Shareware Grab Bag

home *** CD-ROM | disk | FTP | other *** search

/ Shareware Grab Bag / Shareware Grab Bag.iso / 011 / rpn141.arc / RPN.DOC next >

Wrap

Text File | 1985-09-11 | 40.0 KB | 977 lines

RPN CALCULATOR PROGRAM - VERSION 1.41 Copyright (c) 1984,85 Tom Ginn NAME: RPN COMMAND STRING: RPN [RAMFSBn=x] SUMMARY: RPN is a calculator program written in C (mostly) and compiled with the DeSmet C development package. It runs on the IBM PC under PC-DOS and hopefully on most IBM PC compatible machines using PC-DOS or MS-DOS. Either color or monochrome display may be used. RPN uses up about 54K of memory so you will need total system memory of at least 96K to run it, and more if you plan to install it in the background. The features of the program include: 1. Reverse Polish (RPN) or algebraic operation 2. Operation in number bases from 2 to 16 3. Easy number base conversion 4. Fixed point or Scientific (base 10 only) notation 5. Financial calculations 6. Summation, mean & standard deviation 7. User defined macro keypoke sequences 8. 8087 math chip support 9. Trig, Log, Exponential operations 10. Calendar display 11. Ability to remain memory resident in the background 12. Ability to use BIOS routines or go directly to screen memory for faster output. Please note: the memory resident and direct screen output features depend on IBM-PC screen memory and video mode compatibility. While vanilla RPN has worked on every PC compatible I've tried it on including PC jr, and I have not yet had any problems running the memory resident and direct screen features on the machines I've been able to try, there is still room for these features to fail on some machines. PROGRAM DISTRIBUTION: RPN is distributed on a single sided 8 sector MS-DOS/PC-DOS diskette containing the following files: 1) RPN.EXE - The calculator program with software floating point for systems without the 8087 math co-processor. 2) RPN7.EXE - The calculator program with 8087 floating point support for systems with the 8087 math co-processor. 3) RPN.DOC - This documentation file. 4) RPN.R1, RPN.A1 and RPN.R2 - Sample macro files in the format saved by the macro save operation. We are all familiar with the current marketing strategy for software products: use copy protection, accept the fact that there will be unauthorized copies made, and charge an arm and a leg to try to compensate for it. Well, with RPN there is no copy protection, all copies are authorized, and, by the way, you can keep your leg and most of your arm. You are welcome to make copies of RPN for your personal use and encouraged to pass it along to your friends. If you find RPN useful and decide to keep a copy, I will hope to receive a small honorarium from you. I suggest $10.00 as an appropriate amount but leave the final decision to you. Procrastination is my worst enemy in this endeavor, so please don't put it off. Send to: Tom Ginn P.O. Box 24842 San Jose, CA 95154-4842 If you prefer, you may order a program diskette of the current revision level from me for $15.00. Orders from California residents should include sales tax. This method of software distribution can succeed only with your co-operation. If you would like to see greater use of this medium to distribute useful software at minimal cost, please support this program and others like it. Please Note.... ** When you copy RPN be sure to pass along all the files. ** Do not circulate modified copies of RPN. ** Do not remove this message from the RPN files. While I have made every effort to locate and remove insect life from this program, there is an unwritten law that says any non-trivial software package will have bugs, and I don't expect this program to be exempt from the laws of nature. Therefore I must distribute RPN "as is" to be employed at the user's risk, and can assume no liability for any losses or damages of any kind resulting from use or misuse of the program. No express or implied warranties regarding merchantability or fitness of use shall apply. Comments, suggestions, and/or reports of encounters with any remaining insect life may be directed to the author at the above address. If your correspondence requires a reply, please enclose a self addressed stamped envelope to help speed things up. Thank you for your interest in my calculator program. I hope you will find it as useful as I do. DESCRIPTION: RPN is executed by typing its name and optional parameters specifying the desired mode of operation in response to the prompt from the MS-DOS command processor: RPN [RAMFSBn=x] (RPN7 for the 8087 load) Where: R - RPN mode. The program will run in RPN mode and will attempt to load a macro file RPN.R0 (this is the default mode). A - Algebraic mode. The program will run in algebraic mode and will attempt to load a macro file RPN.A0. M - Override color attributes with white on black. F - Fast mode. This parameter causes the screen output to bypass the BIOS and go direct to screen memory. You'll be amazed at the difference. S - Stay resident in memory. RPN will install itself in memory and be ready to run when it sees a special activation code from the keyboard. When running resident, fast mode is implied. B - Blink scroll handling. You may also be amazed by the amount of noise fast mode makes on your screen if you have the IBM color card or a similar one that doesn't like to have the cpu access screen memory during display. If so, use this option when you are using F or S to tell RPN to synchronize output to the screen with the horizontal retrace so it doesn't show. It slows down fast mode a bit but it's worth it. =x - The default activation code is CTRL-ALT-C but this parameter will change it to the letter you specify immediately following the = sign. n - Macro file number from 0-9. Allows you to specify the macro file number to be loaded. See section on saving macros for more on macro files. Examples: RPN vanilla RPN RPN7 8087 load of RPN RPN A algebraic mode RPN AF algebraic and fast modes RPN S memory resident RPN SB memory resident with blink scroll handling RPN S=Z memory resident with CTRL-ALT-Z activation RPN F5 fast mode, load macro file RPN.R5 Please note: If the 8087 load of RPN is executed on a system without the 8087 co-processor present it will hang the computer and you will have to re-boot (CTRL-ALT-DEL will do it). If the non-8087 load is executed on a machine with an 8087 installed there are no ill effects but you will not take advantage of the 8087 speed. If no parameters are specified the program will determine the operating mode by the macro files that are present. It looks first for RPN.R0 and if it is present then RPN mode is used and RPN.R0 is loaded. If it doesn't find RPN.R0 then it looks for RPN.A0 and if it is found then algebraic mode is used and RPN.A0 is loaded. If neither macro file is found the program will default to RPN mode with predefined color attributes. You may specify a macro file number to determine the macro file that is used instead of the default 0. The program execution is terminated by poking the Escape key (ESC). If RPN is running as a normal program it will return control to the operating system. If RPN is running as a background program entered by the CTRL-ALT-whatever activation code, it will return to the interrupted program. In the default RPN (Reverse Polish Notation) mode, the program behaves in many respects like a Hewlett-Packard RPN calculator with a 4 register stack ( x, y, z, & t registers ), 10 memory registers, and a last x register. All 4 registers of the stack are displayed on the screen and the calculations operate on the x register only or on the x and y registers, leaving the result in the x register or x and y registers. The following keys are used to execute the calculator's operations. A short description will be given here for each key used, and more detailed information on some operations will appear later in this manual. Poking a key for which no operation is defined just rings a bell at you. 0-9 Number entry. A-F Hex number entry. . Decimal point entry. + y+x Since some keyboards don't have a plus you can get to without the shift key, TAB is equivalent to +. - y-x / y/x * y*x Since some keyboards don't have an * you can get to without the shift key, COMMA is equivalent to *. % y%x Remainder of y/x or y mod x. This operation is limited to integer operands less than 2,147,483,648 (2^31). ^ Pow(y,x) y raised to the x power. & y AND x Logical bitwise AND of y and x register values. This operation is limited to integer operands less than 2,147,483,648 (2^31). | y OR x Logical bitwise Inclusive OR of y and x register values. This operation is limited to integer operands less than 2,147,483,648 (2^31). $ y XOR x Logical bitwise Exclusive OR of y and x register values. This operation is limited to integer operands less than 2,147,483,648 (2^31). ~ Negate x ! x factorial \ Alternate operation prefix. When this key is poked, "shift" is written on the message line to signify alternate interpretation of the next key poked. # Macro function flag. This character, inserted at the beginning of a macro, indicates that you want the macro treated as a function which returns a value to the x register. [] Macro comment brackets. Characters between brackets in a macro string will be ignored. This allows short comments to be included in saved macro strings. = Enter key (RPN mode) or Equals key (Algebraic mode). Note several alternates are interpreted as ENTER for keypoking ease. CR Enter key (RPN mode) or Equals key (Algebraic mode). The CR (carriage return) key generates the same code as Control-M and on various keyboards may be labeled "RETURN", "CR", or "ENTER". LF Enter key (RPN mode) or Equals key (Algebraic mode). The LF (line feed) key generates the same code as Control-J and on various keyboards may be labeled "LINE FEED" or "NEW LINE" or may not be present at all. space Enter key (RPN mode) or Equals key (Algebraic mode). F1 Function keys F1-F10 execute assigned macros. \F1 Function keys F1-F10 copy the macro function key definition to the scratch string for macro editing. If the macro string is defined it is displayed on the message line. a Screen setup for normal calculator mode. Can also be used to rewrite the calculator screen in case the display gets trashed. Displays the status line, register stack, message line, and calculator help text. \a Screen setup for financial mode. Enables the financial operation keys. Displays status line, register stack, message line, financial registers, and financial help text. The normal calculator functions are still available, too. b Set number base. The calculator is capable of doing most of its numeric input and output in any number base from 2 to 16. \b Memory display mode. An alternate display screen in normal calculator mode shows the 10 memory registers in place of the help text in the lower half of the screen. c Cos(x) Cosine of x. \c Arccos(x) - Angle whose cosine is x. d Set display format. The display format may be set to either fixed point or scientific notation and the number of places right of the decimal point may be varied from 0 to 9. \d Set degrees mode for trig functions and rectangular/polar conversions. \D Debug toggle. This turns on (or off) some debugging messages that were handy enough to keep available. One is output of the x and y registers so their internal 64 bit storage format can be seen as 4 hex words. Another is intermediate values of the iterative process used to calculate interest rate in the financial mode. e Enter exponent. For entry of numbers in scientific notation when number base is decimal. '-' or '~' changes sign of exponent. Any key not '0'-'9', '-' or '~' terminates entry and is interpreted as the next operation key. \e Screen setup for summation mode. Displays status line, register stack, message line, summation registers, and summation help text. Provides for summation of a number of values entered into the x register and computation and display of the mean and standard deviation. f Frac(x) Keep fractional part of x. The integer part of the floating point number in the x register is discarded. \f Calendar display mode. This draws the calendar screen and switches over to the calendar commands, which are described later. In calendar mode the normal calculator commands are disabled. i Int(x) Keep integer part of x. The fractional part of the floating point number in the x register is discarded. k Rectangular to polar. The x and y registers are assumed to contain rectangular x and y coordinates, respectively, and are converted to polar coordinates (r, theta). The resulting r is left in the x register and theta is left in the y register. If you use this while in algebraic mode, you will have to use the "x" key (exchange) to load the y register. \k Polar to rectangular. This is the inverse of the rectangular to polar conversion. Again, in algebraic mode the "x" key is used to load the y register. l Log10(x) Log of x to base 10. \l Pow(10,x) Raise 10 to the x power (antilog of x). L Last x. Whenever an operation is executed which calculates a result and leaves it in the x register, the old value contained in the x register is saved in the "last x" register. The last x key retrieves the saved value and pushes it on the stack. m Execute macro. \m Define macro. M Edit macro. \M Save macros in a disk file. n Ln(x) Natural log of x. \n Exp(x) Raise e to the x power (exponential of x). o 1/x Reciprocal of x. p Pi Push Pi on the stack. q Sqrt(x) Square root of x. \q x * x Square of x. r Recall from memory register and push on the stack. This operation is also used to move financial and summation register values to the operational stack so they can be used in normal calculations. R Recall from memory register. Same as "r" for convenience. \r Set radians mode for trig functions and rectangular/polar conversions. s Sin(x) Sine of x. \s Arcsin(x) Angle whose sine is x. t Tan(x) Tangent of x. \t Arctan(x) Angle whose tangent is x. v Rotate down. V Rotate down. Same as "v" for convenience. \v Rotate up. \V Rotate up. Same as "\v" for convenience. w Write x in memory register. Optionally add x to memory, subtract x from memory, multiply x by memory, or divide x into memory, storing the result into memory. W Write x in memory register. Same as "w" for convenience. x Exch(y,x) Exchange x and y register values. X Exch(y,x) Same as "x" for convenience. \x Clear the summation registers. \X Clear the summation registers. Same as "\x" for convenience. y Clear financial registers. Y Clear financial registers. Same as "y" for convenience. \y Clear memory registers. \Y Clear memory registers. Same as "\y" for convenience. z Clear x register. Z Clear x register. Same as "z" for convenience. \z Clear stack. \Z Clear stack. Same as "\z" for convenience. The following operations make up the finiancial section of the calculator. The Enter/Compute and \p keys are disabled in normal calculator mode and are enabled when you change to the financial mode. \a Redraw financial screen. I Enter/compute Interest Rate. N Enter/compute Number of Periods. P Enter/compute Payment. S Enter/compute Start (Present) value (PV). T Enter/compute Terminal (Future) value (FV). \p Set the financial period to use. This will be day, week, month, quarter, or year. If it makes sense to do so, the number of periods will be converted to the new period and the new periodic payment will be calculated. r Recall. "r" followed by the register letter is used to move a value from a financial register to the operational stack so it can be used in normal calculations. y Clear financial registers. The following operations make up the summation section of the calculator. \e Redraw screen. u Summation (+). The value from the x register is summed in the summation registers and the mean and standard deviation values are computed and displayed. \u Summation Correction (-). If an erroneous value is summed with the "u" operation it can be removed from the summation by entering the value to be removed in the x register and using "\u". \x Clear the summation registers. r Recall. "r" followed by the register letter is used to move a value from a summation register to the operational stack so it can be used in normal calculations. The following commands are available in calendar display mode. When the calendar display mode is active, the rest of the calculator functions are disabled. 0-9 Enter a year. Each time a digit is typed the existing year is multiplied by 10, the new digit is entered into the ones position, and a calendar is generated. The calendar has a range from year 0001 to 10000. F1/F2 Go back/forward 1 month. For convenience the left arrow and right arrow cursor motion keys are also plugged into these functions. F3/F4 Go back/forward 1 year. For convenience the down arrow and up arrow cursor motion keys are also plugged into these functions. F5/F6 Go back/forward 1 decade (10 years). F7/F8 Go back/forward 1 century (100 years). Home Displays the calendar for today's date. Today is determined by reading the system time so you should set your clock for this to come out right. The "t" key is also plugged into this action. End Leave calendar mode. The program will re-enter the calculator mode that was active when calendar mode was started up. The "x" key is also plugged into this action. RPN vs. ALGEBRAIC You will see the difference between RPN mode and algebraic mode in operations which act on two operands(+ - * / % ^ & | $). Reverse Polish Notation requires that you first put the numbers to be operated on into the stack and then poke the appropriate key to inform the program what you want it to do. The "Enter" key is used to separate the entry of the two numbers. To multiply 2 times 3 the keypoke sequence is: 2 Enter 3 * As you poke the keys, you will see the stack be updated on the screen with each keypoke. These are the stack displays you will see during the example calculation: 2 Enter 3 * t: 0.00 0.00 0.00 0.00 0.00 z: 0.00 0.00 0.00 0.00 0.00 y: 0.00 0.00 2.00 2.00 0.00 x: 0.00 2.00 2.00 3.00 6.00 A slightly more complicated example (2 * 3 + 4 * 5) adds another product to the result of the first product and shows the use of RPN and the register stack to preserve an intermediate result for later use. We start off with the result of the first product remaining in the x register, use the same procedure to multiply 4 times 5, and then poke the "+" key to add the two product results: 4 Enter 5 * + t: 0.00 0.00 0.00 0.00 0.00 0.00 z: 0.00 0.00 6.00 6.00 0.00 0.00 y: 0.00 6.00 4.00 4.00 6.00 0.00 x: 6.00 4.00 4.00 5.00 20.00 26.00 If you are uncomfortable with the RPN mode, the calculator may be run in its algebraic mode. In algebraic mode only the x and y registers are displayed and operation is modified so the keypoke sequence to multiply 2 times 3 becomes the familiar: 2 * 3 = In algebraic mode the ability to preserve an intermediate result on the stack is gone but the memory registers can be easily used instead. To use the rectangular/polar conversion in algebraic mode it is necessary to use the exchange function ("x") to get the y register loaded with its parameter. MEMORY RESIDENT MODE and FAST MODE RPN can be run either as a normal program or as a resident background utility. You install it in the background by executing it with the "s" parameter on the command line. Instead of the normal calculator screen you will see something like this: A> RPN S RPN/Algebraic Calculator - Version 1.41 Copyright (c) 1984,85 Tom Ginn Installed in background. Activate with CTRL-ALT-C when the running program is waiting for key input. A> From this point on you may get into RPN by typing the activation code whenever (almost) the running program is expecting keyboard input. I say "almost" because the command.com command line input can only be interrupted if it is a fresh prompt and no other characters have been typed before the activation code. RPN will silently refuse to activate itself if command.com is active and you have typed any characters in response to the command prompt. In this case all you need to do is type CTRL-C to clear the entered characters and get a fresh prompt. You may interrupt any standard IBM-PC video mode and return to it when you exit RPN. Some cautions should be observed, however: 1. If you have a video card with non-standard video modes, DON'T interrupt a non-standard mode. RPN doesn't know how to restore anything but standard modes. 2. If you have a mouse running and a mouse cursor on the screen you interrupt, don't move the mouse while RPN is active. The mouse interrupt handler is still running and trying to update screen memory and you may see it on the RPN screen. Your interrupted screen may not be quite as you left it either. 3. If you have two or more background routines installed, it's best not to interrupt one to enter another. RPN and some other background programs change video mode by programming the video card directly rather than call the BIOS to change the mode. I found this to be necessary to keep a mouse cursor from getting lost. The video mode may not be properly restored when returning to another background program. 4. If you use RPN in the background along with Borland's Sidekick, install RPN before Sidekick. You won't have any trouble getting to Sidekick with the CTRL-ALT combination. To get RPN, hold CTRL-ALT down until Sidekick's menu window goes away before pressing the "C" to get RPN. If you install RPN in the background of a PC jr, you will find that it only allows the activation sequence to proceed if the computer is already in the video mode required (mode 2 or 3). You can't interrupt any other mode because RPN does not handle the mode changes on the jr. I think most jr's will not have enough memory available to install background utilities anyway. If you find it necessary or desirable to have RPN use a different activation code than the default CTRL-ALT-C, you can use the =x parameter to specify any letter of the alphabet to replace the default C. For example, to install RPN with an activation code of CTRL-ALT-R, install RPN in memory with the command: RPN S=R The "fast" mode is required by and automatically set by the resident mode. However, "fast" mode is available by itself when the program is executed normally by using the "F" parameter on the command line. Try it both ways. The difference is interesting. Please note: the memory resident and direct screen output features depend on IBM-PC screen memory and video mode compatibility. While vanilla RPN has worked on every PC compatible I've tried it on including PC jr, and I have not yet had any problems running the memory resident and direct screen features on the few machines I've been able to try, there is still room for these features to fail on some machines. NUMBER BASES As previously mentioned, the calculator will operate in any number base from 2 to 16. The number base is set by the "b" key, which will ask for an input base and a conversion base. The input base determines the range of keys that will be accepted from the keyboard, the conversion of input data to the computer's internal format, and the number base of the primary register displays on the screen. The conversion base, if different from the input base, determines the base for a second display of the x and y registers which is put on the screen to the right of the primary displays. The program comes up with both bases set to 10 for use as a decimal calculator. The status line near the top of the screen will say "Base 10". To do number base conversions from base 10 to base 16 (for example), use the "b" key to set the input base to 10 and the conversion base to 16. The status line will say "Base 10:16" and there will be two displays of the x and y registers. The left display is in base 10 and the right display is in base 16. FINANCIAL MODE The calculator's financial mode implements the standard formulae for computing one of five related values from the known values of the other four. The financial mode screen displays the registers reserved for these values on the lower half of your screen. They are: N: The number of periods. I: The annual interest rate in percent (APR). S: The starting or present value. T: The terminal or future value. P: The periodic payment. Using this mode you can figure out mortgage payments, accumulation of savings, annuities, and several other kinds of financial situations that fit the pattern. When you enter financial mode the first time the default period will be months. You can change the period to days, weeks, months, quarters, or years with the \p function. This allows the program to internally compute the periodic interest rate from the annual interest rate you enter into the interest register. Let's start off with the good ol' home mortgage payment calculation as an example. First you would get into financial mode by keying in "\a" and making sure the period is set to "months". You then enter the known values into the financial registers by keying the value in so it appears in the x register on the stack and then poke the appropriate Enter/Compute operation key from the financial help menu at the bottom of the screen. The program watches numeric input and decides to do Enter if you have just keyed in a number and Compute if no number has been keyed in since the last Enter. For a 30 year mortgage, key in "360N". For a 13.5% annual interest rate, Key in "13.5I". For a $90,000 loan amount, key in "90000S". For a fully amortized loan, the future value is 0 so key in "0T" if the register does not already contain 0. To compute the monthly payment, key in "P". The calculator will compute and display "-1030.87". Note the value of the payment is negative. Since this calculation was done from the point of view of the borrower the loan amount is positive, representing an amount being paid in at the beginning of the loan. The negative payment signifies an amount being periodically paid out from the point of view of the calculation. If we looked at the loan from the lender's point of view the signs would be reversed. Now that these values have been entered, try entering different interest rates and recomputing the payment to see the effect. Remember that any one of these variables can be computed from the other four. In the above example, suppose you could afford a $900 monthly payment and use the program to determine what interest rate you would have to shop for. For a $900 payment, key in "900~P". The "~" changes the sign of the payment to negative, which is necessary for our point of view. To compute the interest rate, key in "I". The program will go away and work for a while on this one but it will come back in a few seconds with the answer, "11.63". For the next example, we'll do the IRA (Individual Retirement Account) routine to see how savings accumulation works. Since we are going to enter a complete new set of numbers, first use "y" to clear the financial registers. Now, lets assume you want to put $2000 per year into your IRA for 30 years and you will do it by depositing $500 each quarter (3 months). Key in "\p" to change the period and answer the "What period?" question with a "q" for "quarter". Now enter the known values: For 120 quarters (30 years), key in "120N". For 12% interest rate (APR), key in "12I". For $500 payment to the account, key in "500~P". Remember our point of view. This is an outgoing payment. The present value is 0, so the S register is OK. To compute the future value, key in "T". WOW!! Don't you wish you had started this 30 years ago? You might have had $561,849.79 by now. There are many financial situations that fit this pattern. Remember that the number of periods and interest rate are always positive. Negative values have no sensible interpretation. Present value, future value, and periodic payment are either positive or negative depending on whether cash is flowing to you or away from you. The correct sign for the known values must be entered for the calculation to work correctly. The sign of a computed answer tells the direction of cash flow. MACROS The program's macro capability allows you to define sequences of keypokes which will be remembered by the program and executed on demand. There are 11 macro storage strings defined. Ten of these are used to hold macros assigned to the 10 function keys. The 11th is a scratch area which is used for initial definitions, macro execution, and macro editing. There are two ways within the program to define a macro string. The first is the "\m" operation which starts collecting and remembering keypokes as you actually execute them on the calculator. This collection process will continue until you fill up the macro storage string (76 characters) or until you enter the "\m" keys a second time to signal the end of the macro entry. The status line near the top of the screen will tell you when macro definition is in progress. In this macro entry mode there can be no backing up to correct mistakes. The second way to define a macro string is with the macro editor, which is capable of examining and modifying the scratch string and assigning its contents to one of the 10 function keys. The macro editor is entered by the "M" key and exited either by a second use of the "M" key or by use of one of the function keys. Use of "M" to exit the editor leaves the edited macro in the scratch string where it can be executed with "m". Use of a function key to exit the editor causes the edited macro to be copied from the scratch storage to the string assigned to the function key. When entered, the macro editor will display the macro currently resident in the scratch area. This may be the macro just entered with "\m", or the last macro executed by a function key, or it may be empty. You may edit the macro with the following edit keys: cursor left ---- non destructive left cursor motion cursor right --- non destructive right cursor motion home ----------- moves cursor to beginning of line end ------------ moves cursor to end of line delete --------- deletes character the cursor is currently on backspace ------ deletes character before the cursor position Printing characters in the range from "!" to "~" are inserted into the line before the current cursor position. Characters outside this range, except for the edit keys, are ignored. Aside from excluding non-printing characters and limiting macro size the editor makes no tests on the validity of the characters you enter so you are responsible for the logic and executability of the changes you make. To edit a macro assigned to a function key, it is necessary to move it to the scratch string. This happens when the macro is executed by poking its function key. To move it without execution, poke the "\" key followed by the function key. Then use "M" to enter the editor. Brief comments may be inserted in macro strings by enclosing them in square brackets. This may be useful if you are going to save macros in a disk file but the comment uses up macro keypokes. Here are some sample macros (in RPN) to do the Fahrenheit to Centigrade and Centigrade to Fahrenheit conversions: [C to F]=9=5/*32+ [F to C]=32-5=9/* Notice in the RPN macros above the uses of the ENTER key are stored as equals signs. Inside the calculator all the keys (space, newline, return, and =) which are interpreted as ENTER are converted to the equals sign, partly because it is the only printing character of the group and partly because it makes sense in algebraic mode where the same set of keys is used as EQUALS. The same calculations as they could be entered as a macro in algebraic mode would look like this: [C to F]*9/5+32= [F to C]-32*5/9= Normally, execution of a macro proceeds exactly as if the characters from the macro string were being entered from the keyboard. The values contained in the stack on exit from the macro will be those left from the calculations in the macro and will likely be different from the values on entry. In many cases this makes no difference and in fact may be desirable. Perhaps just as often, however, it is desirable to write a macro to do some sort of a conversion or other function and return a value to the x register much as the built-in functions such as log and square root do. In such a function, you would prefer not to have the stack altered by the macro (except for the x register, of course). The macro function flag (#) character, inserted at the beginning of a macro, indicates that the macro is to be treated as a function returning a result to the x register. This causes the y, z, and t registers to be saved while the macro is executing and restored when the macro ends. Whatever the macro leaves in the x register is assumed to be the result and is left there. Also the flags that control processing of input are set to indicate that a new number input is expected (in other words, any number entry in progress when the macro terminates is ended as if the ENTER key had been poked). In algebraic mode, a pending operation is saved and restored at the end of the macro execution, allowing the result of a macro function to be included in a chain calculation. For example, if you put the radius of a circle in the x register, a macro to find the area of the circle wuold be "\q*p=" in algebraic mode. If you enter this and execute it, you'll see that it changes the y register. A little further experimentation will show that the result can't be used in a chain calculation except at the beginning. Now change the macro so it is "#\q*p=" and try again. Now executing the macro saves y when the "#" is encountered and restores it when the macro finishes. It also saves any pending operation and propogates it across the macro so you can do 1+2.5m+4m= and get the right answer. It is algebraic mode where this is really necessary. In RPN mode you can get along without this because the problem of remembering pending operations does not exist, but it still can be handy to preserve the contents of the y, z, and t registers through macro execution. By the way, I realize that the next extension of this is to allow a macro to use another macro and that may happen in a later revision but is not handled now. SAVING MACROS IN A DISK FILE The macros assigned to function keys may be saved in a disk file and restored whenever the program is executed. Up to 10 sets of macros may be saved in RPN mode and 10 more in algebraic mode (RPN and algebraic macros are not interchangable, hence separation of macro files). The file names are of the format RPN.R0 to RPN.R9 for RPN macros and RPN.A0 to RPN.A9 for algebraic macros. The "\M" function will ask you for a digit from 0-9 to specify the file number and then will write the macros in the appropriate file, creating the file if necessary or overwriting an existing file. When the program is executed it looks by default for macro file 0. You can specify which file you want it to load by a 0 to 9 digit in the first parameter. For example, the following will attempt to find and load macros from RPN.R4. RPN R4 The macro file is clear text and as such can be changed with a text editor. Be careful not to disturb the file structure if you try this. The format of the macro save file is: The color attributes: (comment line) - - (14 lines containing color attributes) - - The function key macros for F1-F10: (comment line) - - (10 lines containing macro definitions) - - End of file. (comment line) CHANGING SCREEN COLORS If you feel like experimenting with screen colors, you can do it without recompiling the program. The screen colors used are written out to the macro file and can be changed by a text editor. When a macro file is loaded during program initialization, the colors stored in the macro file will override the defaults in the program. The defaults, in the order they are written are: 15 --- Title color (normal & memory mode) 13 --- Status color (normal & memory mode) 14 --- Register color (normal & memory mode) 12 --- Message color (normal & memory mode) 3 ---- Help color (normal & memory mode) 0 ---- Background color (normal & memory mode) 0 ---- Border color (normal & memory mode) 15 --- Title color (financial & summation mode) 13 --- Status color (financial & summation mode) 14 --- Register color (financial & summation mode) 10 --- Message color (financial & summation mode) 3 ---- Help color (financial & summation mode) 0 ---- Background color (financial & summation mode) 0 ---- Border color (financial & summation mode) Following the color data are the macros, and the final line says "End of file". ERROR MESSAGES The calculator will eventually be told to do something that it can't do. If I haven't thought of it in advance you may get an error from the floating point routines and possibly get kicked out of the program or get incorrect results. If I have thought of it in advance, you will probably get beeped at and get an error message on the message line of the screen. The error messages defined currently are: "Divide by zero." The denominator of an "o", "/" or "%" function is 0. "Illegal argument." From "%" when the argument is too large or not an integer. From "!", "&", "|", and "$" if the argument is not an integer. "Overflow." From factorial if the argument is greater than 170. "Valid in decimal only." From enter exponent if the number base is not 10. "Illegal during macro definition." Macros are not recursive so it's illegal to hit "m" or a macro function key during a macro definition phase. It's also illegal to enter the macro editor during macro definition. "Macros not recursive." If this happens, an "m" was encountered during a macro execution. It was probably entered into the macro during an edit session. "Macro not defined." This happens when a macro execution or copy is requested and the macro definition is found to be empty. "Did not converge." This means that the interest rate calculation in financial mode did not converge to an answer in a reasonable number of iterations. Make sure you are asking it to do something reasonable. "Does not compute." This error is the result of one of several tests for gotchas during any of the financial computations. What it generally means is that the data in the financial registers is inconsistent and an answer can't be computed. An example of this is the attempt to compute a number of payments necessary to pay off a loan when the periodic payment specified is less than or equal to the periodic interest on the loan. "Floating point domain error." This is from an error code set by a floating point routine when it gets input which is invalid for the function being computed. Example: Arcsin or arccos of a value greater than 1. "Floating point range error." This is from another error code set by a floating point routine when a calculation results in overflowing the range of numbers the computer can represent internally. Example: exp(695) ------------------------- In this document I have made reference to the following: IBM PC, PC jr, and PC-DOS are trademarks of IBM Corp. MS is a trademark of Microsoft. Sidekick is a trademark of Borland International