W3C home > Mailing lists > Public > www-archive@w3.org > December 2003

cwm's tanh builtin

From: Jos De_Roo <jos.deroo@agfa.com>
Date: Fri, 5 Dec 2003 18:55:20 +0100
To: "timbl" <timbl@w3.org>, "connolly" <connolly@w3.org>
Cc: www-archive+n3bugs@w3.org, "Piet Dewaele" <piet.dewaele@agfa.com>
Message-ID: <OF0799D5A9.0B336961-ONC1256DF3.006165E4-C1256DF3.006272AA@agfa.be>

The different conclusions were found, but the bug is in euler :(
I was completely mistaken by mixing up tanh and atan and
forgot about tanh(x) = i.tan(x/i).

Sorry for the confusion :)

--
Jos De Roo, AGFA http://www.agfa.com/w3c/jdroo/
----- Forwarded by Jos De_Roo/AMDUS/MOR/Agfa-NV/BE/BAYER on 2003-12-05
06:43 PM -----
                                                                                                                                       
                      Jos De_Roo                                                                                                       
                                               To:       timbl@w3.org@internet, connolly@w3.org@internet                               
                      2003-12-02 02:09         cc:       www-archive+n3bugs@w3.org, Piet Dewaele/AMEGE/MOR/Agfa-NV/BE/BAYER@AGFA       
                      PM                       Subject:  cwm's tanh builtin                                                            
                                                                                                                                       
                                                                                                                                       



Tim, Dan,

While doing some tests with angle measurements I found
different conclusions with cwm than I found with euler.
I simplified the case to a simpler one:

###################################################
@prefix log: <http://www.w3.org/2000/10/swap/log#>.
@prefix math: <http://www.w3.org/2000/10/swap/math#>.
@prefix : <testP#>.

:M :ratio 1.
{:M :ratio ?R. ?R math:tanh ?A} => {:M :angle_rad ?A}.
{:M :angle_rad ?A. ?A math:degrees ?D} => {:M :angle_deg ?D}.
#############################################################

and cwm thought

    :M     :angle_deg 43.636130838093536;
         :angle_rad 0.76159415595576485;
         :ratio 1 .

but euler got

:M :angle_deg [iw:Variable "_:X_2"; = 45.0].


I experimented a bit and got a likewise result with a change
in cwm_trigo.py
__cvsid__ = '$Id: cwm_trigo.py,v 1.10 2003/10/20 17:31:54 timbl Exp $'

line 123
        return tanh(numeric(subj_py))
               ^^^^atan

i.e. cwm then thought

    :M     :angle_deg 45.0;
         :angle_rad 0.78539816339744828;
         :ratio 1 .

This was with Python 2.2.1 - the is running quite well :)

--
Jos De Roo, AGFA http://www.agfa.com/w3c/jdroo/
Received on Friday, 5 December 2003 12:55:26 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 7 November 2012 14:17:37 GMT