[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: OT - programming logflash Re: [lojban] Logflash
At 07:10 PM 05/08/2000 -0400, Brook Conner wrote:
>From an engineering standpoint, a flash-card vocabulary drilling program
has several different components. Some of these are best left to
existing libraries (because they've been done and aren't specific to the
task at hand - displaying text and inputting it, for example). Others
are more specific, such as generating random permutations of the
vocabulary lists and tracking performance on the vocabulary. Logflash
puts the vocab into "piles" based on how well the user is doing on that
word. Deciding when to move vocabulary from pile to pile is an
interesting part of the problem.
Note that for LogFlash, the algorithm for the above was precisely the one
developed over several years by JCB and his first wife with hardcopy flash
cards specifically for learning Loglan gismu. Apparently this was the
optimal algorithm among many variations tested, though as usual he never
documented what the testing and option alternatives that were found inferior.
The ultimate test is in the results. People have found LogFlash boring or
frustrating, but so far as I know, no one who has completed the program has
had cause to complain that they hadn't thoroughly learned the words.
Indeed almost 20 years after she wrote the first version of LogFlash on a
TRS-80 in BASIC, and 12 years after completely relearning the Lojban
vocabulary, but rarely practicing Lojban, Nora is now brushing up with
LogFlash and scoring 90% plus, and her worst problems seem to be
occasionally false memory links to the Loglan words she learned with
LogFlash before we redeveloped the Lojban vocabulary.
There were many suggestions for alterations to the algorithm, and the
current LogFlash option screen allows several of the easier to implement
ones. Yet even today, I make the same recommendation I followed in 1987
and 1988, use Gaining Control Mode, and the largest New Word Pile size you
can stand, preferably one that gives between 15 and 25 error words per new
word lesson, and do it daily without fail (this will typically take about
an hour a day). In a month you'll finish the first pass through the
words. Within 3 months or so, increasing your lesson size when your error
rate goes down, you'll get 97% or better on the words (in my case doing 500
words in a half hour test).
lojbab
----
lojbab lojbab@lojban.org
Bob LeChevalier, President, The Logical Language Group, Inc.
2904 Beau Lane, Fairfax VA 22031-1303 USA 703-385-0273
Artificial language Loglan/Lojban: http://www.lojban.org (newly updated!)