Glossary

A

A-Factor (absolute factor; AF)

number associated with every element in a collection. A-Factor determines how much intervals increase in the learning process. The higher the A-Factor, the faster the intervals increase. You can see A-Factors in the Element data window. For items, A-Factors reflect item difficulty. For topics, A-Factors modulate topic priority (by affecting the frequency of exposition). The higher the A-Factor the easier the item. The most difficult items have A-Factors equal to 1.2. For items, A-Factor is defined as the quotient of the second optimum interval and the first optimum interval used in repetitions (read more: SuperMemo Algorithm SM-15). For topics and tasks, A-Factor is interpreted as the number by which the current interval should be multiplied to get the value of the next interval (as it was the case with E-Factors in early versions of SuperMemo).

acquisition rate

speed of learning, usually expressed in items memorized per year per minute. For example, if 20 minutes a day result in memorizing 10,000 items a year, the acquisition rate is 500 elements/year/min (10,000/20). In SuperMemo, the acquisition rate may vary from 30-1000 items/year/minute depending on the difficulty of the material, forgetting index and the stage of the process. Acquisition rate may decrease substantially in the first year only to stabilize asymptotically as years pass by. See formula

active recall

process in which pieces of information are actively retrieved from memory as opposed to passive review. For example, in passive review one might read that the highest divorce rate occurs in the 4th year of marriage. In active recall, you would have to retrieve this information from your memory: In which year of marriage are couples most likely to divorce? If you answer correctly, "4th year", then your memory stability will increase and so will the probability of recall in the future. In passive review, this increase is dramatically less pronounced. In literature, in various contexts, active recall may be referred to as practice at retrieval, active learning, learning based on a testing effect, retrieval practice, test-enhanced learning, active repetition, and more. Active recall may be or is not prompted with a specific question. It can then be called cued recall or free recall respectively.

ancestor (ascendant)

element in the Contents window that is located one or more levels above the element which makes its descendant. All the elements at the same indentation in the knowledge tree are children of their single parent (they are siblings to each other) whereas descendants are all the elements below the level of their ancestors.

See also:

• Building the knowledge tree
• Contents window
• Tree structure at Wikipedia

auto-postpone

option available with Learn : Postpone : Auto-postpone (from the main menu) that makes sure that excess repetitions are automatically postponed before the learning begins. Auto-postpone uses user-defined criteria in choosing elements that should be postponed. Most importantly, high priority material is protected from being postponed. Auto-postpone affects only the material that has been left outstanding from previous days. It does not postpone repetitions scheduled for the current day until the next day of learning.

auto-sort

option available with Learn : Sorting : Auto-sort repetitions (from the main menu) that makes sure that repetitions are sorted by priority (if it is checked)

B

back-retention

outdated term for consolidation in previous versions of SuperMemo

binary component (program component)

component that can hold, preview or execute a binary file stored in any format or written in any language (e.g. EXE, PDF, CHM, etc.). Binary components make it possible to extend the range of applications of SuperMemo into any imaginable area. To prevent the automatic execution of binary components, make sure that AutoPlay is unchecked on the element menu.

branch

in the Contents window, an element in the knowledge tree with all its descendants (incl. its children). All elements in a branch can be processed with subset operations. For example, if you would like to review your physics material before an exam, you could select the Physics branch in the Contents window, and click Learn (Ctrl+L) at the bottom of the window. You can add new branches with Add and Insert (Ins) in the Contents window. A branch that is built automatically by adding elements to a concept is called a concept group. All concepts are associated with their own branches/groups. By analogy to Windows, when a branch has no content and is use solely to hold other branches or elements, it may be called a folder.

See also:

• Building the knowledge tree
• Contents window
• Tree structure at Wikipedia

browser

window with a subset of elements. Browsers are mostly available from the View menu and provide a set of operations available from the browser menu. Read more: Browser window

burden

in the Statistics window, a number which estimates the average number of elements that have to be repeated daily (this statistic may overestimate the workload in collections where Postpone is used often):

burden = 1/I1+1/I2+ ... + 1/In

where:

• In - current interval of the n-th item

C

category

predecessor of concepts in older SuperMemos. No longer used as of SuperMemo 17. A category used to mean a named branch of the knowledge tree to which elements belonging to a given class of knowledge were added. For example, you might have kept categories such as: General Knowledge, Family, Internet, Job, etc. Now you can accomplish the same with concepts.

child

element in the Contents window that is located one level below the element which makes its parent. All the elements at the same indentation in the knowledge tree are children of their single parent (they are siblings to each other). Descendants are all the elements below the ancestor level. In incremental learning all knowledge extracts and clozes are created as children of their source.

See also:

• Building the knowledge tree
• Contents window
• Tree structure at Wikipedia

cloze deletion (cloze question; cloze item)

item in the fill-in-the-blanks form. In cloze deletions, the question is replaced by three dots ([...]) and moved to the answer field. For example:

Q: The highest literacy rate in Africa has been reached in [...](country)(2009)?
A: Seychelles (93%)

Cloze items can effectively remedy knowledge complexity. If you have items that persistently cause recall problems, try using cloze deletion. SuperMemo simplifies creating cloze deletions by providing options such as:

• Alt+Z keyboard shortcut,
• Cloze on the learnbar,
• Reading : Remember cloze on the component menu, or
• Remember cloze on the Read toolbar

Read more: Incremental reading

cloze for pictures

see occlusion test

collection

learning material used in SuperMemo. A collection is made of single pieces of knowledge called elements. The simplest elements have the form of a question and an answer. A collection of a given name is stored in a folder/directory that bears the same name, and all its important statistical and learning data are stored in a file with the extension kno. In older SuperMemos, collections of elements were called knowledge systems (SuperMemo 8) or databases (SuperMemo 7 and earlier)

component

a container for objects placed within the visible field of an element. Components can contain texts, pictures, videos, PDF files, etc. They can also have a form of a shape, spell-pad, or script. You can add new components with Edit : Add components on the main menu, or by using the Compose toolbar at the bottom of the element window.

computational spaced repetition

approach to spaced repetition in which optimum intervals are calculated with the help of a computer on the basis of the past record of learning. SuperMemo was the first computer program to implement rudimentary spaced repetition algorithm in 1987. Currently, SuperMemo employs the DSR model of memory to predict when items are likely to be forgotten.

concept

element associated with an idea. Multiple elements can form links to a concept. Links indicate that elements are associated with the idea represented by the concept. For example, if you learn about infections, you can define a concept flu virus. You can link this concept with many elements related to flu. Each time you want to learn about flu, you might begin from the concept flu virus. Concepts can also form concept groups, which are sets of elements located in the same portion of the knowledge tree. For example, you may create a concept called chemistry to group all your knowledge in the area of chemistry. See also: Concepts

concept group

set of elements added to the knowledge tree at a branch related to a given concept. For example, you can add all elements related to gravity to a concept group named Physics. Each time you add elements, they land in a specific location in the knowledge tree. That location is determined by the hook of the default concept group. Concept groups supersede categories known from earlier SuperMemos. The main difference between concept groups and categories is that the former are associated with concepts that play multiple roles in SuperMemo (starting with SuperMemo 17).

concept map

set of concepts and links between concepts. On the element menu, Concepts : Map displays the outgoing links for the current concept. Read more: Neural creativity

consolidation

retention of material repeated on a given day D as measured on the days of successive repetitions of individual elements. Both retention and consolidation are displayed in Toolkit : Calendar under Retention. Calendar displays these values as <retention> -> <consolidation>. If you are sleepy or tired on the day D, your retention will be poor, even if your memories are strong. In other words, measured retention is not an ideal reflection of memory retrievability. Consolidation is not measured on a single day. The measurements are gradually added up on days on which successive repetitions of the material repeated on the day D take place. Consequently, consolidation is less dependent on the variability in your recall readiness. At the same time, being tired or sleepy on the day D can affect the consolidation of the material. It is retention that is less dependent on the variability in your consolidation readiness.

Contents window

window that displays the hierarchical structure of knowledge in a collection. That structure is called the knowledge tree. It is available by clicking Contents in the element window. Read more: Contents

D

D-Factor

decline of O-Factors with successive repetitions can be approximated with a power curve that begins at O-Factor that equals A-Factor. D-Factors are no longer used in SuperMemo Algorithm. When a power regression is used to compute O-Factors on the basis of R-Factors for successive repetitions in a single A-Factor category, the decay constant of the resulting function is called a D-Factor. You can see individual D-Factors for all A-Factors in Toolkit : Statistics : Analysis : Graphs : D-Factor vs. A-Factor. The larger the D-Factor the faster the decline of O-Factors with each repetition. This means that large D-Factors imply more frequent repetitions in a given difficulty category. Naturally, over time, D-Factors tend to be lower for easy item categories (unless O-Factors hit their minimum value of 1.2 for very difficult items). D in D-Factor stands for decay (it is a decay constant of the negative power function)

delay

degree of delay of an element in the learning process. For items, it equals the current interval divided by the optimum interval for the current requested forgetting index. For topics, delay is heuristically scaled to provide for comparable degree of delay in terms of the damage inflicted on the learning process

descendant

element in the Contents window that is located one or more levels below the element which makes its ancestor. All the elements at the same indentation in the knowledge tree are children of their single parent (they are siblings to each other) whereas descendants are all the elements below the ancestor level.

See also:

• Building the knowledge tree
• Contents window
• Tree structure at Wikipedia

difficulty

estimation of item's difficulty in SuperMemo. The three main measures of difficulty:

1. current estimation of item's difficulty computed by Algorithm SM-18. This number ranges from 0 (for easy items) to 1 (for difficult items). It is displayed as Diff in the Element data window (see: details)
2. current estimation of item's absolute difficulty estimated by Algorithm SM-15. This number ranges from 1.2 (for difficult items) to 6.9 (for easy items). It is represented by A-Factor and displayed as A-Factor in the Element data window. This value reflects bounded stabilization for first review executed at optimum interval
3. heuristic measure of difficulty based on selected item data, and displayed in the Element data window as Old (this measure can be used to sort collections that are to be used by other users; starting with easy items first)

Dismiss

operation that removes an element from the learning process (compare with Forget and Remember). Dismiss leaves the element in the collection, however, it will never show up when using Learn

You can use Dismiss for the following purposes:

• Removing parenting articles from incremental reading while keeping them as reference
• Keeping searchable archive of knowledge that became outdated
• Converting topics to tasks (e.g. for prioritized entry into incremental reading), etc.

dismissed element

element that is not memorized, is ignored in the learning process, and is not kept in the pending queue. A dismissed element can be re-memorized with Learning : Remember (Ctrl+M) on the element menu. Priority of dismissed elements is ignored (at Dismiss it is set to 100%).

dragging mode (drag&size mode)

state of an element/component, in which the component(s) can easily be resized or dragged to a new location in the element window. The other two basic modes are: presentation mode (components are displayed like during repetitions) and editing mode (components can easily be edited, e.g. by typing in new texts, etc.). Components in dragging mode are usually darker than in the other two modes. To enter the dragging mode, Alt+click the component or the element. You can also press Ctrl+E twice. To drag a component in the dragging mode, press the mouse button over the component, and, without releasing it, move it to a new location within the element

DSR model of memory (3 component model of memory)

extension of the two-component model of memory at the neural network level. The two-component model of long-term memory says that a status of memory in a synapse can be described with variables: stability and retrievability. The DSR model, first used in Algorithm SM-17, adds the third variable called memory difficulty, which is an expression of the complexity of the synaptic pattern involved in storing a given memory. The more complex the net of connections involved in a memory, the harder it is to maintain the memory in the long term using spaced repetition. In SuperMemo, memory difficulty is expressed by item difficulty, i.e. a number that says how difficult it is to remember an item. There are more variables involved in storing memories that are not part of the DSR model. For example, both homeostatic and circadian components of sleep propensity have an impact on encoding and retrieving memories at the network level. For more see: DSR model

E

E-Factor (easiness factor; EF)

number related to the difficulty of a given element in early versions of SuperMemo, up to and including SuperMemo 7. In the earliest versions of SuperMemo (up to and including SuperMemo 3), new inter-repetition interval was determined by multiplying the old interval by E-Factor. A-Factors associated with topics and tasks are used in the same way as early E-Factors, i.e. to determine the value of the new interval by multiplying the A-Factor by the old interval. To better understand the role of A-Factors in new versions of SuperMemo, see: Incremental reading

editing mode

state of an element/component, in which it can easily be edited (e.g. by modifying the texts). All components except for HTML components can also be resized in the editing mode. The other two basic modes are: presentation mode (components look the same way like during repetitions) and dragging mode (components can easily be dragged with the mouse). The easiest way to distinguish between presentation and editing modes is that in the latter the components are enclosed by a sizable rectangle (except for the HTML component, which may instead be marked by a bluish status border). To enter the editing mode, press Ctrl+E

element

single page of information stored in SuperMemo (e.g. an article, a question-answer pair, etc.). All elements kept together are called a collection. Elements may have the form of topics (articles, extracts, summaries, etc.), items (testing material), concepts (general ideas used in semantic learning), or tasks (elements representing to-do jobs). A topic presents a larger part of the learning material, e.g. an article about the greenhouse effect. Items provide specific testing questions, e.g. How thick is the cerebral cortex? In the simplest case, topics have the form of a page of text while items are formulated as questions and answers (see: Topics vs. Items). Every element is represented in the Contents window as a single leaf or branch of the knowledge tree. The content of individual elements is displayed in the element window. Read more: Items, topics, concepts, and tasks

element browser

see browser

element subset

set of elements. Subsets are often saved into files with the default extension sub. You can create an element subset file by using the element browser (e.g. Subset : Save all or Select : Save selection in the browser menu). You can view an element subset file with View : Subset from the main menu. Read more: Using subsets

element window

window that displays a single element in a collection. In the default state this window displays buttons Contents, Search, History, etc. in its navigation bar along the top, as well as Learn, Add new, and the control bar at the bottom. Read more: Element window

extract

in incremental reading, a portion of text taken from a larger article and scheduled for review and processing as a separate topic (located in the knowledge tree as a child of the parenting article). In incremental learning, you can also extract portions of videos, images, sounds, etc. For example, you can extract a picture of Ghana when learning about Ghana and you already have a political map of Africa in your collection. Use Zoom if you want to retain Africa, extract Ghana, and save space by not generating a new extract. For more about extracts see: Extract

F

filespace

system of folders for holding multiple files in SuperMemo. Folders in the filespace are organized using a complex algorithm that ensures a minimum folder nesting and minimum access time. Files in the filespace are named 1.*, 2.*, 3.*, etc. Folders and subfolders are named [1], [2], [3], etc. Individual files are placed in individual slots. All files have their extensions retained (e.g. *.htm, *.jpg, *.mp3, *.mp4, etc.). The filespace is held in the subfolder called [ELEMENTS]. When using individual files, SuperMemo usually opens them directly from the filespace. Some files may need to be modified before they are displayed in SuperMemo. Those files are copied to a temporary folder first. The formula for converting filespace slots to folder names is complex and it is easier to search files by their slot names. For example, to find the file in the slot #279310, search [ELEMENTS] for 279310.*. That particular slot falls in the following subfolder: <collection name>\ELEMENTS\30\30\30\ as ELEMENTS\30\30\30\279310.*. Filespace is used in all SuperMemos starting with version 8. In SuperMemo 17, the filespace is limited to 4 folder levels and to 8 million files (exactly: 8,379,310).

filespace slot

single place for a single file allocated in the filespace in the [ELEMENTS] subfolder. For example, when allocating space for 300,000 files, SuperMemo may place a JPEG image file #288846 in the following slot ELEMENTS\1\2\3\4\288846.jpg

final drill

last (optional) stage of the learning process on a given day, in which only items that scored a grade less than Good (4) are repeated as long as they continue scoring less than Good. On the next day, final drill queue is again shifted to the end of the learning day. The final drill stage is not executed if Toolkit : Options : Learning : Skip final drill is checked. Unused final drill queue is deleted after 3 days, and can also be deleted manually. To delete the final drill queue, use Learn : Cut drills from the main menu

final drill queue

queue of elements that scored a grade below Good (4) in repetitions. This queue is used to review poorly scoring elements at the end of the learning day. For this queue to be generated, Toolkit : Options : Learning : Skip final drill must be unchecked. As you progress with incremental learning and build up a solid overload, the final drill plays lesser and lesser role. Advanced users with high learning overload should skip the final drill entirely.

flashcard (card)

term often used by users of SuperMemo to refer to an item or an element

Forget

operation that removes an element from the learning process and places it at the end of the pending queue (compare with Dismiss and Remember)

forgetting curve

SuperMemo: Toolkit : Statistics : Analysis : Forgetting Curves graphs for 20 repetition number categories multiplied by 20 A-Factor categories

describes the decline in the probability of recall over time. A forgetting curve graph in SuperMemo shows how fast the memorized information gets forgotten. A forgetting curve graph shows time on its X axis, and percent recall on its Y axis. It can be viewed with Toolkit : Statistics : Analysis : Forgetting Curves. SuperMemo collects 400 forgetting curves for 20 levels of knowledge difficulty and 20 levels of memory stability.

forgetting index

proportion of elements that are not remembered at repetitions (usually expressed as percentage). The forgetting index can be programmed to fall between 3% and 20%. This way, the speed vs. retention trade-off in learning can be controlled by the student. You can set the default forgetting index with Toolkit : Options : Learning : Forgetting index (default) and individual element forgetting indices with Forgetting index in the Element parameters dialog box. To understand the difference between the requested, measured, default, individual, expected and the estimated forgetting index, see: Forgetting index

free running sleep

sleep that is not artificially regulated. Free running sleep is a form of chronotherapy that can be used in curing a number of sleep disorders. Most of people in the industrial world cannot afford free-running sleep. Only a small proportion of population can sleep in a perfect 24 hours cycle and in synchrony with duties such as work and family. The most typical violation of the free-running sleep is the use of an alarm clock. Another violation is going to sleep too late in reference to one's natural bed-time hour. Going to sleep late in condition of little sleepiness does not violate the free-running sleep principles. Going to sleep too early (e.g. to force longer sleep before early arising) may also disturb the free-running sleep cycle. See also: Good sleep for good learning

G

grade

number from 0 to 5, which you provide when scoring your performance in recalling an item in learning. If you have an excellent recall of the item, you will score your review at 5 (grade 5 is called Great by default). If you perform poorly, you will score 1 (called Bad by default). You can even issue a Null grade of 0 from the keyboard if you perform dismally.

Note that the term first grade refers to the grade issued at first repetition (i.e. Repetitions=2), not at memorizing (i.e. Repetitions=1). The first issued grade does not affect the learning process except for determining which items enter the final drill

graphic cloze deletion

see occlusion test

graphic deletion

see occlusion test

H

hierarchy

see knowledge tree (outdated term that might be found in some older documentation files)

hierarchy node/branch/leaf

see node, branch, or leaf (hierarchy is an outdated term that might be found in some older documentation files)

hint

message that makes it easy to understand the use of a given interface element in SuperMemo. To view hints pause the mouse pointer over a button or menu item. Hints are displayed on the status bar along the bottom of the screen (you can show the status bar with Window : Status bar from the main menu).

You can turn hints on/off by:

1. checking Window : Hints, or
2. double-clicking the hint panel on the status bar

hook node

place that determines where new elements are added in the Contents window (e.g. with Add new). A hook is a knowledge tree node to which new elements are added (as children). The hook node is either the same as the root node, or is a child of the root node, or is a descendant of the root node. Once the children limit of a hook is passed, a new hook is chosen. The newly chosen hook is always a descendant of the root of the current concept group. If SuperMemo runs out of branches that might become new hooks, it will restructure the concept branch and start a new level of branches under the root node with new space for new hooks and new elements.

hyperlink

connection established between a component and an element different than the element owning the component. Hyperlinks make it possible to navigate in knowledge hyperspace by clicking individual components or HTML links. You can set a hyperlink by using Links : Hyperlink on the component menu or Insert HTML Link : Element on the HTML component menu

I

incremental audio

learning technique pioneered by SuperMemo, in which you can listen to hundreds or thousands of audio files in parallel with substantial benefit to the speed of learning, and the quality of acquired knowledge. Incremental audio is analogous to incremental video and uses the same tools. The only difference is that it uses audio files instead of video files. YouTube-based incremental video can also be used as incremental audio (e.g. for learning languages, learning to sing, etc.). See: Incremental learning

incremental learning

set of learning techniques such as incremental reading, incremental video, visual learning, incremental audio, etc. Incremental learning easily ensures 95% recall of top-priority learning material for lifetime. For more see: Incremental learning

incremental reading

reading technique pioneered by SuperMemo, in which you can read hundreds or thousands articles at the same time with substantial benefit to the speed of learning, and the quality of acquired knowledge. See: Incremental reading

incremental video

technique of watching and/or learning multiple YouTube and local videos at the same time. Incremental video is to video as incremental reading is to texts. See: Incremental video

incremental writing

process that combines incremental reading, creative writing, and elaborative writing, in which a text of an article or book is created incrementally. For a more detailed description, read: Incremental writing

interval (inter-repetition interval)

period of time between successive repetitions. In incremental reading, intervals are also periods between the review of topics. Initially, repetitions occur in intervals ranging from 1-15 days; however, with time, intervals can reach well into thousands (which corresponds to decades). Intervals in SuperMemo are measured in days. Occasionally they may be displayed as seconds (if they are very short), or years (if are very long), or with the use of other units

item

simple element, which often has a form of a question and an answer. Items ensure long-term recall of information. Items take part in active learning, which is opposed to passive review or reading (done with the help of topics). Items are often created by means of cloze deletion. See also: Topics vs items

K

knowledge tree

the tree structure in which particular elements of a SuperMemo collection are organized. The knowledge tree is presented in the Contents window. Particular nodes of the tree can hold up to a thousand children each, but for performance reasons it is recommended not to keep more than a hundred children elements in a single node. Some authors use the term knowledge hierarchy to refer to the knowledge tree.

See also:

• Building the knowledge tree
• Contents window
• Tree structure at Wikipedia

L

lapse

memory lapse is a failure to recall an item. When we say "number of lapses", we mean the number of times an individual item has been forgotten, i.e. scored less than Pass (3) in repetitions

layout

position and size of windows and toolbars in SuperMemo. Layouts can be changed with Window : Layout menu. In particular, use Ctrl+Shift+F5 to save your favorite layout. 9 first-defined layouts are accessible through the Window menu. There is no need to save layouts for Contents and browser layouts as they are saved automatically. Find out more: Layouts.

Compare: Template

leaf

one childless element in the knowledge tree presented in the Contents window (an item, topic, concept or task in the collection). You can add new leaves with Add and Insert (Ins) at the bottom of the Contents window, or with options on the Edit submenu of the main menu.

See also:

• Building the knowledge tree
• Contents window
• Tree structure at Wikipedia

leech

particularly difficult item that causes problems in learning. The definition of a leech is specified by means of the Element filter dialog box used in View : Other : Leeches (Shift+F3) from the main menu. A semi-leech is an item that is not a leech but will become one once it is forgotten. See: Leeches

Leitner system

rudimentary method for spaced repetition used in a number of flashcard applications. The method has been proposed by Sebastian Leitner in the 1970s and can be considered a forerunner of SuperMemo

level

degree of complexity of SuperMemo exposed to users of different familiarity with the program. There are four levels with increasing complexity available from File : Level. These are: Beginner, Basic, Middle and Professional. Find out more: Levels

link

connection between components, elements, web pages, etc. There are five main kinds of links in SuperMemo:

1. element-to-element link: link that connects elements, e.g. as created with Concepts : Link contents on the element menu
2. element-to-concept link: link that connects an element to a concept, e.g. as created with Concepts : Link concept on the element menu
3. registry link: link from a component to an object in the corresponding registry such as a picture in the image registry, e.g. as created when pasting a picture to a component
4. component hyperlink: link from a component to another element, e.g. as created with Links : Hyperlink on the component menu
5. HTML hyperlink: link from an HTML text to a webpage, or to an element in SuperMemo

M

matrix smoothing

mathematical procedure that converts a matrix of numbers into a "smoother version" (e.g. by averaging the neighboring entries). For example, if the row of the matrix is 1, 2, 3, 4, 666, 6, 7, 8, smoothing might convert it to 1, 2, 3, 6, 99, 9, 8. For more, see: Smoothing

measured forgetting index

proportion of elements that are not remembered at repetitions. This is the forgetting index as actually measured during repetitions. See: Forgetting index

Memorize

see Remember

memorized element

element that takes part in the learning process (i.e. is repeated in intervals suggested by SuperMemo). A memorized element can be made a pending element with Forget (e.g. Learning : Forget in the element menu). It can also be made a dismissed element with Dismiss (e.g. Learning : Dimiss (Ctrl+D) from the element menu).

memory lapse

see lapse

mid-interval repetition

repetition that takes place ahead of its planned time. Newer SuperMemos make it possible to execute repetitions ahead of time, e.g. before an exam. The algorithm takes the necessary correction for the spacing effect.

minimum information principle

principle of effective learning which says that simple elements formulated for active recall bring much better learning results than complex elements even though one complex element may be equivalent to a large number of simpler elements.

Read more:

• 20 rules of formulating knowledge
• Incremental learning and the minimum information principle

mnemonic hyperspace

extension of Tony Buzan's concept of mind maps by application of hyperlinks between the mind map components and mind map editability. Mind maps are considered an excellent form of representing knowledge for the purpose of learning. SuperMemo makes it possible to create simple mind maps that contain multimedia objects. It can also use mind maps generated with Mind Manager (via OLE in-place activation)

N

neural creativity

creativity induced with the help of neural review. In neural review, SuperMemo successively feeds the user with knowledge associated with a selected topic or a chosen subset of topics. When a network of conceptual links connects areas of knowledge, they form semantic space that can be explored in a neural manner that helps forming new associations and generate ideas. This process if conducive to research, invention, problem solving, etc. Metaphorically speaking, neural creativity helps emulate a thinking process in a human brain extended by knowledge stored in SuperMemo. Unlike it is the case with semantic review in SuperMemo, or simply googling for knowledge, the whole process of associating pieces of knowledge is automated. To give it a try, pick a subject and Go neural. For details see: Neural creativity

neural queue

queue of elements scheduled for neural review. The sequence of elements in the queue is determined by spreading activation from a single element, subset of elements, or a registry member (e.g. a picture)

neural review

subset review that follows meaningful connections between pieces of knowledge. It takes inspiration from how the brain follows associated ideas. Neural review automates the review of knowledge associated with selected concepts or ideas, e.g. in problem solving, creative writing, etc. For example, in a neural review of dogs, there is a good chance of learning about puppies or puddles, but there is also a chance of learning about cats and somewhat lesser chance of learning about cars.

Neural review follows the spreading of activation in a network of inter-element connections. Elements are linked with concepts, other elements, or form connections with their neighbors in the knowledge tree. All those links and connections receive weights based on their importance and the priority of connected elements. Neural review follows those connections and is initiated by executing Learn : Go neural (Ctrl+F2). Neural review carries a degree of randomness. As such it can be very helpful in associating ideas in neural creativity. Find out more: Neural creativity

node

one element in the knowledge tree presented in the Contents window (a single item, topic, concept or task in the collection). You can add new nodes with Add and Insert (Ins) at the bottom of the Contents window, or with options on the Edit submenu of the main menu. If a node does not have children, it is called a leaf. If the node has children, it is called a branch. If a node is used as a container for other nodes, it may also be called a folder. The contents menu provides some operations on entire branches such as: Learn, Review all, Statistics, etc.

See also:

• Building the knowledge tree
• Contents window
• Tree structure at Wikipedia

O

O-Factor (optimum factor; OF)

number which tells you how much intervals should increase to reach recall of 90%. O-Factors differ for different levels of memory stability and different item difficulty. They are normalized for the forgetting index of 10%. For the first repetition, i.e. there was no prior interval, O-Factor is assumed to be the same as the first interval (as if the prior interval was 1 day). For the first repetition, different O-Factors are computed for a different number of memory lapses. For example: if the O-Factor is 2.5, and the prior interval was 20 days, at repetition time, the new interval should be set to 2.5*20 days, i.e. 50 days.

OF matrix

matrix of O-Factors for different levels of difficulty (expressed as A-Factor) and stability (expressed as repetition category). The OF matrix is used by SuperMemo Algorithms SM-5 through SM-15 in computing optimum intervals in spaced repetition. In the newest algorithms, the role of the 2-dimensional OF matrix is played by the 3-dimensional SInc[] matrix (with the added dimension of retrievability)

occlusion test

item that uses a picture whose part is deleted or occluded. During a repetition cycle, the occluded part is exposed at answer time. Occlusion tests can be used to learn geography, anatomy, mind maps, technical graphs, and other forms of knowledge that is presented in graphic form. Also called: graphic deletion test or cloze for pictures.

See also:

• Watch: Learning Human Anatomy with SuperMemo
• Read: Creating occlusion tests

optimum interval

in SuperMemo, an inter-repetition interval that is likely to result in the probability of forgetting equal to the requested forgetting index. In other words, the optimum interval is the best possible interval that will ensure the speed of learning chosen by the user

ordinal number

number assigned to each element indicating the designed ordinal position of the element in the collection. Ordinals are important when writing commercial collections. Instead of ordinals, users should use priorities in their own collections. Some collections are sorted by their ordinal numbers and this is the order in which they should usually be learned. Ordinal number comes with a collection, while the priority is set for each element by the user using his or her own criteria. By default, ordinal number is set to the number of elements in the current collection plus 10,000. This way, the first element you add will have the ordinal 10,002, second element will get 10,003, etc. Ordinal numbers are useful in sorting commercial collections or their subsets

outstanding element

element that is awaiting repetition (e.g. to satisfy the criteria of optimum interval). Each element in the learning process has its next repetition date determined by SuperMemo or by the user. On that scheduled review date and later on, the element is considered outstanding

outstanding material

all outstanding elements, i.e. elements whose next repetition date is less or equal today's date

outstanding queue

queue of elements scheduled for today's review. If you use Add to outstanding (available on the subset processing menu), not all elements in the outstanding queue are actually outstanding

overload

a situation when the student has more outstanding elements to review than they can handle. Few users can sustain more than 200 item repetitions per day (let alone thousands of topic in a heavily overloaded incremental reading process). When the Outstanding parameter (in the Statistics window) starts going above that number, overload is likely. Overload can best be handled with Auto-postpone. However, one-time big loads can be resolved efficiently with Postpone (delaying all elements) or Mercy (spreading all review in time)

P

parent

element right above a level in reference to a given element in the knowledge tree in the Contents window. Use Ctrl+Up to see the parent of the currently displayed element. Root of the tree is the only element that has no parent. All children of element X have element X for parent.

See also:

• Building the knowledge tree
• Contents window
• Tree structure at Wikipedia

passive review

process in which pieces of information are read passively without asking questions (as opposed to active recall). For example, in passive review, one might read that 50% of marriages in the US divorce. In active recall, you would have to retrieve this information from your memory: What proportion of marriages in the US divorce? Active recall is far more effective in learning than passive review. In incremental reading, you always begin with a passive review of the learning material, and then gradually (incrementally) convert its portions to questions and answers suitable for active recall

pending element

element that awaits memorization in the pending queue. To move an element to the end of the pending queue, you can do choose Learning : Remember and Learning : Forget from the element menu (in that sequence)

pending queue

queue of elements that are waiting to be memorized. Option Remember removes the current element from the pending queue while Forget adds a memorized element back to the end of the pending queue. Pending queue determines the order of learning new elements. You can forget all the elements in a given collection by choosing File : Tools : Reset collection. You can also make a subset of elements pending by using options of the browser menu

post-lapse stability (PLS)

stability of an item after a repetition that scored a failing grade. Post-lapse stability can be interpreted as optimum interval after forgetting and rarely goes beyond mere days. On occasion, an item might be forgotten by "accident" and still remembered years after the lapse, however, statistically this is an insignificant phenomenon as visualized by Toolkit : Statistics : Analysis : Graphs : First interval and Toolkit : Memory : 4D Graphs : First interval. See more: SuperMemo Algorithm

presentation mode (display mode)

state of an element/component, in which it looks as when seen by the user during browsing or learning the collection. The other two basic modes are: editing mode (components are ready for editing, e.g. deleting texts, etc.) and dragging mode (components can easily be dragged with the mouse)

primary storage

place on the user's hard disk where the current collection is located. Compare: secondary storage

priority

number that reflects the importance of an element in SuperMemo. Most important elements have the priority 0%, while the least important elements have the priority 100%. Priority can be changed with Priority : Modify (Alt+P) on the element menu. For more, see: Priority queue. Note that the term Priority is also used in reference to tasks where it represents Value/Time

priority bias

false sense that all learning subjects are very important and equally important. Only months of prioritization training can help overcome this bias/sensation. A good user of SuperMemo will avail of the full spectrum of priorities from 0% to 100%, and will apply them naturally and instinctively

priority queue

queue of elements ordered by their priority. Use priority queue to always start from learning the material that is most important for you. Do not despair if you do not finish learning for a day. With priority queue, you know you did your best and only lower priority material was left behind. Remember to use Auto-sort and Auto-postpone to make the most of the priority queue. See: Priority queue

Q

Q&A format

format of text files with questions and answers that can be imported to various versions of SuperMemo. See: File : Import : Q&A text

R

R-Factor (retention factor; RF)

the number which says at which U-Factor (i.e. the measure of interval increase) the measured forgetting index (i.e. the measure of forgetting) is approximated to be 10%. This number is unique for different item difficulties and for different repetition categories (i.e. the measure of memory stability). R-Factors can be seen as a vertical green line on forgetting curve graphs in SuperMemo

R-Metric (Recall Metric)

SuperMemo: Algorithm SM-18 performance metric

absolute measure of performance of two spaced repetition algorithms based on their ability to predict recall before a grade is scored. As of SuperMemo 18, R-Metric is used solely to compare Algorithm SM-15 (known from SuperMemo 16) and the new Algorithm SM-18. It is shown as percentage in Statistics and Toolkit : Statistics : Analysis : Use : Efficiency : R-Metric. R-Metric is a difference between the performance of the two algorithms: R-Metric=LSRM(Alg-15)-LSRM(Alg-18), where LSRM is the least squares predicted recall measure for a given algorithm. R-Metric greater than zero shows superiority of Algorithm SM-18. R-Metric less than zero indicates underperformance of the new algorithm. LSRM is a square root of the average of squared absolute differences in recall predictions: abs(Recall-PredictedRecall), where Recall is 0 for failing grades and Recall is 1 for passing grades. PredictedRecall is a prediction issued by the algorithm before the repetition. In Algorithm SM-18, the prediction is a weighted average of the value taken from the Recall[] matrix, and R (retrievability) computed from S (stability) and the used interval. The weight used is based on prior repetition cases which inform of the significance of the Recall[] matrix prediction (the prediction becomes more meaningful with more prior repetition data)

RF matrix

a matrix that holds all R-Factor values, both measured and theoretical (i.e. those that could not be measured yet). R-Factors are arranged in rows of repetition category (i.e. the measure of memory stability) and columns of A-Factor (i.e. the measure of item difficulty). You can see your RF matrix with Toolkit : Statistics : Analysis : Matrices : RF matrix

read-point

recurring selection made by SuperMemo in an article. Each time you come back to the article, the selection will be restored to show you the fragment that should be processed next. To set your own read-point, use Ctrl+F7. To clear it, use Ctrl+Shift+F7

reading list

prioritized list of articles imported to SuperMemo and scheduled for reading in the sequence of priority. A reading list is a form of a tasklist in which each task is an article to read. In new SuperMemos, reading lists have been largely made obsolete by incremental reading

Recall matrix (Recall[])

the matrix that shows how recall depends on retrievability (R). In an ideal case, recall and retrievability should be the same, however, there are factors that make it hard to achieve. For example, difficult items produce non-exponential forgetting curves, SuperMemo finds it hard to sort items by difficulty, difficulty is fluid, etc. The Recall matrix is then used to predict recall on the basis of retrievability for items with varying difficulty and stability. The matrix can be used to produce the approximation of the recall function which helps estimating recall even if no recall data is available. The Recall matrix can be compared to the forgetting curve matrix. In older algorithms, forgetting curve matrix was used with time measured in U-Factor unit. The Recall matrix measures time in retrievability units. You can inspect the Recall matrix using: Toolkit : Memory : 4D Graphs : Recall

recognition

ability to passively recognize a memory. For example, you may not remember the answer to What is the capital of Sierra Leone?, however, you may recognize Freetown as the capital once you hear the name. Or you can say I knew that once someone tells you that Freetown is the capital of Sierra Leone

registry

sorted set of named objects (called registry members) that are used in creating elements in SuperMemo. Objects stored in registries may have the form of text, HTML file, image, sound, video, font, executable program, DLL, PDF file, etc. You can link particular objects with components in elements by means of Links : Registry member (Shift+Ctrl+K) on the component menu. Upon choosing the appropriate object name in the registry, choose Accept and the object will appear within the selected component. Important! Registries in SuperMemo have nothing to do with the Windows registry (SuperMemo have been using registries before the release of Windows 95)! Read more: Registries

registry index

ordinal number of a given member in the sorted order of registry members in a given registry. For example, the lexicon registry might start with aardvark at the registry index 1 and end with Zoloft at the registry index 23293. Registry index can be inspected at the bottom of the registry window. See: registry position

registry member

one of the objects stored in a registry (e.g. a picture in the image registry or a word in the lexicon registry)

registry object

file or a record associated with a registry member. For example, 12363.jpg may hold the picture of the president for the member of the image registry named Barack Obama. Arial 16p may be a member in the font registry that holds the record of properties of a font used in the collection

registry position

number corresponding with the physical position of a given registry member in a registry. This number can be seen at the bottom of the registry window. See: registry index

registry subset

set of registry members saved in a file. Registry subsets are mostly used in search operations. For example, in a registry window, you could use Search : Find texts (Ctrl+S) on the registry menu to find all texts containing the word virus. Those will be displayed as a registry subset in the text registry window. You could use Ctrl+S again to search for the word AIDS and generate a narrower subset that would contain both the words virus and AIDS. Finally, you can use View : Browse selected (Ctrl+Shift+B) on the registry menu to display a browser with all elements that use the members of the selected registry subset

registry window

any of the several windows that display the contents of individual registries. These windows are available on the Search menu (e.g. Search : Texts, Search : Images, Search : Other registries : Font, etc.). Read more: Registry window

Remember

operation that introduces an element to repetitions. Compare: Dismiss and Forget

remembered element

element that can effectively be retrieved from memory. Remembered elements should not be confused with memorized elements, i.e. elements that take part in the learning process. In a well-executed learning process, probability of recall (retrievability) for memorized elements should be above 100% minus the forgetting index (or exactly that number at repetition time). By definition, a remembered element has the probability of recall verging on certainty

repetition

act in which a given item is rehearsed by going through the following stages:

1. show the question (or the stimulus)
2. respond to the question (or react to the stimulus)
3. compare the response with the correct answer and grade yourself (or be graded by the program for your reaction to the stimulus)

Note that all repetitions take place on a date selected by SuperMemo by using the spaced repetition algorithm (see: SuperMemo Algorithm)

repetition category

number that represents the repetition number corresponding to a given interval in the theoretical learning process in which the current status of the OF matrix is used. If you make many repetitions in short intervals (e.g. as a result of mid-interval review), you may have a large repetition number for a short interval. Such repetition number cannot be used in the SuperMemo Algorithm. Instead, SuperMemo will look at the optimum interval and see how many times an item would need to be repeated to reach that interval. That number is called the repetition category

repetition spacing

see spaced repetition

requested forgetting index

proportion of items that you accept not to be remembered at repetition. This is your planned/desired value of the forgetting index. Your actual measured forgetting index depends on systematic work and the formulation of knowledge in learning. See: Forgetting index

retention

proportion of knowledge retained in memory at any given time. Retention is greater than 100% minus the forgetting index. The forgetting index refers to the probability of forgetting at the moment of a repetition while retention refers to the average recall probability between the last and the next repetition. For an exact formula linking the forgetting index and the retention see: Theoretical aspects of SuperMemo. Retention equals 100% minus the forgetting index only when measured on items repeated on a single day (e.g. as displayed in Toolkit : Calendar)

retrievability

one of the two variables describing memory traces in learning (the other is stability). Retrievability is most often expressed as the probability of recall. Retrievability is subject to negatively exponential decline whose speed depends on memory stability. Stability is often expressed as the optimum interval for the forgetting index equal to 10%. This term was proposed by Wozniak, Gorzelanczyk, and Murakowski (1995) to help differentiate memory stability from the imprecise and all-encompassing term memory strength. See: Two components of long-term memory

review

process in which the learning material is reviewed. In SuperMemo, the review may be active (items) or passive (topics). Usually, passive review means reading, while active review means recalling of information from memory with grading (repetitions). See also: Subset review

root node

root is a branch in the Contents window that keeps all elements of a given concept group. For example, a node named Chemistry might be a root node of all elements related to chemistry.

The root node determines how deep SuperMemo can interfere into the structure of the tree when working with a concept group. All elements under the root are subject to restructuring. All elements above the root are protected from changes.

The root is a node that keeps the hook that parents all newly added elements in the concept group.

The root of the whole collection is the root of the knowledge tree. This master root has a physical position of 1 in the collection (Element #1)

S

script

a list of commands available via the script component to manipulate other components in a given element (i.e. to display components, change their position, animate them, use them in learning tests, etc.). This way you can create interactive presentations, animation, various tests, etc.

search&review

in SuperMemo, a review of a subset of elements that contain a given search phrase. For example, before an exam in microbiology, a student may wish to review all his knowledge of viruses using the following method:

1. search for all elements containing the phrase virus (e.g. with Ctrl+F)
2. review of all those elements (e.g. with Ctrl+Shift+L)

The review may include all searched for elements (e.g. Learning : Review all in the browser with Ctrl+Shift+L), or only the outstanding elements (e.g. Learning : Learn in the browser with Ctrl+L).

Before you execute the review, you can randomize the review material (Ctrl+Shift+F11), sort it by priority, by recency, by interval, etc.

Compare: Neural review

secondary storage

place where multimedia files of a collection are stored. For example, in very large collections, multimedia files can be stored on CD-ROM (secondary storage), while all the frequently accessed or modified files are stored on the user's hard disk (primary storage). When you first create your collection, the primary storage equals the secondary storage. Secondary storage can be inspected or modified with Toolkit : Options : Access : Secondary storage

sibling

element in the Contents window that is located on the same level of the knowledge tree as the element in question.

See also:

• Building the knowledge tree
• Contents window
• Tree structure at Wikipedia

spaced repetition (repetition spacing; expanded rehearsal)

technique of optimizing the learning process by computing optimum intervals that should separate repetitions of individual pieces of knowledge. Those pieces of knowledge are called elements in SuperMemo. SuperMemo pioneered the use of optimization methods in spaced repetition (first implemented in 1987) and has gained acclaim through its impact on the effectiveness of learning. See also: computational spaced repetition

spacing effect

property of memory which makes us remember things better if they are repeated in a spaced manner rather than condensed or cumulative manner (e.g. 7 x 10 min. learning daily as opposed to 70 min. on the weekend). Spacing effect is the basis of spaced repetition in learning. SuperMemo uses the spacing effect by trying to keep intervals as long as possible before forgetting increases beyond a selected threshold called the forgetting index

splitmark

boundary mark by which a large article will be split into smaller portions for faster loading and processing in SuperMemo. It can be a short text (e.g. ###SplitMark### or any custom string), an HTML tag (e.g. HR (also called splitline), H1 through H6, DIV) or a Wiki headline. You can quickly insert a splitmark by choosing Reading : Split : Insert splitline (Shift+Alt+H) from the HTML component menu

spreading activation

method of activating nodes in a neural network, semantic network, directed graph, etc. In SuperMemo, the graph is formed by semantic connections of the knowledge tree, concept links and inter-element links. Spreading activation is used to sequence elements for neural review. This is a new form of semantic review in SuperMemo, where new elements are fed neurally for review. This approach is useful in enhancing creativity and problem solving. To get a taste of neural review choose Learn : Go neural (Ctrl+F2) from the main menu. For more, see Wikipedia: Spreading activation

stability

one of the two variables describing memory traces in learning (the other is retrievability). Stability determines the speed of the negatively exponential decline of memory traces. Stability can be expressed, for example, by the optimum interval in spaced repetition (often normalized for the forgetting index equal to 10%). This term was proposed by Wozniak, Gorzelanczyk, and Murakowski (1995) to help differentiate from the imprecise and all-encompassing term memory strength (see: Two components of long-term memory). At each repetition, stability increases along the Stabilization function.

stability increase matrix (SInc matrix; SInc[])

the matrix that represents the function of interval/stability increase in SuperMemo algorithm. For any item at any time, the increase in stability depends on:

1. current stability,
2. current retrievability, and
3. current item difficulty.

This function can be inspected using: Toolkit : Memory : 4D Graphs : Stability in SuperMemo. The SInc[] matrix can be considered an extension of the OF matrix (from older SuperMemos) into the dimension of retrievability. This means that the new SuperMemo algorithm can handle any repetition at any time, including freshly reviewed items, as well as long-neglected items. Retrievability is the dimension that reflects the time that has elapsed since the last repetition relative to the optimum interval

stabilization curve

shows the dependence of stabilization on retrievability.

The formula for the stabilization curve:

SInc=SIncMax*e-Gain*R

where:

• SInc - stabilization (aka stability increase) is the ratio of memory stability after and before the repetition
• SIncMax - maximum possible increase in memory stability. SIncMax depends on current stability and memory complexity (i.e. item difficulty). Maximum stabilization can be achieved at the verge of forgetting
• Gain - a constant that expresses the gain in stabilization with passing time (gain is a measure of the spacing effect)
• R - retrievability, i.e. the expected probability of recall at a given point in time. It is determined by the forgetting curve

startup interval (SI)

common first interval for all new items determined by the forgetting curve for new heterogeneous material. See more: SuperMemo Algorithm

startup stability (SS)

the estimate of stability after the first review determined by the best match to grades obtained in later repetitions (that estimate may change with the arrival of new data). See more: SuperMemo Algorithm

subset

1. element subset: set of elements saved in a file, or
2. registry subset: set of registry members saved in a file.

You can create an element subset by using the element browser (e.g. Subset : Save all or Select : Save selection on the browser menu). You can view an element subset with View : Subset from the main menu.

Registry subsets are mostly used in search and sort operations. For example, you could use Ctrl+S to find all texts containing the word virus. Those will be displayed as a registry subset in the text registry window. You could use Ctrl+S again to search for the word AIDS and generate a narrower subset that would contain both words virus and AIDS

subset review

review of elements in a subset. Subset review may, for example, review all items related to geology, or all elements related to Nuclear power in the United States. Subset review may be based on a variety of subsets. For example, it may have a form of search&review, neural review, branch review, and more

SuperMemo

1. method of fast learning based on computing optimum intervals between repetitions
2. computer program implementing the SuperMemo method

See: Introduction to SuperMemo

T

task

element that represents a to-do task. For example, a task may have a form of an article that is waiting for reading on a prioritized tasklist called a reading list.

See also:

• Break free from work overload - rationale behind tasklists,
• Tasklists - how to use tasklists, and
• Incremental learning - alternative to tasklists in creative jobs

tasklist

form of to-do list used in SuperMemo. Tasklists are made of a set of tasks (each corresponds with a single element in SuperMemo). Tasks in a tasklist are sorted by their priority. Task priorities in SuperMemo are determined by the Value/Time ratio. See: Tasklist manager

template

a reusable definition of an element's appearance. It may define the type and properties of individual components such as size, color, font, image files, alignment, etc. All templates used in a given collection are kept in the template registry available with Search : Templates from the main menu. Read more: Using templates

topic

element that presents a synthetic overview of knowledge (e.g. an article to read). Knowledge stored in topics is gradually converted into items (e.g. in the process of incremental reading). Optimally, topics introduce you to the learned knowledge by providing a synthetic overview. You later keep the knowledge in your memory by only reviewing items. In a well-structured collection, topics will always be parents to items derived from their contents. Each time a student loses the sense of context during repetition, he or she can press Ctrl+Up to view the parent of the current element. This way a quick review of the synthetic material in the topic is possible. See also: Topics vs items

U

U-Factor

number associated with each memorized element. It equals to the ratio of the current interval and the previously used interval. If the element has not been repeated yet, i.e. the previous interval is not defined, U-Factor equals the first interval. The greater the U-Factor the greater the increase of the interval between repetitions. For items, U-Factors are determined by SuperMemo Algorithm. For topics, U-Factors are determined by A-Factors. For topics that have already been reviewed at least once, that have not been subject to postpones, U-Factors equal A-Factors. U in U-Factor stands for used interval increase

V

visual browser

SuperMemo browser in a mode that shows pictorial snapshots of individual elements. For pictures, see: Visual browser

visual learning

equivalent of incremental reading applied to pictures as well as texts stored as pictures. For details see: Visual learning

Z

zoom&trim mode

state in which pictures in SuperMemo can be processed with mouse operations that include zooming, trimming, moving the zoom, etc. To enter this mode, Alt+click the picture. The picture will be surrounded with a thin lime border. To test the mode, 'middle-click the picture to zoom in