The first method is, by far, the fastest, but it is seldom the one which gives the best results. It is thus interesting when there is a need to compose quickly a lot of complex formulas, for a document which is not intended to be a quality publication.
There is, to my knowledge, five systems on Atari to compose "simply" mathematical formulas.
All these methods suffer from the same drawbacks for a quality composition:
This is the reason why direct creation of formulas under Calamus seems inevitable - only if for short formulas mixed with text, such as "square root of 2" or "1/2". The following part gives some advices and tips to gain time for these creations.
First some preliminary remarks. Even if the direct creation of formulas with Calamus SL 93 without additional modules is possible (on the other hand, it seems like masochism to do it with version 1.09n, with which you cannot use anchors), if you must compose rather frequently formulas, I would advise to use at least version SL 98, which makes work much easier, thanks to the possibility of subscript and exponent configuration, and the added possibilities of the frame module. The position module (or toolbox(+), but I think position is more interesting in this case) is also to be valued, to be able to move frames with the keyboard at the pixel level, but it is easier to do without. I will try, however, in the following, to say also how to use these methods for versions older than SL 98.
To be able to compose mathematics correctly, you must first have access to a Greek font, if possible devised for the font you use for the text, and also a mathematical font or set of symbols (integration sign, partial derived, radical,...). You will find in this archive (in zip format) a complete Greek font extracted from the calamaximus CD (I hope there will not be a problem of copyright, otherwise let me know and I will withdraw this font) and a Calamus document containing bitmap images in 600 dpi, already in actual size, of some usual mathematical symbols (computed with tex and imported via calipso). If you need a more important resolution (they should be fine for lower resolutions, if you do not ask for the optimal size for printing), let me know, and I will try to compute them. You will probably need also a script font; it seems to me there is one included with the fonts delivered with calamus.
Even if, when reading the following, it seems long, the result is worth it (in my opinion) and, with a little training, composition of simple formulas can be quite fast. As an example, every formulas in the book "Architecture de la matière" have been done this way.
Theoritically, for each writing font there is a mathematical or Greek font - thus, for Times there is Greek times, which is not the same as Greek Garamond,... - However, it is quite rare to find the Greek font corresponding to the font chosen for the text. Thus, one must resort to trickery.
One of the main reasons for the existence of the various Greek fonts lies in the difference of the relative size between uppercase and lowercase: in certain fonts, uppercase are twice as high as uppercase; in others, there are only one and a half higher;... Mixing without precaution two fonts quite similar in size gives nearly always ugly results.
The simplest to harmonize all that is to use two Greek styles instead
of only one. One will be use for the Greek uppercase, the other for the
For uppercase, it is very simple: the size of the Greek text is the same as the normal text size (provided you work with the uppercase sizes in Calamus, which seems however the most logical for me).
For the lowercase, some trial and error has to be done. The simplest: type some lowercase letters in normal text, then Greek text. Apply a huge zoom factor on these letters (use the magnifying glass), place two horizontal help lines positioned around the lowercase letters of the text (just like the lines in a copy-book). After that, vary the size of the lowercase Greek text style are also positioned on the Greek lowercase letters.
Use the same method to make small uppercase, if you do not have a font for that, or if you need to mix two different fonts in a text (for instance, if you must highlight a word with a special font). In the three cases, the result is less pretty than using a dedicated font, but the difference is rather small, and the result is far better than without the tip.
Of course, the same doubling must be done for each kind of style: titles, texts; exponents, subscript,... Fortunately, the size ratio between text font / size of the lowercase Greek font remains constant, which means you only have to use trial and error once.
The first problem comes from the default parameters of subscripts and
exponents: size of 50%, vertical offset of 50%. This leads to too small
subscripts and exponents, hard to read in small sizes (10 p) and
which oblige to use a large leading out not to overlap on the next line.
Under Calamus SL 98, correction is very simple, these values being configurable. As an indication, a size of 75% and a vertical offset of 0% for subscripts, -20% for exponents, gives good results. Warning, in your setup:
The second problem is to make subscripts and exponents on several
levels (subscript of subscript...).
With Calamus SL 98, it is quite simple: just create the right style with the relevant parameters for the size and the position of the subscript or the exponent. A size from 50 to 60% seems reasonable - after the second level, it is not always wise to continue to diminish the size (it all depends on the original text size and your motivation...).
With and older Calamus, you must fiddle with the vertical position using the manual approach codes; it is better not to try to modify the size as well, especially without Eddie and its search and replace functions.
The third problem is to make subscripts and exponents allocated to the
same character (the examples I can think of are rather from Chemical, as
CO_3^2-, but it should be used in maths also).
For that, there is to my knowledge only one solution: use the manual approaches, but horizontal this time. Do not forget to put the shortest of the two first, then the longest (i.e., for the above example, CO[Subscript style]3[Exponent style]2-[Normal style]) to avoid bad surprises...
There is two ways to make fractions in Calamus. The first is well adapted to small fractions (one or deux characters each side of the fraction bar, no fractions on several levels), the second for complex fractions - even if both may be used for every fraction.
The first method relies on the possibility of horizontal and
vertical move of block of characters in a text (manual approach functions,
called with Esc then Ctrl-cursor and Shift-Cursor in the text or text
styles module). The fraction bar is made of superimposition of minus signs
or long (alt-254) and short (alt-253) dash; the numerator is risen a few
points, the denominator lowered a few points and the whole is put on a
same vertical thanks to the horizontal approach corrections.
To prevent the justification to modify the horizontal position in case of corrections, do not put a space around the fraction bar. For an a/b fraction, the sequence is thus something like, in Eddie notation, [Manual approach: + 1p up]a[AM -1p vertical -5p horizontal]-[AM : -2p V -5p H]b[AM : +2p V]. Keep track of the exact number of vertical Ctrl-cursor and Shift-cursor, to keep the text aligned! Help lines may help.
Be very careful, afterward, when you alter the text directly in the text frame: it becomes quickly difficult to know exactly to what character the cursor corresponds, and a destruction of a part of the fraction is easily done. It is much safer, to alter the fraction or the near characters before and after, to use the text editor module.
It the same fraction is used often, it is very efficient to define a macro from the first fraction carefully composed. To select the whole fraction is quite easy with Eddie or Eddie light; on the other hand, it is a little difficult in the text frame. You have to be sure the whole fraction is in reversed video; it is safer to select the spaces around the fraction as well.
The second method relies on anchors. The fraction is made from a text
frame for the numerator (centred justification), a text frame for the
denominator (centred also) and from a horizontal line frame for the
fraction bar (the line lined up with the middle of the frame seems the
most practical to me; a thickness of 0.2 p for a font in 10 p
gives good results). These three frames must have exactly the same width,
equal to the width of the longest of the two texts - the position module
rather helps, but giving manually the coordinates may be sufficient.
Once the three parts of the fraction are completed, group the three frames and anchor them in the text.
This method is very handy when a centred formula including one or several fractions has to be done; it is nearly mandatory if there is to have parenthesis around the fraction, since these parenthesis must have the same height as the fraction and thus have to made specially (see below). Use a help line to correctly line up the fraction bar to the rest of the formula (at the middle of the =). If the fraction has to be inserted in the text, manual approaches have to be used to place the fraction correctly.
In both cases, it is often necessary to fiddle with the leading out if the fraction is not a centred formula. A ruler has to be placed at the beginning of the line preceding the fraction, with a greater leading out, then the normal ruler just after the fraction. Be careful, if the text has to be altered, it may be necessary to place again the first ruler to keep a correct leading out.
Algebraic measures are rather simple to make; it is sufficient to define an underscore style with correct parameters (with a negative distance from the base line). As underscore is normally not used otherwise in typography, and as (by luck) Calamus gives a correct underscore even when using subscripts and exponents, this presents no special problem.For all similar kind of notations (vectors, arcs, angles,...), the solution which seems the most efficient to me is to use anchors, which gives the immense advantage of following the text in case of modifications. To obtain a good result, you have to work with an absolute leading out (otherwise the gap between the lines becomes greater for the lines including a vector) with the leading out oriented to the top). The text under the symbol is next correctly placed by modifying the manual approach. If the pattern occurs often, it is very advantageous to define a macro to build the vector; use Eddie as much as possible to define the macro properly (include the anchor, the text, the approach correction and the possible changes of text style).
In the case of small radicals (square root of 2, cubic root of
a,...), the method which is by far the simplest is to create a group
of frames containing the radical symbol gathered externally (or drawn;
the maths document available here contains
one defined for 600 dpi), a surface frame (black rectangle) for
the horizontal line and one or two text frames for the value under the
radical and the power. The group is next anchored in the text (be
careful to use an absolute leading out to avoid surprises with leading
out) and the vertical approaches modified to have a good line up. The
first one to make takes a little long, but thanks to the macros the
following are obtained quickly.
Warning! If you must rotate this group, do not forget the radical symbol is a bitmap image, thus it can only be rotated by steps of 90°! So it has to be vectorised (with Speedline or manually) before.
Furthermore, if you insert the macros in Eddie, please not the anchored frames are a virtual copy of the same frame, thus be careful if you insert a root(2) and you want to alter it to a root(3), the frame must me unanchored to make the modification. On the other hand, if you enter the macro from the text module, Calamus will ask the nature of the copy which means no surprise.
If you must use radicals for larger formulas, the most reliable solutions are to compute again the symbol with an adequate software or to vectorise the symbol and change its shape. In the second case, I would advise to separate the "V" part from the horizontal line, in order to keep an homogeneous thickness.