From: Mark Hendy, markhendy1@compuserve.com TWIMC: For a thin space in Word just use a superior or inferior (that's old-style printerese for superscript or subscript) space. ------------------------ Date: Thurs, 13 Dec 2001 From: David Penfold, penfold@eps-edge.demon.co.uk Lane P. Lester wrote: > Do you mean Linotype? That's the name I think of when I think > of "real" type. Perhaps the Monotype company had something > similar. Yes, Linotype as well, although Monotype (in which one character was cast at a time, rather than a whole line as in Linotype) was much better for setting complex equations. Subsequently, systems, such as Penta and Atex, were developed for phototypesetting of equations, taking over much of the logic from the hot-metal systems and developing it further. (I should also mention TeX, which is probably the most developed system for setting mathematics, but is different in kind from programs such as Ventura and FrameMaker.) > I use Corel Ventura which, like other good desktop publishing > packages, provides control over character spacing and thin > spaces, too. The manuscripts which come to me have already been > edited and proofread, so I'm more concerned about layout than > wording. I am not Ventura expert (Ian K should be able to help), but you should certainly be able to add fixed spaces of the appropriate width around any character. You could experiment with thin spaces and number spaces. If I remember correctly, the usual space used in technical journals was either 16 or 24 units, with 100 units to the em. > I'm not a mathematician, but I don't think of the plus/minus > (±) character as being part of equations. I see it most often > in reports of measurements, indicating a range in which a value > falls, e.g., average inches of rainfall: 46 ±5. I'm leaning > toward this type of spacing because of this consideration. My personal opinion is that space around the ± should be equal, but it is really a question of the style you decide to adopt. Incidentally, I do not feel that there should be a line break before ± unless it the best of a number of poor options. ---------------------- Date: Fri, 14 Dec 2001 From: Alex Gray, wordworks@gairloch.co.uk David Penfold wrote: > In the old days of 'real' (Monotype) typesetting, each > character such as carried its own space on either side, so that > this was not a problem. They still do in PostScript and TrueType - it's usually referred to in font design as 'side-bearing'. However, this is just a starting point, and for most flexibility it is usually set to be quite tight (as it was with hot metal type). Beyond that, fonts may contain a table of 'kerning pairs', which is a list of spacing modifications (increase or decrease) to be inserted between specified pairs (the most obvious examples being such as AV, AW). Good wordprocessing and DTP software will respect these kerning pairs, and the really good ones such as Ventura allow you to modify the kerning tables in use for specific purposes (for example, I find that in Palatino I need to modify the spacing of numerals and commas for a good appearance in numerical lists). The Symbol font has 'proper' plus and minus signs with appropriate default spacing around them, and these are often useful to set occasional equations with the minimum of fuss and reasonable spacing. The plus and hyphen signs in most 'regular' fonts are not good for maths. In particular the hyhen, which is set too low, too short and set too tight (it is optimized for fitting in between lowercase letters, after all). The en-dash with hair spaces either side is an improvement, but it is too long and still set a bit too low in most faces. The minus sign (in the same code position as hyphen) in the Symbol font suffers none of these problems. > I would say that, if you have the facility to create one (Word > does not, unless later versions than Word97 have introduced the > facility), then use a thin space. If not, then ideally use a > non-breaking space. The most accurate answer to inserting an arbitrary space in most systems is to use a fixed-width space, and then adjust its nominal point size as required. Lane Lester wrote: > I use Corel Ventura which, like other good desktop publishing > packages, provides control over character spacing and thin > spaces, too. Ventura is superb, and offers unparalleled typographical control in the DTP world. However it has a few glaring errors, and one of them is the thin space. A 'thin space' (Ctrl+Alt+T, coded <|>) in Ventura is in fact a fixed width standard space (i.e. the same as the nominal space in the font used, but not subject to justification widening or narrowing). And it offers no hair-space as such. Sad, but true. Of course, there are several ways to achieve any desired effect in Ventura, but there are no simple fixed thin or hair spaces. The most reliable and controllable technique for arbitrary spacing is to insert an em-space and size it to the width required (because it is square, an em-quad set at, say, 3 points, is exactly 3 points wide also). To approach a tradiitional typesetting thin space as closely as possible, one would set an em-quad at -ne-fifthe the running text height. So, far example, in 10pt text in Ventura, an em-space set at 2 points will achieve a classical thin space. If you are entering this as plain text coding, it would be <_> (the P255 means return to default type height). Of course, if you are doing fancy equations in Ventura, then you would probably be using one of its two built-in equation setting systems, and these have different control over spacing, optimized for math setting. > I am not Ventura expert (Ian K should be able to help), but you > should certainly be able to add fixed spaces of the appropriate > width around any character. You could experiment with thin > spaces and number spaces. If I remember correctly, the usual > space used in technical journals was either 16 or 24 units, > with 100 units to the em. 100 units would be an impractical and inefficient system for mechanical use, as 100 has so few factors. I thought the predominant system for book setting (certainly the Monotype and Oxford University Press system) was 96 units, conventionally divided as follows: 96 units Em quad 48 units En quad 32 units Thick space 24 units Middle space ('ordinary' nominal space) 19 units Thin space 8 units One point 4 units Half point 2 unita Quarter point Note the point designations are based on a nominal Em size of 1 pica (12 points), but of course vary according as the font size, and are not fixed units as used in the context of type height. > Note that if you use MathType (of which Word's equation > facility is a subset), then you can indeed adjust the spacing > as you wish. The problem with in-line equations, however, is > then getting the spacing around the equation right. As you can in Ventura's equation setting. Ventura has two systems. One uses quasi-English coding of equations, such as "cos ( theta sub 1 ) + sin ( theta sub 2 )" which can be entered in a wordprocessor and will be correctly laid out using stanrd maths symbols and positioning. The other uses an advanced interactive designer dialogue base don Mathtype which permits the construction of very elaborate affairs, including complex calculus and matrix expressions, and offers incredible automatic or manual control over spacing. > I'm not a mathematician, but I don't think of the plus/minus () > character as being part of equations. I see it most often in > reports of measurements, indicating a range in which a value > falls, e.g., average inches of rainfall: 46 5. I'm leaning > toward this type of spacing because of this consideration. I think for routine reports and articles, it's not terribly important, but the hyphen does look ugly and inappropriate in almost all cases. The use of + and - from the Symbol font is a relatively painless way to greatly improve the look of matertial that uses these signs a lot in run-of-text equations and quantities. ---------------------- Date: Wed, 19 Dec 2001 From: Eddie Kent, edlineek@aol.com If a line break in an equation is inevitable it should come after an operator, never before. But if an equation comes lower than about the third line of a paragraph you can usually stagger enough takedowns to engulf it, or if it is too long, add a neutral word or two, or remove some. If all else fails, display it. In getting on for thirty years of doing this job I don't think I have allowed more than half a dozen broken equations. ---------------------- Date: Wed, 19 Dec 2001 From: Lane Lester, llester@simpub.com I'm confused. Is the plus/minus character ever found in equations? Is it really a mathematical operator? I never see it in equations. I see it only in tables of values or in sentences, where it serves to indicate that a numerical value falls between a certain range. ---------------------- Date: Wed, 19 Dec 2001 From: Susan Roberts, susan.l.roberts@saic.com I have seen the plus-or-minus sign in algebraic equations. I have not, however, run across it during my professional career. ---------------------- Date: Wed, 19 Dec 2001 From: Kathleen Lyle, edserve@klyle.demon.co.uk Lane Lester wrote: > I'm confused. Is the plus/minus character ever found in > equations? Yes > Is it really a mathematical operator? Yes > I never see it in equations. I see it only in tables of values > or in sentences, where it serves to indicate that a numerical > value falls between a certain range. It must depend what kind of material you are working on. I would have said it was quite a common symbol, although of course it is often used in the mean/range kind of listings you are referring to. Even there, logically is is functioning as an operator too and I can't see why it should be treated any differently from a plus sign or a minus sign. ---------------------- Date: Wed, 19 Dec 2001 From: Alex Gray, wordworks@gairloch.co.uk Yes, it most definitely occurs in equations. Most commonly it appears where the square root of a value is indicated, because the square root of any positive number has two values, of the same numerical size, but opposite signs. For example, the two square roots of 4 are +2 and -2. This would commonly be written using the plus-minus sign in an equation as [square root sign] 4 = ± 2 ---------------------- Date: Wed, 19 Dec 2001 From: Eddie Kent, edlineek@aol.com Lane Lester writes: > I'm confused. Is the plus/minus character ever found in > equations? Yes. For instance in the general solution to the quadratic equation. Take the number 4; this is a square number so it has a square root. But in fact it has two, +2 and -2. So you can write down sqrt 4 = +/- 2. That's about as compulcated as we need to get in a general list. ---------------------- Date: Wed, 19 Dec 2001 From: David Penfold, penfold@eps-edge.demon.co.uk Sorry to cause confusion. We are obviously talking about different things (or at least different contexts). Plus/minus is found in equations as well as in the context that Lane refers to. I won't confuse Lane any further by talking about the differences between in-text and displayed equations. However, in the sort of material Lane is talking about, ideally there should not be a line break either side of the plus/minus. However, I would agree with Eddie, that if a break is necessary, it should come after the operator, the reason being that, if the plus/minus is taken to the next line, the reader is not quite as aware that there is an error in the value (or a range). I hope that clarifies matters. Of course if one simply gives a value of plus/minus and then a value, i.e. no value before the plus/minus, then the break obviously comes before the plus/minus, just as it would if the operator were a minus. ---------------------- Date: Wed, 19 Dec 2001 From: David Ibbetson, isserlis@rogers.com It's a genuine mathematical operator. plus/minus x means either +x or -x. It doesn't of itself indicate a range:"7 plus/minus 2" means "5 or 9". A common use is in connection with square roots. There are two square roots to every positive number +x and -x Many schoolboys and schoolgirls will be familiar with (-b plus/minus (Square root sign) b^2 - 4ac) 2a as the two solutions of a general quadratic equation. ---------------------- Date: Fri, 21 Dec 2001 From: Michael Hall, babash@btinternet.com There is some confusion, here, between the plus/minus symbol that appears in mathematical equations (such as in the famous equation for the roots of a quadratic equation, where the plus and minus operators both produce equally valid results) and that where the symbol indicates an element of doubt in a value. They are not the same thing. The one should be treated as any other binary operator. The other should never be separated from its value. ---------------------- Date: Fri, 21 Dec 2001 From: Lane P. Lester, llester@simpub.com Yes! Perhaps because all of the equations with which I deal involve scientific quantities, it has never come up in my experience a situation where adding and subtracting produce the same result. [smile] Because of the ambiguity in this thread, may I beg your indulgence one more time and ask you to comment on the spaces for the second usage (element of doubt in a value) =only=. What is your preference, standard spaces on both sides of the ±, some other size space, a space only on the left side . . .? ---------------------- Date: Fri, 21 Dec 2001 From: Michael Hall, babash@btinternet.com I can't recall ever using the plus/minus form in a publication. My field was crystallography, and standard errors were always written in parentheses immediately after the value and with no spacing: 1.05(3) for 1.05 ± .03. It's a good many years since I published a scientific paper, so I can't be sure of the current house styles. If I was to use a ± in an error expression, I would tend not to place a space after it, if only to differentiate from its use as a binary operator. There is a slight problem, though, in that any decimal point immediately following it would become next to invisible and make proof reading a nightmare -- half a non-breaking space, maybe. On a lighter note. it is quite easy for plus and minus to produce the same result; try adding and subtracting a term that reduces to zero. ---------------------- Date: Sat, 22 Dec 2001 From: Lane Lester, llester@simpub.com This problem doesn't arise in scientific communication, because a decimal fraction less than one is always preceded by a zero, to insure that the decimal point is not missed. Example: 0.3. ---------------------- Date: Sun, 23 Dec 2001 From: David Penfold, penfold@eps-edge.demon.co.uk This makes good sense. However, as a one-time technical editor of Acta Crystallographica, I should note that (at least in my time) we never put a decimal point without a digit before it, zero in the case that Michael quotes, i.e. 1.05 ± 0.03. Even so, I agree with his comment about half a non-breaking space (a thin space?). ---------------------- Date: Mon, 24 Dec 2001 From: Nancy Boston, boston.editorial@ntlworld.com David writes (about .03 or 0.03) > However, as a one-time technical editor of Acta Crystallographica, > I should note that (at least in my time) we never put a decimal > point without a digit before it, zero in the case that Michael > quotes, i.e. 1.05 ± 0.03. In my experience, scientific journals tend to insist on a zero before a decimal point where the value is less than one (i.e. 0.03) but other disciplines may not. For instance, American Psychological Association style (also followed by the British Psychological Society) is to have no zero before the decimal point (i.e. .03) only if the value cannot be more than one (i.e. probabilities, correlation coefficients, and various other statistical values). However, a standard error or standard deviation, which is what a plus-or-minus sign is most likely to denote, can take a value greater than one, and so would have a zero before the decimal point under this system. Thus the problem of having something like ± .03 would never arise. ---------------------- Date: Mon, 24 Dec 2001 From: David Ibbetson, isserlis@rogers.com I was most certainly taught, partly by example, that all decimal fractions between -1 and +1 require a zero before the decimal point. --------------------- Date: Fri, 4 Jan 2002 From: Eddie Kent, edlineek@aol.com Michael Hall writes: > On a lighter note. it is quite easy for plus and minus to > produce the same result; try adding and subtracting a term that > reduces to zero. When a term reduces to something this is generally taken to mean 'at the limit', that is, never. Until never comes along plus and minus will always be different. On the other hand, consider: On a clock face with the hour hand pointing to six, imagine moving it forwards or backwards by six ! --- Thanks also to Ann Milligan for her contribution. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= END OF EDline 7.9 EDline homepage: < http://www.electriceditors.net/edline/ > ** The views expressed in this mailing list are strictly those of the individual contributors, and do not necessarily reflect the views of the moderators or of the Electric Editors. ** Articles (c) 2001, 2002, by individual contributors Design (c) 1996--2002 Iain Brown Compilation (c) 2002, Iain Brown / The Electric Editors =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=