SuperMemo 2004 beta-testing program-May 2004 - Jan 2005

SuperMemo 2004 beta-testing program
May 2004 - Jan 2005


  1. Please create a back up of all your collections! Do it each time you download a new version of SuperMemo!
  2. Collection upgrade to SuperMemo 2004 is irreversible!
  3. Please note that all information related to SuperMemo 2004 and this beta-testing program is confidential until the release date. Please do not share it with others. In particular, please withdraw from public comment, positive or negative, on the Internet or other media
  4. For all communication related to beta-testing, please use the following address: beta(AT)supermemo(.)com. Do not change the default bug reporting address from bug@ to beta@. These two are processed differently. Do not use other support addresses as this will slow down bug report processing! Please use only English in your messages 
  5. Send screenshots and HTML examples as attachments. Do not compress files less than 2 MB in size. Send collections as ZIP attachments. For files above 10MB, provide a URL for download. Do not submit large files to non-beta addresses
  6. Upon upgrade from earlier version, you must not abort File : Repair collection. Only this procedure will fill out new records with default data. Ignore error messages in the upgrade report
  7. Please send your requests for future versions too. Those cannot be considered in the course of beta-testing but will scrupulously be stored on tasklists built for future releases
  8. Please send comments about SuperMemo help
  9. Please send comments about ABC.kno


  1. To update SuperMemo 2004 re-run your setup program
  2. Help for SuperMemo 2004
  3. What's new in SuperMemo 2004
  4. FAQs
  5. Bugs fixed during beta-testing (below)
  6. Documentation texts to review in order of priority:
    1. 25 Steps to SuperMemo Mastery
    2. Incremental reading: primary introduction to incremental reading for new users (FAQ: Incremental reading)
    3. Using templates and categories (FAQ: Templates and categories)
    4. Postpone (FAQ: Learning overload)
    5. Backup (FAQ: Backup)
    6. Building knowledge tree: knowledge tree for beginners (FAQ: Knowledge tree)
    7. Element menu
    8. Component menu

Bugs fixed during beta-testing

Build Bugs fixed New features
Beta B
  1. Errors when importing from XML with repetitions executed "in the future" (e.g. when using Midnight clock shift)
  2. Cursor would not respond to Left arrow in the tasklist manager
  3. UTF-8 search not possible beyond ASCII (e.g. searching for "Ptain")
  4. Searching for whole word "pp" would also find "ppp"
  5. Author: would receive precedence over From: in incremental e-mail processing
  6. Restoring predefined templates might leave the template registry open
  7. Esc code errors when deleting a template while selecting a category template
  8. Ctrl+click would not work with HTML components
  9. Compose toolbar would show no caption upon docking
  10. Canceling applied template in Element Parameters might apply an empty template
  11. Invalid e-mail address in Help : E-Mail : Support
  1. Customizable start-up picture (Help : About)
  2. Default font replaced with Question font and Answer font for templates where font is not specified
Beta C
  1. [ and ] keys would delete texts in the HTML component
  2. Wrong branch overflow handling in XML import
  3. Setting the default font in the plain text component would not recognize the answer attribute
  4. Wrong setting of U-Factors in XML import to a new collection
  5. Registry Browse all button would only display the list of elements
  6. Import mail dialog would not fit 800x600 screens
  7. Wrong tab order in couple of dialogs (e.g. Repair Options)
Beta D
  1. Wrong branch overflow handling in Q&A import
  2. Files with the extension html would list twice in the file browser
  3. Show Reference would omit line breaks
  4. Program vs. documentation naming discrepancies for E-mail : Element and E-mail : Attachment
  5. Edit Font on default font in plain text components might attempt to modify the registry font
  6. XML import would wrongly interpret missing element IDs
  7. Download images would be visible on non-HTML component menus
  8. Saving contents tree in a text file would erroneously read repetition history 
Beta E
  1. Template : Flags menu would be unavailable in the template registry
  2. Download images would ignore PNG files
  3. Skip final drill would be set to true when starting a new collection
  4. Ctrl+W shortcut would not work on the beginner level
  5. mpeg files would not be converted to mpg on import to the registry
  6. Video : Continuous would fail to respect extract boundaries
  7. Temporary image files would be deleted without confirmation when moving to another element
  8. Ctrl+F10 would be disabled for replay in sound and video registries
  9. Esc on multiple downloads might fail to free resources (progress box, download thread, etc.)
  10. "Gray" columns in Plan might become non-gray in certain color settings in Windows
  11. Cost and Daily Cost in Statistics would refer to elements instead of items
  12. Some hints were missing in Statistics and Postpone windows
  1. PNG compression
  2. Download images allows of the preview of all images before importing to SuperMemo
  3. Instead of cascading, Download images tiles images for easier culling in extracts
Beta F
  1. Cloze deletion would not reuse the existing answer component when used on an item
  2. Templates would not paste correctly if copied directly from the template registry
  3. Textless HTML would not transfer correctly between collections (e.g. <FONT face=Arial color=#080000></FONT>)
  4. Download Images might leave temporary image files in the elements filespace
  5. Sender's address in mail import might be wrongly interpreted if the sender's name used diacritical characters
  6. Errors on mail import would not generate a report
  7. Archiving mail would produce a phantom error count
  8. SuperMemo might allow of loading a new element before the HTML file load completed
  9. Cancel on importing learning process from XML files would still proceed with import
  10. Switching back to the beginner level would not hide the statistics
  1. Localize in Download Images to localize remote images without integrating them with the registry
  2. Text : Full HTML to keep full HTML code when needed (styles, frames, scripts, etc.)
  3. Border coloring in HTML components to visually differentiate between Full HTML, Inner HTML and plain text
  4. Removing Border parameter from components
  5. Removing Scrollbar parameter from components
  6. File : Level : Warrior for one-click advanced level layout


Beta G
  1. Download Images would use absolute instead of relative file paths
  2. Download Images could delete properly localized pictures at import
  3. Shift+Ins and Ctrl+V shortcuts would not work for pictures in HTML components (ignored by mshtml.dll)
  4. Pasted pictures would hide behind other components in presentation mode
  5. Edit : Edit texts menu would ignore HTML components
  6. Upgrading from SuperMemo 2002 might render some MPG files invisible
  7. Alt+F shortcut would not work in the file browser
  8. Spell-Pad would not be considered as an answer in swapping questions with answers
  9. Toolbars too cramped in Warrior layout at lower resolutions
  10. Upgrades from SuperMemo 98 might fail if the upgrade from SuperMemo 8 to SuperMemo 98 crashed in progress
  1. The user can choose if pasted pictures should be kept in HTML or in the registry
Gamma E
  1. Some plain Asian texts would be incorrectly interpreted as valid UTF-8 strings
  2. Collections upgraded from SuperMemo 98 would incorrectly scale templates
  3. Non-Unicode Cyrillic would not import correctly as XML
  4. Long UTF8 encoded texts would not display correctly in the registry or in the contents window
  5. The ancestor window would not apply the contents font
  6. Ctrl+Alt+L in the search dialog box would conflict with AltGr+L
  7. File : Merge collection would not transfer the layout to collections that did not exist before the transfer
  8. Some diacritical characters would not transfer correctly between Windows and Pocket PC
  9. Lexicon registry would ignore non-Latin words
  1. Allowing of mixed ANSI and UTF8 encodings in the text registry
  2. Possibility of converting code page fonts to Unicode on export to XML
  3. Search : Find Unicode
  4. Text : Convert menu allows of encoding UTF-8, decoding UTF-8 and code page conversion to Unicode
  5. Faster member delete at the registry level resulting in 25% faster text edits (3 instead of 4 passes of binary chop on the member index)
Gamma F
  1. Ctrl+Shift+Y would throw an error if there were no components to transcribe
  2. Wrong guess of intervals for substantially delayed repetitions imported via XML
  3. Error message when converting tasks to items via Element parameters
  4. RTL text-flow would be ignored if the element was not first set as Full HTML 
  5. RTL text-flow would not be inherited in HTML extracts 
  6. Wrong relative path determination when extracting localized pictures with URLs that use backslashes instead of slashes
  7. XML import based on the current element's template would report errors
  1. HTML status borders is false by default (#1 complaint among users who upgraded from SuperMemo 2002)
  2. Text : Align : RTL-LTR in HTML components simplifies setting the right-to-left text direction (e.g. for Arabic, Hebrew, etc.)
  3. HTML filter allows of filtering out <?xml> declarations as well as the TEXT-ALIGN style
Gamma G
  1. Click on HTML components would set focus without changing the active component
  2. BOM markers not skipped in Q&A import in UTF-8 encoding
  3. Answers in cloze deletions would not inherit the text direction
  4. Link : Hyperlink would be active but not operational in non-edit modes
  1. Text : Alignment : Default RTL for simple change of text direction with right-alignment button (for Hebrew, Arabic, etc.)
Gamma H
  1. Due to UTF-8 sorting ambiguities, title shortening might result in duplicate registry entries
  2. Pastes from some OLE applications would execute twice
  3. Unintelligible error reporting on failed DDE transactions
  1. Element titles are not stored as separate registry member if longer text equivalent exist. On longer cloze deletions, this reduces the number of text registry members up to 40% and their size up to 30%
  2. F5 shortcut is used now for the Warrior layout instead of the Classic layout
Gamma I
  1. Repair collection would leave UTF-8 encoded texts encoded when removing hidden formatting with Remove hidden formatting check
  2. Some non-UTF-8 Japanese texts would be UTF-8 decoded on Q&A import
  3. Multi-picture and multi-table HTML texts are now less sensitive to HTML selection mshtml bug on Alt+X extract
  4. Outstanding field would not scale down the font for longer strings on the status bar
  5. Minimize button was active in the file browser
  6. Empty icon would show in the statistics window
  1. Decode UTF8 option in Q&A import for choosing between UTF8 decoding and literal import
Gamma J
  1. Selecting words with double-click in HTML would result in erratic mouse behavior
  2. Basic level would disallow of the status bar
  3. Swap Q&A in the browser would display a range error on untitled items
  4. Forgetting index vs. Grade approximation might show inversed monotonicity if uninitialized data were imported from another application (e.g. MemAid). This would occur if a large number of well-rehearsed items were presented as new to SuperMemo (esp. for substantially delayed repetitions)