annotate lisp/calc/README.prev @ 61527:020a384cb96d

(DONTCOMPILE): Remove list. (compile, compile-always): Fix comments. (update-authors): Load the library in which batch-update-authors is defined.

Summary of changes to "Calc" Preceding 2.00
------- -- ------- --  ----  --------- ----


Version 2.00:

 * Changed to compile calc-macs/-maint, to allow "cp *.elc new-dir".

 * Improved calc-describe-bindings to avoid showing redundant ESC maps.


Version 2.00 beta 3:

 * Removed version numbers from most .el files to reduce size of patches.

 * Added a "calc-version" command.

 * Changed `M-# ? ?' to allow for modified describe-function.

 * Changed date parser to accept "Sept" as an alternative for "Sep".

 * Inhibited answers to exercise from showing up in table of contents.

 * Changed Makefile to say "texindex calc.[cfkptv]?" to avoid "calc.el".

 * Fixed up the Makefile in various other ways.

 * Rearranged banner at top of `h h' command's output.

 * Changed "make summary" to print "Calc Summary" on the title page.

 * Added "IntegSimpRules".

 * Added `M-# :', `M-# _', and `M-# Z' options.

 * Changed `^' to evaluate "[-3..-1]^-2" properly.

 * Improved `f g' to give symbolic answers for, e.g., 101:2 and -3:2.

 * Fixed a bug where `h k RET' didn't find the right place on the page.

 * Fixed a bug that formatted "x*(y ? a : b)" as "x y ? a : b".

 * Fixed a bug where defmath translated (< x 0) as (math-posp x)!

 * Fixed a bug that prevented quick-calc from working sometimes.

 * Fixed the `z ?' bug again (maybe this time for good?).

 * Fixed a bug in which `V ^' (vint) was just plain wrong, wrong, wrong!

 * Scanned for and fixed remaining bugs relating to autoloading.


Version 2.00 beta 2:

 * Changed "make info" to try "make texinfo" if "makeinfo" not found.

 * Changed to "New for Calc 2.00"; texinfo.tex chokes on apostrophes.

 * Added List Tutorial Exercise 14 (just in case there weren't enough!).

 * Added a discussion of the `Z F' command to the Programming Tutorial.

 * Improved `H a f' not to lose info if input is partially pre-factored.

 * Improved simplification of expressions like sqrt(3) + 3^3:2.

 * Changed Big mode to omit "*" in expressions like 2 sqrt(3) 5^3:4.

 * Replaced European date format D/M/Y with D.M.Y.

 * Changed `a N' and `a X' to consider the endpoints of the interval.

 * Fixed a bug where TeX mode made c*(1+a/b) look like a function call.

 * Fixed a bug formatting top-level evalto's while using selections.

 * Fixed a bug that caused `z ?' to crash.

 * Fixed a bug where `Z F' broke for argument names "t" and "nil".

 * Fixed several bugs relating to autoloading.


Version 2.00 beta 1:

 * Added "What's new in Calc 2.00" to the manual (summary of info below).

 * Added support for many GNUPLOT 3.0 features.

 * Tweaked the Makefile and calc-compile a bit more.

 * Modified to work with Zawinski's/Furuseth's optimizing byte compiler.

 * Modified Calc to garbage-collect less often (raised gc-cons-threshold).

 * Changed quick-calc to avoid autoloading so many parts of Calc.

 * Changed Calc subfiles to work properly if not byte-compiled.

 * Renamed `M-# s' to `M-# j', made `M-# s' be equivalent to `h s'.

 * Changed calc-quit to avoid reapportioning space among other windows.

 * Added `M-DEL' (calc-pop-above) key, to DEL as LFD is to RET.

 * Added `{' and `}' to scroll vertically, analogous to `<' and `>'.

 * Added `m t' for "total" algebraic mode.

 * Added `d , \' option to group digits with "\,".

 * Improved support of "prime" accent in "eqn" language mode.

 * Changed macedit's read-kbd-macro to accept a string argument in Lisp.

 * Changed calc-check-defines to use a more concise run-hooks linkage.

 * Changed auto-why mode not to say [w=more] if next msg is not urgent.

 * Made `a d' able to differentiate "a?b:c" and "a_i" formulas.

 * Changed probability dist. functions to work with `a f' and `a d'.

 * Added special constants "phi" and "gamma".

 * Added "poly" function, simpler cousin of "gpoly".

 * Added "pdeg", "plead", "pcont", "pprim"; cleaned up "pdiv" and "pgcd".

 * Added `a p' command for polynomial interpolation.

 * Added `a I' command for numerical integration; made IntegLimit variable.

 * Added `a f' to factor polynomials; moved old `a f' to `a "'.

 * Added `a a' to do partial fraction decompositions.

 * Improved `a i' to integrate many more kinds of formulas.

 * Modified `a P' to find numerical roots of high-degree polynomials.

 * Modified `c 0' through `c 9' to convert int-valued floats to integers.

 * Made sinh, arctanh, etc., expandable into exps/logs by `a f'.

 * Added more algebraic simplifications having to do with logs and exps.

 * Changed `s s', `s t', `s x', `s l' to accept an equation at prompt.

 * Changed `s i' not to store Decls if its value is the default, [].

 * Changed `s i' to store in `d O' language mode if in Normal or Big mode.

 * Rearranged `V M'/`V R' matrix mapping modes.

 * Added <#1+#2> notation for lambda expressions.

 * Extended `b l' and other binary shifts to have a 2-argument version.

 * Changed `u c' and `u t' to give unitless result for unitless input.

 * Changed sqrt(1-cos(x)^2)-to-sin(x) to be an unsafe simplification.

 * Improved simplification of sqrts, e.g., sqrt(a^2 x + a^2 y).

 * Changed solver to treat (x-a)(x-b)(x-c) more intelligently.

 * Changed Pascal language mode to use "$FFFF" for hexadecimal numbers.

 * Added support for non-decimal display of floats.

 * Changed `p' to refresh stack display if current float format uses it.

 * Changed Big mode to use subscript notation for log10(x), log(x,b), r#nnn.

 * Changed Big mode to format deriv(u,x) and tderiv(u,x) as du/dx.

 * Changed Big mode to format integ(1/x,x) as "dx/x" instead of "1/x dx".

 * Added "tty" output type for graphics commands.

 * Documented Calc's random number generation algorithm in the manual.

 * Fixed a bug involving having "(setq calc-timing t)" in .emacs.

 * Fixed a bug that incorrectly parsed "|x| - 1" in TeX mode.

 * Fixed bugs and made improvements in `a R' when widening the guess.

 * Fixed a bug that where `a S' didn't solve (x - a)^2 = (x - b)^2.

 * Fixed a bug that sometimes crashed `a P' on systems of equations.

 * Fixed a bug that prevented `b p' (calc-pack-bits) from working.

 * Fixed some bugs in which certain functions didn't get autoloaded.

 * Fixed a bug in which the date <1/1/13> was incorrectly parsed.

 * Fixed a bug which prevented `j D' from expanding (a+b)/c.

 * Fixed a bug in solver: bad inverses for sinh and cosh.

 * Fixed a bug in math-possible-signs that failed for x*0.

 * Fixed a bug where sqrt(-a) was rewritten sqrt(a)*i even if a<0.

 * Fixed a bug in line breaker when first "word" of line was too long.

 * Worked around a makeinfo bug that handled @end group/@group badly.


Version 2.00 alpha 3:

 * Changed logic for locating component .elc files to be even smarter.

 * Changed "make install" to "make compile"; added different "make install".

 * Improved "make compile" to check file dates and compile only when needed.

 * Made output of "make compile" in batch mode more compact and readable.

 * Replaced "Quick Overview" in manual with "Demonstration of Calc".

 * Changed to use keymaps for dispatching M-# and h prefix keys.

 * Added TAGS target to the Calc Makefile.

 * Removed most doc strings from functions; new help commands are better.

 * Got rid of some crufty "fset" calls that were cluttering the code.

 * Split calc-grab-region into two functions, calc-grab-region/-rectangle.

 * Swapped positions of stack and trail in full-calc-keypad display.

 * Improved line-breaking algorithm for displaying long formulas.

 * Improved display of control characters in vectors shown as strings.

 * Changed `d o' to allow fraction format to specify desired denominator.

 * Changed `M-# y' to respect overwrite mode in target buffer.

 * Added `H' prefix to display-mode commands to suppress stack refresh.

 * Changed "calc-why" mechanism to display urgent messages automatically.

 * Handled taking derivatives of symbolic integrals and vice-versa.

 * Handled integrating vectors of formulas.

 * Incorporated Ewerlid's polynomial division and GCD functions into Calc.

 * Improved algebraic operations on "mod" forms, esp. polynomials.

 * Added some more financial functions (sln, syd, ddb).

 * Added nest, anest, fixp, and afixp (`H V R' and `H V U') functions.

 * Added `a .' (calc-remove-equal) command to take apart equations.

 * Generalized dfact to work for negative odd integers; added !! syntax.

 * Changed `k f' to factor 1, 0, and negative integers.

 * Changed `u M', etc., to accept +/- and [ .. ] forms as distributions.

 * Changed `g q' to remove *Gnuplot Commands/Trail* window if present.

 * Added support for Francois Pinard's "dumb terminal" driver for GNUPLOT.

 * Added ":: remember" feature for rewrite rules.

 * Changed rewrites to let pattern "a*b" match "x/2" with a=x, b=1/2.

 * Added ability to put function names like "simplify" in rewrite schedule.

 * Added "Rewrites Tutorial" to the manual.

 * Changed ` to bind RET as newline instead of finish if editing a vector.

 * Added some new exercises to the List Tutorial.

 * Changed `Z F', `V M', etc. not to remove stored vars from def arg list.

 * Added parsing for /1, 2, 3/ notation for Fortran mode vectors.

 * Added a "%%" syntax for comments in formulas being read.

 * Fixed a bug in which failing `h k' removed an existing Info window.

 * Fixed a bug in `j /' operating on subformulas like "a + b".

 * Fixed a bug in which "inf = inf" undesirably evaluated to 1.

 * Fixed a bug that simplified "0 = 1 + a + 2" to "0 = a".

 * Fixed a bug that failed for rewrite patterns like "fib(1 ||| 2)".

 * Fixed a bug that arose because rewrite programs are non-reentrant.


Version 2.00 alpha 2:

 * Changed LFD terminating algebraic entry to push in no-simplify mode.

 * Changed so that `K -' interprets `-' as calc-minus, not neg prefix arg.

 * Improved `h c' command to understand all Calc key sequences.

 * Fixed problems with DistribRules, NegateRules, and FitRules.

 * Fixed several bad node pointers in the manual.

 * Fixed a bug in `h C-w' when used with makeinfo-formatted manuals.

 * Fixed a bug in sqrt(-1) when Polar and HMS modes are enabled.

 * Fixed/improved dscalar and deven functions; added dodd.

 * Fixed a bug in polynomial handling that also affected sum(sin(k),k,1,n).

 * Fixed various other glitches in the manual.


Version 2.00 alpha 1:

 * Calc's tar file now creates a calc-(version) directory to unpack into.

 * Calc now comes with a Makefile; install with "make install".

 * Calc now comes already split into many files; installation is much simpler.

 * Changed base file name of the manual from "calc-info" to "calc.info".

 * Key binding for `M-# w' was documented but not implemented.

 * Bound M-# ' to be synonymous with `M-# f' (used to be `M-# q').

 * Changed M-# M-# to use last interface of C or K; E no longer counts.

 * Changed `i' (and `M-# i') not to return to Top node unnecessarily.

 * Changed `h' to be a prefix key with various help commands.

 * Changed `s' to be a prefix key with various store and recall commands.

 * Keys `i', `r', and `l' are obsolete (moved to `h' and `s' prefixes).

 * Rearranged `K', `X', and `M-RET' keys; `K' is now calc-keep-args.

 * Changed quick-calc to display input formula as well as output if room.

 * Changed quick-calc to interact with the editing buffer and kill ring.

 * Created pack, unpack, unpackt function equivalents of `v p', `v u'.

 * Changed to expand (a/b)^x to a^x/b^x only if b > 0 (not if a > 0).

 * Changed math-possible-signs to understand sqrt function.

 * Changed Z [, rewrites to consider any provably non-zero value as true.

 * Changed normal language modes to accept ** as a synonym for ^.

 * Added "maple" language mode.

 * Changed, e.g., Mathematica "(2 + 3 I)^(1.23*10^20)" to include parens.

 * Generalized math-compose-big properties for all language modes.

 * Introduced "string" and other function for composing expressions.

 * Changed many recursive vector routines to use loops instead.

 * Added evalv, evalvn function equivalents to `=', `N'.

 * Changed "expr =>" not to evaluate at all if in no-simplify mode.

 * Redesigned user interface of `a F' (calc-curve-fit) command.

 * Added "phase" feature to the rewrite rule system.

 * Added "&&&", "|||", "!!!" to the rewrite rule system.

 * Introduced a new notation for rewrites:  LHS := RHS :: COND.

 * Changed `a r' (but not `j r') to repeat 100 times by default.

 * Integrated EvalRules more cleanly into the default simplifications.

 * Added `H v l' [mdims] to measure the dimensions of a matrix.

 * Changed `u c' to interpret "/units" as "1/units".

 * Added `u a' to adjust unit prefix letters automatically.

 * Changed `u s' to enable scalar mode while simplifying.

 * Changed `c f' [pfloat] not to float integer powers or subscripts.

 * Added a three-argument form for the "hms" function.

 * Changed, e.g., sin(90) degrees to produce 1 instead of 1.0.

 * Changed symbolic mode to prefer sqrt(int): abs([1 2 3]) => sqrt(14).

 * Enhanced solver to handle, e.g., x + 1/x = a; exp(x) + exp(-x) = a.

 * Enhanced simplifier to handle, e.g., exp(a+2) / e^a => e^2.

 * Enhanced `a s' to simplify sqrt(x) - x^1:2 and exp(x) - e^x to 0.

 * Added -(a + b) to -a - b as a default simplification.

 * Added rules for differentiating sum() and prod() functions.

 * Added a few more energy units (due to Przemek Klosowski).

 * Added overflow/underflow checking for all floating-point arithmetic.

 * Extended error forms to work with complex numbers.

 * Generalized GCD to handle fractional arguments.

 * Changed graphics routines to evaluate "x" values, e.g., [-pi .. pi].

 * Added `g q', like `g K' but without viewing the Gnuplot Trail.

 * Changed `g p' and `V M' to display better "Working..." messages.

 * Modified `M-# g' to be more robust about grabbing formulas.

 * Added `Y' prefix key reserved for user-written extensions.

 * Added calc-load-hook and calc-ext-load-hook.

 * Prevented calc-install from leaving large ~ files behind.

 * Changed @bullet to @bullet{} in manual to conform to texinfo spec.

 * Rearranged some chapters in the manual to be a bit more logical.

 * Added calc-split-summary command.

 * Fixed several bugs in embedded mode.

 * Fixed a bug in calc-vector-covariance that required a prefix arg.

 * Fixed a bug that prevented parsing "a=>" with no right-hand side.

 * Fixed a bug which allowed incorrectly dividing a vector by a vector.

 * Fixed a bug formatting sum(...)^2 in Big mode.

 * Fixed a bug that prevented Calc from deleting old graphics temp files.

 * Fixed some typos calling calc-inverse-func instead of calc-invert-func.

 * Fixed bugs in the derivatives of conj, deg, and rad; added re, im.

 * Fixed a bug where (r;theta) parsed as r exp(theta i) even in Deg mode.

 * Fixed a bug which gave wrong answer for exp of a polar complex number.

 * Fixed a bug in `Z F' that failed if formula used non-arg variables.

 * Fixed a bad pointer to Info node "Assignments in Embedded Mode".

 * Fixed several errors in the Calc Summary.


Version 1.08 beta 1:

 * Calc's copyright has been assigned to FSF, for inclusion in Emacs 19!

 * Changed M-# to be a two-key sequence; use M-# M-# to start Calc now.

 * Rewrote and expanded the introductory chapter of the manual.

 * Added a key and function summary to the manual.

 * Changed the manual to take better advantage of TeX's math formatting.

 * Changed manual to be printable in @smallbook format.

 * Added "calc-embedded" mode.

 * Added "=>" [evalto] operator.

 * Added facilities for date and date/time arithmetic.

 * Added a set of financial functions (pv, fv, etc.).

 * Added infinite quantities inf, uinf, and nan (plus infinite intervals).

 * Added "EvalRules", "SimpRules", and "ExtSimpRules" variables.

 * Added sum and product commands `a +', `a -', `a *', `a T'.

 * Enhanced `a S' and `a P' to solve systems of equations.

 * Enhanced solver to handle eqns like sin(x) = cos(2 x), sqrt(x) + x = 1.

 * Added `a M' (calc-map-equation) command.

 * Added new statistical functions: mean, standard deviation, etc.

 * Added line, polynomial, and curve fitting commands (`a L' and `a F').

 * Added support for composite units, e.g., "mi+ft+in".

 * Enhanced "Big" mode to format square roots, choose, and powers better.

 * Enhanced "Big" mode to display fractions in large notation.

 * Added several alternate formats for matrix display.

 * Changed TeX mode to write "(1 + x^2)" instead of "\left(1 + x^2\right)".

 * Added support for relational operators in TeX and FORTRAN modes.

 * Added recognition of accents like \dot, \tilde, \underline in TeX mode.

 * Added "eqn" language mode.

 * Added extra control over display justification with `d <', `d =', `d >'.

 * Added calc-left-label and calc-right-label (`d {', `d }').

 * Added "nn%" syntax for algebraic formulas; equivalent to "nn * .01".

 * Added input syntaxes like a = b = c, a != b != c, a <= b < c.

 * Changed "_" to mean subscripts; old use of "_" in vars is now "#".

 * Introduced "matrix mode" and "scalar mode" (`m v').

 * Introduced generic identity matrices (idn(1)).

 * Added a method for declaring variables to be real, integer, > 0, etc.

 * Added `Z S' command for editing stored value of a variable.

 * Added "subst" algebraic function equivalent to the `a b' command.

 * Added `a f' command, changed deriv/integ/solve-for to use it.

 * Improved `a s' to simplify (x + y) (y + x) to (x + y)^2.

 * Improved `a s' to simplify i^2 to -1.

 * Improved `a s' to simplify, e.g., sin(pi/3) in Symbolic mode.

 * Improved `a s' to simplify sqrt(8) to 2 sqrt(2), 1/sqrt(2) to sqrt(2)/2.

 * Moved sin(arccos(x)) from `a e' to `a s'; not unsafe after all!

 * Changed (x y)^z => x^z y^z to be a usually-unsafe simplification.

 * Added thorough documentation of `a s' and `a e' to the manual.

 * Improved `a c' to collect "f(a)" even if "a" also appears elsewhere.

 * Introduced lin, linnt, islin, islinnt functions for linearity testing.

 * Improved `a x' to use binomial theorem to give simpler answers.

 * Improved `j D' to distribute powers of sums: (a + b)^n.

 * Improved `j M' to merge products of powers (may need no-simplify mode).

 * Changed to use defvar for DistribRules etc. so `Z V' works with them.

 * Improved `j *' and `j /' to work properly in a few more cases.

 * Improved `V R' to use identity value when reducing empty vectors.

 * Improved `v p' and `v u' to support more complex packing operations.

 * Disabled automatic simplification of sqrt(2)/2 to 1/sqrt(2).

 * Bound SPC and RET to press, TAB to next-menu in *Calc Keypad* buffer.

 * Added C-u ' to do algebraic entry with language mode forced to normal.

 * Added "$1", "$2", etc. input notation for algebraic entry.

 * Changed unary operators like `n', `&' to treat neg prefix args like RET.

 * Changed ` (calc-edit) to show full precision regardless of float format.

 * Enhanced quick-calc to display integers in several formats.

 * Documented `g H' (calc-graph-hide) command (had been left from manual).

 * Enhanced floor/ceil/trunc/round in several ways.

 * Added rounde and roundu functions.

 * Changed `c 1' through `c 9' to change small floats to 0.0; added `c 0'.

 * Enhanced set operations to work on sets of intervals.

 * Fixed erf(0), utpn(x,x,y), and arccosh(-1) to work properly.

 * Changed complex arctan and arctanh to follow Steele 2nd edition.

 * Expanded "Branch Cuts" section of the manual with some useful tables.

 * Rearranged order of words in mode line to be a bit more logical.

 * Changed `m N' (num-simplify) mode to evaluate constant vectors, too.

 * Changed `a r'/`j r' to prompt twice for separate LHS/RHS if necessary.

 * Enhanced `let(v,x)' in rewrites by allowing arbitrary patterns for v.

 * Changed cursor positioning in second prompt for `a b' (calc-substitute).

 * Changed `y' to omit line numbers more consistently.

 * Changed `o' (calc-realign) to reset horizontal scrolling to zero, also.

 * Added "pred" mode for calc-eval.

 * Added "calc-report-bug" as an alias for "report-calc-bug".

 * Added `Z T' and "calc-pass-errors" to aid debugging Calc-related code.

 * Added "calc-load-everything" (`m X' or `M-# L') command.

 * Enhanced calc-install to pre-build units table, CommuteRules, etc.

 * Changed Calc to interact more gracefully with load-path.

 * Changed Lisp Variable Index in manual to include user variables, too.

 * Fixed a bug that prevented calc-install from working under VMS.

 * Fixed a bug that sometimes crashed rewrites dealing with subtractions.

 * Fixed a bug that prevented `a S' from solving "3 - x = 1 + x"!

 * Fixed a bug in solver that crashed for certain cubics and quartics.

 * Fixed a bug in calc-simplify that crashed for equations and ineqs.

 * Fixed a bug which placed the "[" oddly in `d B' + `v /' mode.

 * Fixed a bug where finishing calc-edit improperly obeyed language mode.

 * Fixed a bug formatting (-1)^n in Big mode after selection commands.

 * Fixed a bug that got ">=" and "<=" backwards in rewrite conditions.

 * Fixed a bug that broke the `"x"' key in calc-keypad mode.

 * Fixed a bug in which `MAP$' in calc-keypad didn't display "Working...".

 * Fixed a bug where matrix division gave bad result for singular matrix.

 * Fixed a bug which closed Calc window if calc-grab-region got an error.

 * Fixed a bug where `a s' failed on formulas containing dimension errors.

 * Fixed a bug that caused `m F' to hang.

 * Fixed a bug in complex arithmetic that caused problems with solver.

 * Fixed a bug which raised intervals to interval powers incorrectly.

 * Fixed a bug in utpp/ltpp (order of arguments did not match the manual).

 * Fixed a bug in which `t y' rounded yanked data with old precision.

 * Fixed a bug in which "in(3, [3 .. 3))" returned true.

 * Fixed a bug which simplified abs(abs(x)) incorrectly.

 * Fixed a bug in which (a^2)^1:3 was unsafely simplified to a^2:3.

 * Fixed a bug in rewrite system which missed pattern "2 sin(x) cos(x)".

 * Fixed a bug in rewrite system which missed pattern "a - a cos(x)^2".

 * Fixed obsolete trail tags gsmp, gneg, ginv to jsmp, jneg, jinv.

 * Fixed some errors and made improvements in units table [Ulrich Mueller].


Version 1.07:

 * Added `m F' (calc-settings-file-name) command.

 * Added calc-autoload-directory variable.

 * Extended Z ` to accept a prefix argument.

 * Added keystrokes (v h, v k) for head, tail, cons.

 * Extended `v e' to accept a vector as the filler.

 * Changed `V M', `V R' to accept mapping-mode keys in uppercase, too.

 * Changed V M ' etc. to accept $, $$, ... as argument indicators.

 * Changed `t y' to accept a prefix argument.

 * Put in a cleaner and safer random number generator for `k r' et al.

 * Fixed a bug which completely broke `a r' command!

 * Fixed "0 * matrix" to generate a zero matrix instead of 0.

 * Fixed a bug in `a R' which sometimes caused it to crash.

 * Fixed a fatal typo in the TeX version of the manual.

 * Fixed a bug that prevented C-k, C-w, M-w from working in Trail buffer.

 * Fixed another bug in `Z P' command.

 * Fixed a bug in `u s' which incorrectly simplified subtractions.

 * Fixed an argument-name aliasing bug evaluating lambda( ) formulas.

 * Fixed overfull hboxes in the manual.

 * Fixed various other bugs in the manual.


Version 1.06:

 * Added "calc-keypad" mode for X window system users (try it!).

 * Improved "calc-eval" for calling/operating Calc from user-written Lisp.

 * Moved vector accumulate command to `V U' (old `H V R' still supported).

 * Added right-to-left reductions: `I V R' and `I V U'.

 * Added set operations on vectors: intersect, union, diff, xor.

 * Added `I v s' to remove a subvector from a vector.

 * Introduced `H |' to append two vectors with no magical special cases.

 * Introduced rhead, rtail, and rcons for isolating last vector element.

 * Changed `g p' to keep temp files around until data actually change.

 * Improved `a S' to solve many higher-order polynomial equations.

 * Added `a P' to produce a vector of all solutions to an equation.

 * Enhanced `a v' and `j v' to allow top-level-only evaluation.

 * Changed `j DEL' to delete a side of an eqn or ineq, leaving other side.

 * Fixed binding for keys `j 1' through `j 9'.

 * Introduced "let" marker in rewrite rules.

 * Enhanced the "sign" function to provide a two-argument version.

 * Changed "max-specpdl-size exceeded" error message to be user-friendly.

 * Put "<Aborted>" in the trail in above case and when user presses C-g.

 * Changed TeX mode to generate \ldots instead of \dots, recognize both.

 * Changed "sin(0)" etc. (for integer 0) to generate "0" instead of "0.".

 * Enhanced Programming Tutorial exercise 2.

 * Fixed an error in the answer to Types Tutorial exercise 3.

 * Fixed several bugs relating to head, tail, and cons functions.

 * Fixed some other minor typos in the manual.

 * Fixed several bugs in `Z P' (calc-user-define-permanent).

 * Fixed several bugs that broke the `g P' command.


Version 1.05:

 * Created a calc-install command to ease installation.

 * Added lots of exercises to the Tutorial section of the manual.

 * Added ability to select and operate on sub-formulas.

 * Substantially improved the algebraic rewrite-rule system.

 * Added a set of graphing commands that use GNUPLOT.

 * Added a command (`a R') for finding numerical roots to equations.

 * Added several new math functions, such as erf and Bessel functions.

 * Added key bindings for miscellaneous commands using the "f" prefix key.

 * Added lots of new vector operations, many of them in the spirit of APL.

 * Added more control over vector display, including an abbreviated mode.

 * Improved keyboard macro editing; added read-kbd-macro to macedit.el.

 * Introduced the `m S' (calc-shift-prefix) command.

 * Enhanced the calc-edit command in several ways.

 * Made it possible to hit ` (calc-edit) during numeric/algebraic entry.

 * Enhanced the calc-solve-for command to handle inequalities.

 * Enhanced calc-simplify to handle equations and inequalities.

 * Taught log10 and log to look for exact integer or rational results.

 * Added ability to take Nth roots directly.

 * Added "increment" and "decrement" commands for integers and floats.

 * Added "full-help" command, changed "h" key to invoke it.

 * Added special help for Inverse and Hyperbolic prefixes.

 * Added an optional prefix argument to `o' (calc-realign).

 * Changed `t s' and `t r' to use RET as the search exit key.

 * Made handling of operator keys for V M, V R, etc. more regular.

 * Improved TeX mode; added support for \matrix format.

 * Added a variant of `m a' mode that only affects ( and [ keys.

 * Fixed "Mismatch" message for algebraic entry of semi-open intervals.

 * Trimmed fat from calc.el to speed loading, moved more to calc-ext.el.

 * Fixed a bug in which minibuffer entry rounded to out-of-date precision.

 * Fixed a bug which crashed Calc 1.04 under Epoch.

 * Fixed a bug which messed up Calc Trail's mode line, among other things.

 * Fixed a bug which caused trail ">" to show only when in Trail buffer.

 * Fixed a bug in which "calc" called "calc-grab-region" with too few args.

 * Fixed bugs in both implementation and documentation of calc-perm.

 * Fixed a bug in which calc-simplify-extended always used radians.

 * Fixed a bug where calc-comma failed to override "polar" mode.

 * Fixed a bug doing mixed arithmetic on rectangular+polar complex numbers.

 * Fixed several bugs in transcendental functions with complex arguments.

 * Fixed a bug in which `a s' simplified "x / .5" to ".5 x".

 * Fixed numerous other bugs in various parts of Calc.

 * Completed the "Hooks" section of the "Internals" chapter of the manual.


Version 1.04:

 * Included a copy of revision history (from README) in calc.el.

 * Added the "calc-split" feature to split calc-ext.el into smaller bits.

 * Changed calc-unpack to unpack floats and fractions, too.

 * Added "mant", "xpon", and "scf" functions for decomposing floats.

 * Fixed a bug in the "y" command with positive prefix arguments.

 * Rearranged binary shift/rotate command keys to be a bit more convenient.

 * Fixed a bug in which simplifying "(0/0) * 2" crashed with a Lisp error.

 * Made `H F' [ffloor] and friends faster for very large arguments.

 * Made calc-define-del more robust.

 * Handled pasting of data into the Calculator using the mouse under X.

 * Made overlay-arrow variables buffer-local to avoid interference.

 * Fixed a problem in which Calc Trail buffer got stuck after a C-x C-w.


Version 1.03:

 * Changed math-choose to compute n-choose-m faster when m is large.

 * Fixed some problems with TeX mode.

 * Fixed a bug that prevented `b s' from working without a prefix argument.

 * Added "calc-eval" function.

 * Improved calc-grab-region.


Version 1.02:

 * Fixed a bug in Tutorial: telephone pole height/distance were switched!

 * Fixed a few other things in the manual.

 * Added "full-calc" command.

 * Added "calc-insert-variables" (`Z I') command.

 * Quick Calc now works even if you are already in the minibuffer.

 * Fixed a bug in math-mul-bignum-digit which affected math-and, etc.

 * Definition of "Hectares" was wrong in units table.

 * Fixed a bug in calc-execute-kbd-macro concerning undo and refresh.

 * Bound "calc-undo" to `C-x u' as well as `C-_' and `U'.

Version 1.01:

 * Added a tutorial section to the manual.

 * Next and Prev for node Strings in the manual were reversed; fixed.

 * Changed "'bignum" in calc-isqrt-bignum-iter to "'bigpos".

 * Fixed a bug that prevented "$" from working during algebraic entry.

 * Fixed a bug caused by an X (last-X) command following a K (macro) cmd.

 * Fixed a bug in which K command incorrectly formatted stack in Big mode.

 * Added space between unary operators and non-flat compositions.
   (Otherwise, "-(a/b)" in Big mode blended the minus sign into the rule!)

 * Fixed formatting of (-1)^n in Big mode.

 * Fixed some problems relating to "not" operator in Pascal language mode.

 * Fixed several bugs relating to V M ' and V M $ sequences.

 * Fixed matrix-vector multiplication to produce a vector.

 * Introduced Z ` ... Z ' commands; renamed old Z ' to Z #.

 * Fixed various other bugs.

 * Added calc-settings-file variable suggested by C. Witty.


Version 1.00:

 * First official release of Calc.

 * If you used the Beta test version (0.01), you will find that this
   version of Calc is over 50% larger than the original release.
   General areas of improvement include much better algebra features;
   operations on units; language modes; simplification modes; interval
   arithmetic; vector mapping and reduction.  Other new commands include
   calc-fraction and calc-grab-region.  The program has been split into
   two parts for faster loading, and the manual is more complete.