Suggestions Not In Consideration

From SuperMemopedia
Jump to: navigation, search
  • Rich texts in Plan
  • More options for leech management
  • Remove the midnight shift concept
  • Commander is wrongly designed
  • Change the way Tab behaves at Show answer
  • No menus on toolbars
  • Separate SuperMemo and create a few applications
  • Release SuperMemo Light
  • Printing support
  • Auto-update SuperMemo?
  • Put all collection files into a single file
  • Include knowledge structure in the algorithm
  • Let SleepChart and SuperMemo detect time zones
  • Inserting marked text into question or answer components
  • Make final drill configurable
  • Integrate SuperMemo with speech synthesis products
  • Plugin or Active X for processing PDF in SuperMemo
  • Text-to-speech support
  • implement neural network promised. On supermemo website, the author said he considered adding it to supermemo in the FUTURE. I am not sure WHEN supermemo will encompass that feature. next 5 year, 10 years...? (neural networks are not well-suited to replace the algebraic algorithm - see FAQs on the algorithm)
  • Replace forgetting index bar with retention rate. I used to assume forgetting index 10% = 90% retention. It turned out to be 97%. That is really misleading. More importantly, We are more interested in how much we can remember. "Forgetting index, A factor, B factor.." are too technical.
  • Make the font size compatible with Microsoft Word. Somehow, size 10 in supermemo looks as big as size 12 in microsoft word.[if this refers to HTML components: the size of the font depends on Internet Explorer settings, not SuperMemo]
  • Advanced Backup option - zip the entire collection as supermemo is performing backing up. Then insert into the archive the time the database is backed up. [it was implemented and voted out from earlier versions - it seems more practicable to use one's own favorite backup/compressor program]
  • Remove levels such as Basic, Professional, Warrior, from SuperMemo
  • Remove Next repetitions stage
  • Extracting words to a separate category
  • Supermemo should send out more news and let people know what you are doing and what the result might be. Right now, I don't know anything about supermemo's new move, or whether all the programmers are enjoying afternoon coffees or plan to write supermemo 2006 ["no announcements" is a very old policy to prevent vaporware - probably not likely to change, "no mass mailing" is also several years old, if a software line was to be terminated that would certainly be announced hence, in 2006, you can always safely assume that sm2006, or at least sm2007 is being worked on, even if the direction of change is not always to everyone's liking]
  • the old Undo problem: History Manager - Newest changes in supermemo will be saved as restoration points. If users did silly things, they can go back [changes are too complex to effectively store for undo purposes, the size of files stored would be very large, or garbage collection would slow down operations, millions of bugs could show up, copying the entire collection would be simpler, but this ... user can execute on his own at his own choosing and his own time][some FAQs at supermemo.com explain this problem, i.e. how many files and in how many places are modified when a simple change is done such as reformatting an element]
  • lock the occlusion shape with image item that is not set to stretch mode. So if the window is resized, the occlusion will still cover the image item that is set to normal mode. It has been a nightmare to me that all the occlusion shapes are misplaced when I maximise the window on computer with different screen resolution [this would require checking shapes and images on each resize and moving them around accordingly - often with results unexpected by the user; all resizes would then be very slow on weaker computers; instead of chosing "lock occlusion" option, choosing the right stretch mode once seems a more practicable solution]
  • Separate presentation and editing modes
  • Analysis function should be available for every item, branch, rather than just for whole collection. I have two folders set to different retention rate, so analysis for the entire collection is no longer accurate. A simple statistic function is avaialble, but not as advanced as the 2 D curves. The analysis window should also clearly indicate which folder or collection it is analysing [data used in Analysis is HUGE, it cannot possibly be kept for individual branches on the tree, it can only be collected upon request; 3D graphs refer to HUMAN MEMORY not LEARNING DATA, i.e. if they were to be kept for indidivual branches, they would only be less accurate, i.e. not an iota more informative]
  • Make "workload", "plan" available for each subfolder, rather than exclusive for the entire collection [this data is too big to be kept for collection subsets, see: workload.dat; however, a slow-running subset Workload could theoretically be added, it would probably open for a few minutes ...]
  • allow user to choose what error to fix, how to fix, what items will be affected BEFORE doing the repair. It should also explain what does "parent mismatch" mean and the consequences if it is not fixed [this is not specific enough, number of possible data errors is so enormous that if each was to receive dedicated handling, SuperMemo would quadruple its size; this would not add to user's learning power and as such would compete with usability functions]
  • A shortcut for the "Queue Cloze" and "Schedule Cloze" options. Manually forgetting the items one by one is annoying ,I am currently forced to interrupt my daily cloze generation (because the next day workload is already too full) or do complicated operations with the browser to do something that could be done with a simple shortcut. [pending queue should not be used as a prioritization tool; incremental reading includes a long-list of more sophisticated prioritization mechanisms; pending queue concept is retained only for the possibility of creating "unused" collections for others]
  • Drill the postponed items first,so that some items will not end up being postponed twice if I have no time to learn all of them [learning should proceed along a strictly defined priority criteria, those should be defined by the user, postponed-first is just one of many possible choices, always sorting the collection by default would take time and penalize all those who opt for other criteria]
  • Ad-ware
    • Suggestion: I don't mean the kind of software that peeks at users' computers. Supermemo can have an ad banner in the program. I have the feeling that supermemo program project is severely underfunded from the quality of the program. Since open source is not quite practical, introducing a small ad banner can at least provide more money to fully polish the program. The author at least is able to employ a REAL & PROFESSIONAL delphi programmer who writes programs for WINDOWS! I prefer seeing a much better written program with low price to having another crippled version with staggering cost
    • SMW: No advertising policy: Unfortunately there is a policy in force (since 1999) that SuperMemo is not advertised through commercial channels (except via partners who opt for their own solution). It relies entirely on supermemo.com and word-of-month. The pricing is consistently kept below $40 and all older versions are available at lower price or as freeware. For similar reasons, there is no advertising allowed neither in software nor at supermemo.com
      • "Unfortunately there is a policy in force (since 1999)" Welcome to 21st century! Option 1: some less annoying Ads + better faster cheaper Supermemo Option 2: no ads, little feature and design upgrades--> not so many people choose to upgrade --> less profit --> no ads, little feature and design upgrades==> so on. you can't have both.
  • remove plain text and RTF components
    • Suggestion: remove rich text and plain text field in "edit" menu. Supermemo has progressed to HTML. Why do we still need those options?
    • Cons: Consumers who use code page fonts for learning Chinese, or RTL text in RTF for learning Arabic would eat you alive! :)
  • Retain user modifications to the statistics display. One can move the various items (retention, burden etc.) around within the statistics display, but the changes are not retained if SuperMemo be closed and re-opened
    • This option has been removed from SuperMemo due to confusion caused by reshuffled statistics (incl. simple errors in interpreting the statistics once fields changed places)
  • Replace component types with just a component
  • Compile the entire set of SuperMemo Help pages into a single file, e.g. PDF. Also add an index to the file. This would facilitate searching for guidance on a Supermemo feature. I find the current help pages hard to navigate. The Search utility that goes with the web site is also not very useful: The results of a search are a list of phrases in very tiny text that are hard to read. Furthermore the phrases are not linked back to their specific web content, but rather to the main FAQ page to which they belong (PDF is more suitable for printing, HTML is more suitable for "navigating" - the documentation is hyperlink-rich, HTML docs are regularly updated)
  • Implement automatic import from CHM
  • Collection mirroring
  • Enable the import of a power point file. In the optimal case each power point page would compose a child topic of the top knod where it was imported.
  • implementation of RSS streams in IR, conveniently making text-extracts possible from the presented live pages
  • Possibility to predefine learning days. I have my copy of SM installed on my corporate machine. Thus I will definitely not learn on saturdays and sundays. It would be helpful if the computation of spacings could be accomodated with regard to a predefined availability schema. Adding holidays to this would be even greater!
  • Implement Interpreter for basic LaTex commands especially the math mode (\frac{}{}, \sqrt ^ _ \dot{}, \bar{}, ...) and special characters (greek,...). Latex provides a fast and memory saving way to input formulas, and much of my documentation is already available in Latex. It would be also nice if the same interpreter could be implemented on the ppc.
  • Improved Q&A text import to be able to create articles and hierarchical structure
  • text-to-speech support
  • Integrate Plan with Outlook
  • Mark images as not stored in the registry
  • Shortcut for RTL conversions
  • Better printing support. The whole idea of exporting and then printing is very sloppy. Ability to print jest questions and questions with answers (good for preparing tests for students)
  • Native Linux support. -Dempf
  • Make the "thumb preview"-option available for Windows XP
  • Auto copy text on a component into the Windows clipboard when it appears?
  • Allow of secondary storage for categories
  • Extend Plan to the entire week
  • Actively space out sister cloze deletions (through maximising difference that is normally left to randomness)

see also: Suggestions