Version history
Version |
Comments |
V 1.0 (march 1999) |
Yeeeeaooooh ... the first »official« ;-)) version
is out! First versions are never good, but actually, we were quite satisfied... ;-)) |
V 1.1 (february 2000) |
Several improvements:
-
The same word can be recalculated by entering '=' (its just comfortable)
-
Some few and inefficient debug functions added (available by pressing <d>
in the main menu; I added them 'cause I thought that it would help me to
find one of the bugs..., but it didn't :-( )
-
Statistical functions added: new log-files are created:
-
RULES.LOG: shows what rules have been used and how often
-
WORDS.LOG: shows what rules have been used for each word
-
Screensaver (... well, it's probably not very useful but anyway... it made me feel better while programming it...:-))
I also tried to put a little bit of order
in the source-files, but... you know :-))
O, by the way: even after days of debugging the program I didn't succed
in finding "multiple evolution"-bug described below... :-( |
V 1.2 (may 2000) |
The first version which can be considered bug-free... (well, touch
wood!;-))
Improvements:
-
The "multiple evolution bug" has been fixed (Yeeeeaaah, gotcha!:-)): all branches of the evolution
tree are now correctly calculated
-
A bug in the corpus parser has also been fixed: commentaries and blank
lines can be added without causing calculation errors
|
V 1.3 (october 2000) | Some minor improvements (nothing substantial, actually). |
Known bugs and weak points
Although we have succeeded in fixing the most important bugs in version 1.3 (including the "multiple evolution"-bug which caused incomplete calculations), there remain some un unpleasant problems that have not been corrected until now. The most important one is a memory problem which leads to wrong results when the program is asked to process formulas containing wildcards (* or .) and/or empty sounds (&). In many cases, the formulas are not correctly executed: they are applied to words they shouldn't be applied to which normally causes sounds to fall out which normally shouldn't fall out. If you notice irregularities of that type, do the following in order to avoid the problem:
- Start the program and load a rule catalogue.
- Do a calculation (e.g. a corpus calculation)
- Exit the program (by entering 'e')
- Restart the program and reload the rule catalogue.
- The calculation should now be correct
The memory-bug may also lead to complete system crashes but this happens very rarely: personally, I have been able to work during hours (even when calculating large word-corpora) without even crashing once. I just mention it here in case it happens to you.
Please note, that there are also the following problems:
-
formalism @*V > @*V doesn't work. (Must be entered: @#*V
> @*V) (BUG)
-
the program doesn't warn if the user has committed a spelling
error (e.g. a[long] instead of a[+long]) in the corpus (WEAK POINT)
-
errors in file WORD.LOG: rules for some words missing (BUG)