lout-users
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

First version of the `eq2' package available


From: Ludovic Courtès
Subject: First version of the `eq2' package available
Date: Wed, 18 Apr 2007 18:34:08 +0200
User-agent: Gnus/5.110006 (No Gnus v0.6) Emacs/21.4 (gnu/linux)

Hi,

The first version of the `eq2' package is available (for Lout 3.35).
This package is a (partial) rewrite of `eq' to better honor some of the
equation typesetting rules described by Donald E. Knuth in _The
TeXbook_.  It is available from:

  http://www.laas.fr/~lcourtes/software/lout/eq2
  http://www.laas.fr/~lcourtes/software/lout/eqf2

Additionally, changes to the User Guide, as well as an (admittedly ugly,
yet useful) example are available at:

  http://www.laas.fr/~lcourtes/software/lout/lout-eq2-3.35.diff
  http://www.laas.fr/~lcourtes/software/lout/eq.lout

The SHA-1 sums of these files are:

  a21c1fab3c411bb27ba651bf1332f8c2cafa2eea  eq2
  f40ec3570d3615615ba6f18292eb2baa0cd89c6f  eqf2
  02f161b93a827d80cc53b63027334bd2bf3db014  lout-eq2-3.35.diff
  c6de519dec1bccf936cffd5468a1cbe9400666c4  eq.lout

`eq2' implements the 4 styles described by Knuth (namely "display",
"text", "script" and "subscript"), along with the exponent "cramping".
In `eq', these styles were essentially "emulated", leading to
inappropriate layout in some cases.  In particular, layout of in-line
equations (i.e., equations appearing within a paragraph) was unsuitable.

Sums, products, integrations and other similar operations are now laid
out according to the current style.  For instance, the limits of a sum
in styles other than "display" appear to the right of the "sigma"
symbol.  Also, `big' is no longer needed when displaying a sum, product,
etc.: instead, the "sigma" symbol (or other symbol) is automatically
enlarged if needed.

address@hidden' now truly yields equations for in-line layout.  The produced
equation is unbreakable, unlike those produced by address@hidden' in `eq'. 
(Knuth
seems to imply that in-line equations _can_ be broken onto several lines
in certain cases, but I haven't seen any example or set of rules.)

The base font and symbol font used by address@hidden' can (in theory) be 
changed.
The base font is always explicitly installed, so that all equations are
laid out with the same font, no matter the context they appear in.  The
default is "Times".

`eq2' is not compatible with `eq'.  In particular:

  * `from' and `to' as stand-alone symbols have disappeared.  One can
    now use `atop' and `below' instead.

  * for `sum', `int', `prod', etc., `from' and `to' are now named
    parameters.  This means that their value must be enclosed in
    braces.  So, instead of "sum from i=0", one must write "sum from {
    i=0 }".

  * Binary operators like "+", "-", "<=", etc., now take a left and
    right parameter (this was needed to achieve proper spacing around
    them).  Their precedence follows that of the corresponding
    mathematical operators, as visible in [0], for instance.

    For example, to get:

        n + 1
       -------
          2

    one must now write:

       { n + 1 } over 2

    instead of just:

       n+1 over 2

    Hopefully, this should not cause too much troubles as this is
    consistent with mathematical conventions.


I have not updated the documentation to reflect the changes and new
symbols available in `eq2'.  The package itself could probably be
improved and I would welcome suggestions.  Hopefully it does not make
any regression compared to `eq' (modulo the incompatibilities).  Please,
do report issues such as regressions or broken layout.  Beware: it is
clearly not as good-looking as the equations in _The TeXbook_; I believe
fonts are partly responsible, though.

To my knowledge, it is the second user of address@hidden' and the first
user of macros for precedence.  ;-)

Speaking of that, I suspect there is a bug in address@hidden' that
prevents space style to be properly inherited; there's an easy
workaround, though (see the definition of address@hidden').

That's it!

Thanks,
Ludovic.


[0] http://en.wikipedia.org/wiki/Order_of_operations


reply via email to

[Prev in Thread] Current Thread [Next in Thread]