Re: TI-83 matrix memory management bug


[Prev][Next][Index][Thread]

Re: TI-83 matrix memory management bug



scott goehring wrote:
>
> my wife has discovered a rather pervasive (and rather nasty bug) in
> the matrix memory management routines on the TI-83.  if you do just
> about anything with a matrix that results in a data type error, the
> calculator goes into an unstable state, leading to any number of
> really weird errors, including, occasionally, fascinating
> video-game-like displays.  in nearly every case, battery removal has
> been required to recover the calculator; in at least one case, we had
> to remove the lithium backup battery as well.  she also managed to get
> a "Err: Link" out of it; this error is not documented in the manual,
> so we have no idea what it means.
>
> to reproduce this error, enter any matrix in [A] (or any other matrix
> slot).  divide it by a constant (IMO, this is a reasonable thing to
> do; the HP 48 series allows this, by the way).  you will get a "Data
> Type" error.  hit quit.  now try to change the dimensions of any
> matrix (doesn't matter which one).  now, remove the batteries: your
> calculator will be frozen solid.  (you should not do this if you have
> anything on your calculator you care about, obviously.)
>
> this has been replicated on several other units, and TI has been
> notified.
I don't know if this has any relation to a bug on the ti-82, but maybe
you'd like to check if there is nothing important on the 83.  This bug
is documented by myself at
http://www.millcomm.com/~bendzick/ti82bugs/ti82bugs.html .
I suspect nothing will be weird on the 83 with this bug, but I haven't
heard of anyone checking yet.


-Michael Bendzick
bendzick@millcomm.com


References: