Voici la réponse de ti-cares à mon mail :
Dear User ,
Thank you for your recent mail to Texas Instruments.
What you are experiencing is not a bug. The calculator is doing it
correctly. If you take e^3, it is
equal to 20.085536. Now take the natural log of that answer, you get 3.
Then you are subtracting 3 from it,
you will get zero. This will be the case for any integer greater or
equal
to 2. But when a is equal to 1, the
calculator gives an answer as "undef", with questionable accuracy
written
at the bottom of the screen.
The developers will try to implement it on the next version.
For any additional support , please don' t hesitate to contact the
Texas
Instruments
Customer Support Center by e-mail at TI-Cares@ti.com or call us at
0155212508
If e-mailing please copy this message in your response for faster
service
on replies.
Best Regards,
Marika Olsson
TI Customer Support Center
http://education.ti.com
========================= ORIGINAL MESSAGE ==========================
[ Clément <keelize@yahoo.fr> on 29/Nov/01 16:00:34 ]
Hello, I wantto report a bug :I have some strange results when calculating a
limit (under AMS 2.05) :
lim((ln(x)-a)/(x-e^a),x,e^a) givesthegoodanswer : e^-a
lim((ln(x)-a)/(x-e^a),x,e^a)|a=1 also gives the good answer 1/e
lim((ln(x)-a)/(x-e^a),x,e^a)|a=any integer > 2 gives 0 !!!
lim((ln(x)-a)/(x-e^a),x,e^a)|a=2. gives 0. !!!
lim((ln(x)-a)/(x-e^a),x,e^a)|a=3. gives undef !!!
lim((ln(x)-a)/(x-e^a),x,e^a)|a=5. gives the good approx value !!!
I just can't understand how the ti89 work !
Could you reply to this mail to show me you've read this.
Thank you much
keelize
keelize@yahoo.fr
Je sais vraiment pas quoi leur répondre, ils sont vraiment trop cons ou ils se foutent vraiment de ma gueule
