Copyright (c) SuperMemo World 1991 All Rights Reserved This file is NOT the manual mentioned in other SM6 files (README.TXT, OFFER.TXT, etc.) This file is just a selection of the most significant pages from the SM6 help. The following topics are covered: 1 Introduction 2 Basic operations in SM6 3 Main screen 4 Browsing mode options 5 Menu options 6 Hot keys 7 SuperMemo databases 8 General principles of SuperMemo 9 Hints and tips for SuperMemo users 10 Error reports 11 Passwords For a detailed description of terms used here refer to the SM6 help index (press F6 when in SM6). 1 INTRODUCTION ============================================= SuperMemo 6.0 Professional Help File, ver. 1.0, Aug 12, 1991 (c) SuperMemo World 1991 Welcome to the world of SuperMemo - your gateway to better life, better self-image and better future for all of us. In the breathtaking days of unprecedented growth of our civilization, the individual demand for new knowledge and better understanding is overwhelming. Capabilities of human memory became a bottleneck of creative thinking. To keep pace with new ideas and solutions we need a revolution in the methodology of learning. SuperMemo lays ground for such a revolution. Given a dose of persistence, you can witness an amazing boost of your intelectual powers, memory capacity and knowledge of whatever you choose to be worth knowing. This is not a commercial hype. SuperMemo has been developed on purely scientific basis, and in strictly scientific terms it can defend its underlying principles, as well as to demonstrate its incredible effectiveness. Get it working now! Knowledge that is to be learned by means of SuperMemo 6 must be represented in a text form, as a collection of question-answer pairs, later called items. A collection of items will later be referred to as a database. SuperMemo allows to learn and retain knowledge represented in the database at the speed that is close to the maximum natural capability of human memory. This is achieved by means of repetition scheduling, i.e. computing exactly when and which item should be repeated in order to produce the maximum memory formation effect. In spite of what it may seem at first, the SuperMemo-based learning is straightforward. There are only two simple operations you have to learn in order to start: 1. How to create a database 2. How to learn and retain the knowledge To ensure the security of your databases, you should also learn few facts about database maintanance. And that's all! All the rest is optional! Do not delay. Start working now. You have much to win. BEST OF GOOD LUCK! 2 BASIC OPERATIONS IN SM6 ========================================== 2.1 CREATING A NEW DATABASE ------------------------------- You do not have to create your own databases. You can obtain them from your colleagues or from the SuperMemo World's Database Bank. To create your own, customized database do the following: 1. Initialize the files of your database by choosing the Miscells:Initialize option (press Alt-I, and provide an 8-letter name for the database) 2. Add new items to your database by choosing Expand (press Alt-E) 3. Close the database (e.g. by Quit). 2.2 EXPANDING THE DATABASE ------------------------------ You can add new items to the database by means of the Expand option. Follow the steps: 1. Select Expand (e.g. by pressing Alt-E) 2. Type in the question 3. Press PgDn to get to the answer pane 4. Type in the answer 5. Press ESC 6. If you want to terminate the Expand option then press ESC again, otherwise go to Step 2 See also: Editor Simplicity 2.3 LEARNING --------------------------- In order to learn new items and repeat items that have already been memorized, go through the Learn option (press Alt-L to start) REMEMBER: TO MAKE YOUR WORK TRULY EFFECTIVE YOU SHOULD PASS THE LEARNING OPTION EVERY DAY OR ALMOST EVERY DAY!!! It is better to work moderatly in regular intervals than to rack one's brain to death just from time to time. See: Advice Optimization Each day, the learning process is composed of the following stages: 1. repeating items scheduled for repetition on the current day 2. learning new items from the database 3. repeating again items that appeared to be most difficult in Steps 1 and 2 Each of these steps is a series of single repetitions that proceed as follows: 1. displaying a question 2. answering the question (verbally or mentally) 3. displaying the answer (for comparison) 4. inputting the response quality <0..5> The response quality provided by the learner after each repetition is intended to tell how good the response was. The grades/scores are: 5 - excellent response 4 - correct response after a hesitation 3 - correct or slightly distorted response produced with an effort 2 - incorrect response where the correct answer seemed obvious 1 - incorrect response where the correct answer can hardly be recalled 0 - complete blackout In SuperMemo, you are unlikely to use response qualities 0 and 1! 2.3.1 REPEATING OUTSTANDING ITEMS Repetition of outstanding items comes as the first stage of the Learn procedure. In this stage, all items scheduled on the current day, as well as all outstanding items from days on which repetitions has not been made, are repeated in turn and scheduled on other dates (depending on the corresponding values of optimal intervals). All items that score below 3 are considered forgotten, and items that score below 4 are scheduled for additional repetition on the same day (Final drill). 2.3.2 MEMORIZING NEW ITEMS The second stage of the Learn procedure consists in memorizing new items, and is passed through only after an explicit request from the learner. New items taken from the Intact group are memorized in the sequence of appending to the database. Items that scored below 4 are scheduled for additional repetition on the same day (Final drill). 2.3.3 FINAL DRILL This is the third and the last stage of the Learn procedure. Items that scored below four in the two previous stages, Outstanding/Current and Memorization, are repeated again and again until they all score at least four in the response quality assessment. 3 MAIN SCREEN ============================================ The organization of the main SuperMemo screen is as follows: ÉÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ SuperMemo 6 ÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ» º Options of the main menu º ÌÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍËÍÍÍÍÍÍÍÍÍÍÍÍ͹ º Question pane º Process º ÌÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ͹ pane º º Answer pane º º ÌÍÍÍÍÍÍÍÍÍÍÍÍÍÍËÍÍÍÍÍÍÍÍÍÍÍÍÍ͹ º º º º º º Item pane º Test pane º º º º º º ÈÍÍÍÍÍÍÍÍÍÍÍÍÍÍÊÍÍÍÍÍÍÍÍÍÍÍÍÍÍÊÍÍÍÍÍÍÍÍÍÍÍÍͼ 3.1 QUESTION AND ANSWER PANES ----------------------------------- These are the two parts of the main SuperMemo screen where the question and answer of the currently processed item are displayed. If the answer pane is not empty, then SuperMemo is in the browsing mode (except for the moment when the response quality is being input at learning). In the browsing mode, the item currently presented in the question and answer panes may be edited, deleted, duplicated, reset, printed etc. 3.2 ITEM PANE -------------------------- The item pane is the part of the SuperMemo main screen where the information about the currently processed item is displayed. The information pertains to the item whose content is presented in the question and answer panes, and is as follows: Number - number of the item Grade - recently input response quality for the item (if any) Repetition - number of the repetitions the item have been subject to, separated by the colon from memory lapses (number of times the items has been forgotten) Factor - E-factor characterizing the difficulty of the item Used factor - recent ratio I(n)/I(n-1) where I(n) is the current interval of the item (for Repetition=1, Used factor equals the current interval) Interval - current interval that separates the last and the nearest repetition of the item Last date on which the item has been repeated (if any), separated by a hyphen from the next date on which the item is scheduled for the nearest repetition. 3.3 TEST PANE ------------------------------ The test pane is the part of the SuperMemo main screen where the information pertaining to the currently performed repetition is displayed. It consists of the following: RF - old and the newly computed value of the retention factor corresponding to the relevant E-factor and repetition number OF - old and the newly computed value of the optimal factor (the optimal factor matrix is a smoothed equivalent of the retention matrix) OF support - number of repetition instances used to compute the RF and OF factors displayed above New factor - newly computed value of the E-factor for the repeated item Opt interval - next optimal interval for the repeated item (it equals to the value of the previous interval multiplied by the relevant entry of the OF matrix) New interval - next value of the current interval computed for the repeated item (it equals the optimal interval +/- a random deviation) Next rep - date of the next repetition 3.4 PROCESS PANE ----------------------------- This is the rightmost part of the main SuperMemo screen where most important information about the currently used database and its optimization process are displayed. An empty process pane indicates that no database is in use. The information presented in the process pane is as follows: Database - name of the database in use Date of initialization of the database in use separated by a hyphen from the current date Day - number of the day of the process (since initialization) Memorized - number of items in the database that have already been memorized by means of Learn) Intact - number of items in the database that have been added by means of Expand and have not yet been memorized Total - total number of items in the database (Total=Memorized+Intact) Outstanding - number of items that should be repeated on the given day separated by a colon from items that will be repeated again because of response quality below 4 Burden +/- change of the Burden parameter in the current session (see below) Burden - average number of items that are repeated per day Mean time - average time used for repeating a single item (in seconds) Workload - average time used for repetitions per day (Workload = Mean time * Burden) Interval +/- change of the Interval parameter in the current session (see below) Interval - average current interval assigned to items in the database Factor - average E-factor assigned to items in the database Repetition - average repetition number assigned to items in the database separated by a colon from the average number of memory lapses per item Lapses +/- change of the Lapses parameter in the current session (see below) Lapses - forgetting index recorded in repetitions (the number in parentheses indicates the requested value of Lapses) Retention - estimated retention of knowledge for the database in use Grade - average response quality Time - repetition time for the current session 4 BROWSING MODE OPTIONS ========================================== Browsing mode is the state of the SuperMemo program in which such database operations as Edit, Delete, Select, Print, etc. are available (press E, D, S, P, etc.). SuperMemo enters the browsing mode in the following contexts: - after each repetition in Learn - after finding an item by Find - after choosing an item by Select - after displaying an item assigned on a given day in Burden - after each repetition in Random test. Browsing mode operations are summarized in the status line and are as follows: Edit - edit the currently displayed item Delete - delete the currently displayed item Select - display an item of a given number and return to the browsing mode Print (^P) - print the currently displayed item Reset (^R) - reset the currently displayed item, i.e. remove it from the set of memorized and treat it as an intact, not yet memorized item Duplicate (^D) - duplicate the currently display item, edit the copy and return to the browsing mode Last (^L) - repeat the last operation of substring replacement Renew - display the updated information in the Item pane (useful during the test to inspect the new value of the Used factor) Arrows - move one item forward or backward PgUp/PgDn - move a couple of items forward or backward Home/End - move to the first or last item Enter - continue (or move one item forward) ESC - exit 4.1 EDIT -------------------- The item editor is called by the following commands: - Expand (adding new items to the database) - Edit (in the browsing mode) - Duplicate (in the browsing mode). The editor makes use of the Question and Answer panes of the main screen. The following commands are available: Arrow keys - move the cursor (only within the current pane) Ctrl-> - move one word right Ctrl<- move one word left Backspace - delete the symbol preceding the cursor, and paste lines if the cursor is placed in the first column Del - delete the symbol at the cursor's position, and paste lines if the cursor is placed at the end of the line PgUp/PgDn - change the current pane (from Question to Answer and vice versa) Enter - cut the current line and move to the beginning of the next line Ins - change the mode from Insert (new characters are inserted into the text) to Overwrite (new characters overwrite the old ones) Tab - swap the panes Home/End - go to the beginning or end of the line ^Y - delete the current line ^T - delete the word to the right of the cursor ^QY - delete all the characters from the cursor's position to the end of the line ^S - swap the current line with the next one (used for reordering enumerative answers) ^R or ^QA - replace a string of characters with another string ^L - repeat the last replace operation F7 or ^KB - mark the beginning of a block F8 or ^KK - mark the end of the block and move the block to the buffer ^KC - copy the block stored in the buffer to the current cursor's position ESC - move to the Answer pane (if in the Question pane), or exit the editor (if in the Answer pane) Note that when using Expand, the ESC key will allow to leave the editor, but Edit will be reentered in order to allow to input more items. Therfore, to quit the editor while using Expand press ESC twice (or three times if in the Question pane). 4.2 DELETE --------------------- This operation is available only in the browsing mode. It is used to delete the currently displayed item. Before deleting is performed, SuperMemo will ask you to confirm your decision. Delete preceded by Duplicate may be used to move the item to the end of the database. 4.3 SELECT --------------------- This operation of the browsing mode allows to display the item of a given number. It may be used whenever any of the SuperMemo procedures reports a database integrity error. In such cases, the number of the defective item is often displayed, and can be used with Select to inspect the problem. 4.4 PRINT --------------------- This operation of the browsing mode should be used only if a printer is connected to the computer (LPT1). Print sends the currently displayed item to the printer, and can primarily be used to simplify items in the database. Inexperienced learners are inclined to produce monster items which should be split into many simpler ones. In case of discovering an item that causes persistent problems, print it, delete it, and use the printout for formulating an equivalent set of simpler and univocal items. 4.5 RESET -------------------- This operation of the browsing mode allows to transfer an item from the Memorized to the Intact category. In other words, the reset item will be treated as a new, intact item, and all the process parameters will be updated accordingly. The reset should be used each time an item has been substantially reformulated. It can also be used in case of mistakenly providing a pass grade to a forgotten item when the new interval is very long. 4.6 DUPLICATE ------------------------- This operation of the browsing mode allows to use old items to produce new, similar items. For example, if the old and the new item differ by just only one phrase, it is easier to duplicate the old item, and change the phrase in the newly created duplicate. Duplicate acts as follows: - make an intact copy of the duplicated item (placed at the end of the database) - enter the editor to edit the newly produced duplicate - return to the old item - reenter the browsing mode. 4.7 LAST -------------------- This command of the browsing mode allows to repeat the last replace operation used by the item editor. If no replace operation has been made since loading SuperMemo, the Last command will have no effect. Similarly, it will have no effect if the replaced string is not a part of the current item. In combination with Find, Last is used to replace a given string of characters throughout the database. For example, if you want to replace the string '(mathematics)' with '(math)' throughout the database, do the following: 1. use Find to search for '(mathematics)' 2. as soon as the first item has been found, enter the editing mode (E) and define the replacement string by pressing ^R (it will replace the string in the current item) 3. exit the editor and press Enter to continue the search 4. if another item has been found, decide if you want to replace the string, press ^L if yes, press Enter and go to 4 4.8 RENEW ---------------------- This is a specialized option of the browsing mode used only during repetitions. It is used to inspect the value of the U-factor (Factor used), which is: - before the Grade is provided: the planned value of the U-factor - after the Grade is provided: the real value of the U-factor (used in optimization) - after using Renew: the new, planned value of the U-factor. 5 MENU OPTIONS ============================================ 5.1 OPEN --------------------- To open a database use Open (F3 or Alt-O). Open is also automatically called when you want to use any of the options that need an open database. Open selects and prepares a database for use. It displays a window with the names of all databases stored in your database directory. Use arrow keys and Enter to choose the database you want to work with. If you want to change the drive or path to the database directory press TAB to move between panes of the Open window. The panes are as follows (clockwise): Files, Drive, Path and Template. You can define a database filename template using wildcard characters ? and *, e.g. if you have the following database in your directory (geography and economics: GEOG_A, GEOG_B, ..., GEOG_N and ECON_A, ECON_B, ..., ECON_N, you can define the template GEO*.INF to work only with databases on geography. Do not change the filename extension in the template pane. If no database is listed in the Open window then either you haven't created any database or the database path is wrong. 5.2 LEARN AND EXPAND --------------------- For information about these options see chapters 2.2 and 2.3 in this file. 5.3 FIND ----------------- Find is used to search the database for items containing a defined string of characters (press Alt-F to call it). Upon chosing this option, input the string you want to search for (search string). As soon as an item has been found, it is displayed in Question, Answer and Item panes, and SuperMemo enters the browsing mode. The most recently used string is stored in in the SM6.PAR file as soon as you use the Write parameters option. 5.4 SELECT ------------------- Select is used to display the most recently used item (press Alt-S to call Select from anywhere in the SuperMemo program). Upon displaying the item, SuperMemo enters the browsing mode which its operations available at a keystroke (e.g. Edit, Delete, Reset, Print etc.). You can also select an item by its number. In such a case press Alt-S to get into the browsing mode and press S again. 5.5 MISCELLS ---------------------- Miscellaneous options of the Miscells menu: - Date - setting the DOS date - Random test - random presentation of items from the database in use - Garbage - database garbage collection used also to test database integrity - Initialize - creation of a new, empty SuperMemo database - OS Shell - temporary exit to DOS - Parameters - menu of SuperMemo parameters - Approximate - procedure used to compute the OF matrix on the base of the RF matrix - Close - closing the database in use. 5.5.1 DATE The Date option allows to set the current date used by the SuperMemo program. The DOS date in your computer will also be changed accordingly. The date format depends on the setting of the Dates parameter in the Parameter menu. Therefore, it may be DD.MM.YY or MM.DD.YY, depending on your choice. The Date option is automatically called if SuperMemo discovers that the DOS date setting is much different from what could be expected from data stored in the database. REMEMBER TO KEEP THE DATE SETTING ACCURATE! 5.5.2 RANDOM TEST This option randomly selects items from the database and presents them by displaying the question, and after you press a key, the answer. Press ESC to leave Random test. You can use Random test to: - test your colleagues or students - inspect the content of the database - compute the participation of particular topics in the database - check if the proportion of knowledge you remember corresponds to the value approximated by the Retention parameter 5.5.3 GARBAGE This option is used to remove superfluous data from database files. This process may substantially reduce the size of some of the files (esp ITM). As Garbage performs a rough database integrity checkup, it is recommended that it be performed once a month or two. The garbage directory stores temporary work files (comparable in size to the database). As Garbage is very read-write intensive, it should be performed on fast media. WARNING! Back up your files before garbage collection. 5.5.4 INITIALIZE This option of the Miscells menu allows to start a new SuperMemo database, and is activated by pressing Alt-I (or Alt-T:M:I). It creates a set of four files that are marked by a unique filename. These files will store the collection of items and the information about the learning process. Initialize asks for the filename, and if it conflicts with already existing databases, you will be requested to confirm the decision to override. 5.5.5 OS SHELL You can exit SuperMemo temporarily by using the OS Shell option (Alt-X). This option loads a new command interpreter, and allows you to issue commands of the DOS operating system. As soon as you wish to return to SuperMemo, type Exit at the DOS prompt and press Enter. The database in use is automatically closed in case your forget returing to SuperMemo. If your have a number of resident utilities in memory of your computer, there may be not enough space for loading COMMAND.COM, and the 'Out of memory' message will be displayed 5.5.6 PARAMETERS The Parameters menu is displayed by Alt-P, and allows to define the following parameters Lapses - requested forgetting index, i.e. proportion of items that are allowed to be forgotten at repetitions (this value may vary from 3-20%) Paths - directories used by SuperMemo Sound - generation of sound signals Color - color display Dates - date format (DD.MM.YY or MM.DD.YY) It also allows to Write parameters to the disk or Read parameters from the disk. 5.5.7 APPROXIMATE This option of the Miscells menu allows to reevaluate the current state of the OF matrix on the base of the RF matrix, and should be passed through each time the Lapses option is used to request a new value of the forgetting index. The procedure makes use of a hill-climbing algorithm which is intended to find parameters of a predetermined function of optimal factors in order to find the closest fit to the matrix of retention factors. This function is then translated into the matrix of optimal factors. The predetermined shape of the function of optimal intervals was established on the base of results yielded by working with earlier versions of SuperMemo (particularly SuperMemo 5). For the repetition number equal 1, the function of optimal intervals has a close-to-linear nature and is approximated by a procedure akin to linear regression with the number of repetition instances considered in the computation. For the repetition number greater than 1, the approximation procedure is iterative in nature. In this case, the four parameters of the function of optimal intervals are initially set to average, expected values, and the algorithm proceeds toward minimizing the objective function called the deviation (sum of square differences between OFs and RFs). During iterations, all the parameters are displayed on the screen, as well as steps used in the hill-climbing algorithm (akin to the Rosenbrock method). The counter variable indicates the number of steps and iterations of the algorithm as it progresses. The progress variable indicates the decrease of the objective function (DEVIATION) in particular steps and iterations, as well as the trailing average of the iteration progress. Iterations proceed until the trailing progress drops below 0.01 or ESC is pressed. In the former case, the newly computed parameters of the function of optimal factors are displayed along the final value of the objective function. Selected entries of the RF and OF matrices are juxtaposed in order to provide ground for accuracy judgement. In case the fit is satisfactory, Ctrl-S will store the newly computed matrix of optimal factors in the database. Apart from changing the Lapses parameters, the Approximation procedure may also be used in cases where the smoothing scheme seems to be unsatisfactory and the edge of the matrix of optimal intervals, displayed by Progress on the Analysis menu, assumes a jagged, irregular shape. WARNING! The approximation algorithm should only be used after changing the Lapses paramater, unless the learner truly understands the optimization procedures. 5.5.8 CLOSE You can close the database in use by means of Close (F2 or Alt-C). Close updates all database files and closes them. Close is automatically called by: - Quit - Open, if you decide to open another database - OS Shell (to protect your database in case you forget returning to SuperMemo) 5.5.9 USER DATA The User option on the Miscells menu allows to inspect information about the person, institution or company that bought the marked copy of SuperMemo. This information does not necessarily pertain to the end user of SuperMemo. A given copy of SuperMemo may be purchased for retail sale, and the end buyer becomes a legal user upon registering his copy at SuperMemo World. 5.6 ANALYSIS ------------------------ The Analysis options allow to see: - Burden - number of items to be repeated on particular days - Factor distribution - distribution of E-factors across the database - Interval distribution - distribution of current intervals - Optimal factors - the OF matrix - Support for Optimal factors - matrix of repetition cases used to compute OFs - Prospects - matrix of optimal intervals - Retention - approximation of the OF matrix based on sketching forgetting curves 5.6.1 BURDEN The Burden option can be found in the Analysis menu (press Alt-B to call it). It is used to display how many items are assigned for repetition on particular days and which these items are. The following keys can be used when in the Burden option: - arrow keys - move the selection bar between days of the current page - PgUp/PgDn - present the next/previous page of days - Ctrl-PgUp/PgDn - move one year forward or backward - Home - move to the earliest day in the process that has not yet been gone through - End - move to the furthest day on which any repetition is/was scheduled - Enter - display the list of items scheduled on the day indicated by the selection bar (see Assignments) - ESC - exit to the Analysis menu Use the Burden option to see how many items you have in the schedule for the nearest days. You are advised not to inspect these items as it may interfere with repetitions. ASSIGNMENTS You get to the Assignments option by using Burden (e.g. Alt-B), positioning the selection bar on a given day and pressing Enter. In the right part of the screen, the list of items scheduled on the selected day is displayed. The selection frame can be moved to any of the items, and the content of the Question, Answer and Item pane will change accordingly. The selection frame may be moved by means of arrow keys, PgUp and PgDn. Upon moving to the item of interest, most of the operations of the browsing mode are available: - Edit, - Delete, - Reset, - Print, - Select, - Duplicate, etc. By pressing Enter, the browsing mode may be entered with all its operations at hand. The browsing mode may be exited by ESC. 5.6.2 FACTOR DISTRIBUTION This option displays the number of items belonging to particular difficulty categories (determined by E-factors). The range of E-factors is displayed leftmost, followed by the number of items in each category, their proportional participation in the whole database and the histographic representation of the latter. A 'flatfoot' distribution with a large number of 1.3 items is the best indicator of an ill-structured database, and should be, by all means, avoided. 5.6.3 INTERVAL DISTRIBUTION This option presents the distribution of current intervals attributable to particular items. The interval category is displayed in the left column, and is followed by the number of pertaining items and their proportional participation in the database. Initially, soon after creating the database, intervals are on the order of days, but in a short time, most of them increase to months and years. Then, only a small proportion of items is scheduled in short intervals, and is accounted for by marginal, but existent process of forgetting. 5.6.4 OPTIMAL FACTORS This option displays the matrix of optimal factors (OF matrix). This matrix is a tabular representation of the function of optimal intervals. Columns are designated by particular E-factor values, while rows correspond to repetition numbers (memorization is counted as the first repetition). PgUp and PgDn can be used to display lower and higher half of the matrix, corresponding to lower and higher E-factors. See: Optimization 5.6.5 OPTIMAL FACTOR SUPPORT This option of the Analysis menu allows to see how many repetition instances have been used to compute particular entries of the matrix of optimal factors. The top row displays E-factor categories (from 1.3 to 2.5). The leftmost column displays repetition nos (from 1 to 20). The parameter displayed right to the window header is the number of instances used to compute the Lapses parameter (FI stands for Forgetting index) 5.6.6 PROSPECTS This option allows to see the approximate values of optimal intervals used in the process of learning. In the top row, E-factor categories are displayed (from 1.3 to 2.5). The leftmost column displays repetition nos (from 1 to 20). Use PgUp/PgDn to inspect both halves of the matrix (E-factors equal up to 3.2). First optimal interval in a given E-factor category equals the first optimal factor for the same category. All subsequent optimal intervals are computed by multiplying the preceeding value of optimal interval by the corresponding value of optimal factor: OI(1,EF)=OF(1,EF) OI(n,EF)=OI(n-1,EF)*OF(n,EF) where: OF(n,EF) is the n-th optimal factor for E-factor equal EF OI(n,EF) is the n-th optimal interval for E-factor equal EF Intervals longer than the average human lifespan are not displayed. (M stands for months, Y stands for years) 5.6.7 RETENTIONS This option allows to see the content of the matrix of retention factors. Retention factors correspond to O-factors, but their values are computed directly from forgetting curves sketched for particular E-factor categories and repetition numbers. Retention factor expresses the optimal interval (first repetition) or the increase of the optimal interval (further repetitions) for which the expected value of the forgetting index equals the desired value of Lapses set by the Lapses option. For example, if the Lapses parameter equals 5% and the repetition number is one, then the corresponding retention factor will equal the interval for which the forgetting curve indicates the expected 5% of memory lapses. Because the forgetting curve sketching is subject to high degree of noise, the retention factor matrix undergoes smoothing before it can be treated as the matrix of optimal intervals. The smoothed form of the matrix of optimal intervals may be inspected by means of the Optimal factors option. You can use arrow keys, Home and End to move across the matrix of retention factors. You can see the current state of the forgetting curve corresponding to a given retention factor by pressing Enter, while in the Retentions option (Alt-R). If you have a color display, you may see that the color of retention factor entries changes along with the increasing number of repetition instances uses to sketch the curve Obviously, you may wish to inspect only those entries whose color intensity indicates that the forgetting curve sketching is in an advanced stage. See: Optimization FORGETTING CURVES You can view forgetting curves by using Retention (Alt-R), positioning the cursor on a selected entry of the matrix of retention factors and pressing Enter. The top row of the forgetting curve window indicates the values of the R/U-factor (equivalent to the interrepetitory interval if the repetition number is 1). The bottom row of the window displays the number of repetition instances used to sketch a given section of the curve. The vertical axis corresponds to the forgetting index (top-0%, bottom-100%). The broken lines that cross inside the forgetting curve panel indicate: - horizontal - desired value of the forgetting index (top corresponds to 0%) - vertical - Retention factor (factor for which the forgetting curve crosses the line of the desired forgetting index) Texture of rectangles used to plot the curve reflects the number of instances used to compute the retention for that part of the graph: the more solid the rectangle, the greater the number of instances, and in consequence, the greater the accuracy of plotting. 5.6.8 MONTHLY BURDEN This option allows to see how many items are scheduled for repetition in particular months and years (press PgUp/PgDn to see the next or previous panel of four years). 5.7 HELP MENU ----------------------- Help menu provides the following options: - Front page - process help starting from its front page - Index - process help starting from the index which lists all the most important topics - Recent page - process help starting from that page that has recently been used - Help on Help - process help starting from the page that provides instructions on how to use help - Hints and Tips - process help starting from the Hints-and-Tips page 5.8 QUIT --------------------- This option of the main menu should always be used to terminate the work with SuperMemo. To terminate press Alt-Q, and confirm the termination by pressing Y(es). Quit closes the currently used database (if any), restores the video attributes from before running the program and disposes memory area occupied by SuperMemo. 6 HOT KEYS ========================================= The following hot keys may be used in most of contexts of SuperMemo: F1 - context-sensitive help F2 - Close the database in use F3 - Open a database F4 - recent page of help F5 - front page of help F6 - help's index F10 - main menu Alt-O - Open a database Alt-L - Learn Alt-F - Find Alt-S - Select Alt-E - Expand Alt-D - change Date Alt-B - daily Burden Alt-M - Monthly burden Alt-C - Close the database in use Alt-T - Top options (main menu) Alt-P - Parameters Alt-X - eXit to DOS (OS Shell) Alt-H - Help menu Alt-Q - Quit 7 SUPER-MEMO DATABASES ============================================ 7.1 DATABASE FILES --------------------------- The following files make up a SuperMemo database: .INF - info about the process .ITM - text of items .ITI - info about items .DAT - list of pointers to lists storing items scheduled on particular days Optionally: .REP - list of items scheduled for the Final drill (this file can always be erased). 7.2 DATABASE SECURITY PRINCIPLES ---------------------------------- Upon loosing your database files, you are doomed to forget 60-100% of your knowledge in the first year depending on how long you have been using SuperMemo. To make sure its is not the case, please do obey all the database security principles listed on the next page. Recently, computer viruses have become a scourge of hard working and honest computer users. SuperMemo databases are particularly sensitive as they store highly complex data produced by months of strenuous work of a SuperMemo learner. YOU CANNOT BE TOO CAREFUL! Security principles: - always leave SuperMemo by Quit - keep your passwords handy - in case of abandoning the program in a prohibited way (e.g. power failure), use your backup copy - do not modify database files manually - keep a month and two months old database backups on diskettes - create a temporary backup each time you work with SuperMemo (e.g. use RAMdisk) - perform Garbage collection once a month - do not create databases greater than 300 kB or 3000 items - keep the DOS date setting accurate - upon DATABASE INTEGRITY ERROR message, backup your database (do not override older backups) and perform Garbage collection - do not work with databases that report integrity errors, use your backups - recover from any damage only be means of tools provided by SuperMemo World - in case of a serious damage, send your database to SuperMemo World for recovery (remember to attach postage and the processing fee as indicated in files accompanying SuperMemo) 8 GENERAL PRINCIPLES OF SUPER-MEMO ============================================= All forms of learning produce molecular changes in neuronal synapses which form connections between nerve cells. These changes are gradually obliterated in the process of spontaneous forgetting. Forgetting affects all synapses able to learn and can be naturally prevented only by one mechanism: repetition. Every attentive learner knows, that forgetting can ruin the delicate fabric of knowledge that may take months and years to build. The saying 'Repetitio est matter studiorum' is as old as the art of learning. Not everybody, however, realizes the fact that: IT IS NOT POSSIBLE TO LEARN ANYTHING ONCE AND REMEMBER IT FOREVER! Even one's own name could be forgotten if it were not used, or in other words repeated, as often as it is. The problem with repetitions is that they consume one of the most valuable assets in the modern lifestyle: time. Therefore the key to effective learning is to find ways to reduce the number of otherwise ineluctable repetitions. SuperMemo makes learning approach the maximum natural capacity of the human brain to form memories. This is done by the optimization of repetition scheduling, i.e. finding out when and which portions of knowlege should be repeated. Considering a given piece of knowledge, there are two criteria that can be used to determine the length of optimal intervals that should separate repetitions: - minimization of the number of repetitions - maximization of the overall knowledge retention. In other words: - intervals should be as long as possible to reduce time necessary for repetitions - intervals should be short enough to make sure that the repeated knowledge is not forgotten. Because of the fact that forgetting has a stochastic nature, i.e. it cannot be easily predicted when a given piece of knowledge will be forgotten, the statistical approach to the process of learning must be applied. By means of statistical analysis, we can determine when a given proportion of memorized knowledge will be forgotten. Thus, the following definition of optimal intervals could be accepted: OPTIMAL INTERVALS ARE INTERVALS THAT RESULT IN A SMALL, PREVIOUSLY DETERMINED FRACTION OF KNOWLEDGE BEING FORGOTTEN. The mentioned fraction of knowledge, later called the forgetting index, may be chosen by the learner, and usually falls into the range from 5% to 10%. By application of optimal intervals in the process of learning, the SuperMemo method produces an incredible increase in the rate of knowledge acquisition without affecting knowledge retention. A little problem arises when we observe that optimal intervals differ between different sorts of knowledge and different learners. The former problem is dealt with in the SuperMemo method by splitting knowledge into smalles possible pieces called items. The optimization procedure, i.e. computation of optimal intervals, is then applied to each of the pieces separately, producing a unique repetition intervalization in each of the cases. The principle of applying items of maximum simplicity is later called the minimum information principle. The problem of differences between learners is solved in SuperMemo by application of self-modifying algorithms that adjust repetition intervalization to individual needs. In other words, the procedures used by the SuperMemo program can detect what sort of learning and what sort of learner are subject to optimization. The net result is that a determined level of knowledge retention may be maintained in the process of learning that approaches the maximum natural speed with which the learner's brain can form memories. This is an outline of how SuperMemo allows to learn fast: - the learned knowledge is split into smallest possible pieces called items - items are formulated in the question- answer form - items are memorized by answering questions as long as it takes to provide correct answers to all of them - first repetition is scheduled at the same time for all items, after the interval which statistically produces FI lapses of memory (FI stands for the forgetting index and equals 5% be default) - the first interval is computed as for an average learner, but as soon as the FI deviates from the requested value, the length of the first interval is modified - after each repetition, learner produces a number called the response quality which determines how difficult it was to answer the question - on the base of response qualities, items are classified into difficulty categories, and their difficulty is reestimated in all successive repetitions - different optimal intervals are applied to items of different difficulty - different intervals are applied to items that have been repeated a different number of times - the function of optimal intervals is constantly modified in order to produce exactly FI memory lapses - the function of optimal intervals is represented as the matrix of optimal factors OF: I(1,EF)=OF(1,EF) I(n,EF)=I(n-1,EF)*OF(n,EF) where: I(n,EF) - n-th interval for difficulty EF OF(n,EF) - n-th O-factor for difficulty EF - matrix of optimal factors is produced by smoothing the matrix of retention factors RF - each entry of the RF matrix equals the current estimation of the optimal factor OF that produces FI lapses of memory - RFs are computed from forgetting curves sketched on the base of repetitions - each difficulty category and repetition number has its own record of repetitions in the form of a forgetting curve sketch - intervals used in learning are slightly dispersed round the optimal values for the accuracy of the forgetting curve sketching 9 HINTS AND TIPS FOR SUPER-MEMO USERS ========================================== To be truly satisfied with your SuperMemo workouts, you should obey the following principles of a successful learner: - regular work - formulation of simple and univocal items - moderation - concentration - proper selection of material - multifaceted approach - application of mnemonic techniques - health - pleasure - knowledge of the above principles. 9.1 REGULAR WORK ------------------------- There is no way of avoiding the fundamental principle of SuperMemo: REGULAR WORK. It is better to work 3 minutes every day, than to spend an hour on weekends. Regular work may be against the human nature, it may be hard to bear for some, but the mechanisms of memory are ruthless, and do not provide any escape from regular brain-racking Moreover, the positive feedback between the pile-up of outstanding repetitions and the urge to procrastinate is the main source of SuperMemo casualties. DO NOT POSTPONE! 9.2 SIMPLICITY AND UNIVOCALITY OF ITEMS ---------------------------------------- One of the main ways to increase the speed of learning with SuperMemo is to consequently formulate items that contain minimum information and are unambiguous. Minimum information does not mean minimum number of items. An ill-structured item may often be replaced by a dozen of simple items, and the resulting redundancy is likely to greatly reduce the learning time! One should never learn convoluted formulas, lists containing more than 3-5 elements, sets containing more than 2-3 elements, descriptions, etc. All of them should be replaced by a set of simpler items. The notable exception are short epigrams, however, their use may be constrained by the SuperMemo's limit on the size of an item. Most of SuperMemo beginners display the lack of sense of item simplicity. Despite all warnings, they formulate monster items that by no means can be reproduced at repetitions. Usually, the learner will have to learn the hard way that there are different knowledge representation strategies that may render the same piece of knowledge a pushover in one case, and a nightmare in another. 9.3 MODERATION ---------------------- A dangerous snag awaits all inexperienced SuperMemo learners as a result of two phenomena: - it is easier to memorize items than to retain them in memory over a longer time - in the first year of work, the learning rate will gradually be reduced as a result of accumulation of items scheduled for repetitions. In consequence, the learner, bedazzled with the initial progress, may overestimate his or her ability to sustain the working regime. Soon, he or she will be overwhelmed by the number of repetition, and the crisis may be amplified by the increasing difficulty to recall the ill-structured items. One of the first things a SuperMemo learner must do is to test his capability to work with SuperMemo over a longer period of time. The best daily dose for beginners is 5-10 min Only after a month or so, it could gradually be increased to 30-40 minutes per day. It is worth noting that the decrease in the speed of learning becomes practically undetectable as soon as after the first year of the process. 9.4 CONCENTRATION ----------------------- It is important for the quality of knowledge obtained by means of SuperMemo, that repetitions are conscious and semantic, i.e. they do not just reproduce the answers automatically, without truly grasping all their implications. Automatic repetitions not only reduce the value of the knowledge. They also make the work with SuperMemo monotonous, and are less engaging for memory. In consequence, the repeated items are more difficult to recall. A truly successful SuperMemo learner finds a challenge in each repetition. 9.5 SELECTION OF MATERIAL ------------------------------ One of the most important prerequisites of a success with SuperMemo is the proper selection of the learned material. The art is not to know thousands of facts, but to know just those of them that can be directly applied in the day-to-day activities of the learner. Memorizing a phone number of a colleague may produce more profit than knowledge of a sophisticated mathematical formula. A simple rule for drug application may be of greater use than a hundred of drug dosages. The question: 'What do I want to learn and why?' should be a commonplace in the SuperMemo learner's daily schedule. SuperMemo allows to learn fast, and therefore may easily lead somebody to learn a lot of useless garbage without ever stopping to ask WHAT FOR? Applicability must be the primary criterion in choosing new items to learn. Applicable knowledge may be a source of hugely increased effectiveness and personal satisfaction for the learner. Garbage items only soak up learner's time. 9.6 MULTIFACETED APPROACH ----------------------------- It can easily be seen that a dose of smart redundancy can greatly reduce the learning workload, and improve the knowledge quality. Imagine that you want to learn the following piece of knowledge: 'The cost of smoking to the world's economy is $100 billion, while the antismoking campain fund is $10 million' The best way of learning the above facts is to produce a set of items, that ask for all the most important aspects and implications of the learned knowledge. It may appear sufficient to ask two questions about the losses produced by smoking and the cost of the antismoking campain. However, an important implication may soon be lost from the learner's memory: 'What is the mutual cost relation of the disease and the remedy?' Answer: 10,000 !!! Not only implications may be lost. A learner able to tell what disease is caused by HIV, may not be able to answer the question what virus causes AIDS. Mutlifaceted questions increase the quality of knowledge and reduce the learning time! 9.8 MNEMONIC TECHNIQUES ------------------------------ One of the principles of SuperMemo is to formulate simple and univocal items. However, even the simplest item may be quite hard to remember. Numbers are a good example of intractable polluters of SuperMemo databases. The problem with numbers is that they are all similar to each other and can easily be confused. Here, the help comes from techniques which are as old as the art of learning itself: the mnemonic techniques. Any good mathematician may demonstrate that remembering numbers is much less of a problem to him than to an average man in the street. Having had to learn hundreds of constants and formulas, a mathematician more or less consciously knows how to deal with them. The simplest technique is the learn a list of 10, 20 or 100 pictures associated with numbers. For example: 1 - harpoon, 2 - coin, 3 - tripod, 4 - dog, 5 - hand, etc. Knowing such associations, numbers may easily be represented as colorful pictures. For example: 4315 may be seen as a dog on a tripod harpooning somebody's hand. Despite what may seem at first, time spent on forming the pictures is incomparabe with gains produced by successful memorization. By virtue of the brain circuitry developed in the course of evolution, vivid pictures are by far easier to retain in memory than dry numbers. In SuperMemo, it is always better to invest heavily in formulation of items, in order to have only a couple repetitions in the lifetime. Ill-structured or intractable items may be repeated as many as 30 times in the course of a single year!!! 9.9 HEALTH -------------------- Health is an obvious prerequisite of success in any form of learning. Here, only the greatest natural allies of SuperMemo are signalled: - sleep - nothing can ruin a SuperMemo session as effectively as drowsiness - physical exercise - jogging and swimming have salubrious neutrotrophic effects when accompanied by mental effort - oxygen - without fresh air, a SuperMemo learner will experience signs of extreme exhaustion as soon as in the first 5 min - abstinence from smoking and drinking. 9.10 PLEASURE --------------------- The most sure way of telling if all the principles of effective work with SuperMemo are satisfied is user's pleasure experienced during repetitions. This is an infallible litmus test and should constantly be examined in order to detect the causes of unsatisfactory progress. All problems mentioned earlier such as backlog of items, recall problems, monotonous repetitions, garbage knowledge, superficial items, tiredness, etc. are certain to take the gilt off the SuperMemo work and result in disillusionment. MAKE IT FUN! 9.11 KNOWLEDGE OF THE PRINCIPLES ----------------------------------- Each of the ten principles of the effective work with SuperMemo is of a huge importance on its own. The problem is that there are nine of them, and none is less important than the others. Why not to make use of SuperMemo now, and learn the principles to make them available whenever you wish to recall them. By obeying the principles, you may increase your effectiveness manifold. Perhaps, the gain may be as big as from applying SuperMemo in the first place. Be on your guard, the reserves are immense. 10 ERROR REPORTS =========================================== 10.1 INTEGRITY ERROR ------------------------ If SuperMemo displays the message: 'Database integrity error!' you should not continue working with your database as it must have been damaged. Use your backup copies to see if the problem reappears there. If you don't have a backup that reports no errors then you may try to go through Garbage collection (make a backup copy beforehand), and see if it could help to remedy the problem. If the Garbage collection procedure does not work, then you may try to use database recovery tools provided by SuperMemo World. Unless you are an experienced programmer, and you fully understand the structure of SuperMemo databases, any attempt to correct the damage manually or programmatically may cause further loss of data! If you appear to be unable to solve your problem then consult your dealer, or send your damaged database to SuperMemo World (add postage and the processing fee as indicated in the files accompanying SuperMemo 6 on the distribution disks). 10.2 OUT OF MEMORY ----------------------- If SuperMemo displays the 'Out of memory' message, you should quit the program and remove some of your resident utilities. A frequent reason for running out of memory is just a failure to remember that SuperMemo has already been loaded, and could be reactivated by typing Exit (after using the OS Shell option). The typical symptom of memory deficit is unability to restore the content of windows used by the program. 10.3 CANNOT OPEN ... -------------------------- The two most frequent reasons of the 'CANNOT OPEN ...' message are: - the CONFIG.SYS file indicates a too little number of files allowed to open (add FILES=20 line to CONFIG.SYS) - some of the database files are missing from the database directory (note that the COPY command of DOS fails to copy files whose size is zero) You may also be unable to open a database because of a wrong format. 10.4 WRONG DATABASE FORMAT ----------------------------- This error appears whenever you attempt to open a database that was created by a copy of SuperMemo marked by a different registration number. To convert a database to another format you will have to send it to SuperMemo World for conversion. Upon your permission, the database will be added to the Database Bank and distributed worldwide with 50% of net profit coming back to you. 10.5 UNCLOSED DATABASE -------------------------- In case a database has not been closed for any reason, SuperMemo will display a warning message. In such a case, you should use your backup copy, as the unclosed database might have lost its integrity. In case you don't have a backup, try to run the Garbage collection procedure for a rough integrity check-up (make a backup beforehand). If Garbage reports no integrity errors, you stand a 3:4 chance that your database is OK. 11 PASSWORDS =========================================== In order to protect SuperMemo from piracy, SuperMemo World introduced the system of password checking. Every month, SuperMemo will ask the learner to input the relevant password. This password is unique for different learners (different registration numbers) and for different months. Each user is periodically sent a collection of passwords for the coming months, to make sure that illegal copying is slowed down. Because the password is asked for each of the databases, it is recommended that the password be saved in the parameter set as soon as the check-up procedure is over. This will make sure, that in case of all the remaining databases, all you will have to do is to press Enter. To save the password press Alt-P (Parameters) and W (Write parameters). You may also need to press Y to confirm overwriting the parameter file (if it exists). Remember to have your passwords handy each time your work with SuperMemo (e.g. keep the passwords in the SuperMemo backup sleeves). On no condition should you reset or turn off the computer at the moment of a password check. This evasive action is likely to cause a damage to your database (consequences depend on the context of the check-up). SUPER-MEMO WORLD TAKES NO RESPONSIBILITY FOR ANY DAMAGE THAT MIGHT RESULT FROM THE USER BEING UNABLE TO PRODUCE THE PASSWORD AT THE CHECK-UP TIME. SuperMemo World appologizes for any inconvenience caused by application of password checking procedures.