[Date Index] [Thread Index] [Author Index]

Re: Re: Inconsistent evaluation

• To: mathgroup at smc.vnet.net
• Subject: [mg61989] Re: [mg61925] Re: Inconsistent evaluation
• From: leigh pascoe <leigh at cephb.fr>
• Date: Tue, 8 Nov 2005 03:41:31 -0500 (EST)
• References: <dkfcjb$eo5$1@smc.vnet.net> <200511050652.BAA01936@smc.vnet.net>
• Sender: owner-wri-mathgroup at wolfram.com

Jens-Peer Kuska wrote:

>Hi,
>
>how can we "reproduce" the problem on our own 
>machines ?
>how can we see what your initial cell does ?
>
>In general you should set up a different cell for
>Get[] and Needs[] to be sure, that the variables
>in the packages are parsed and added to the symbol
>table first. And than make a second cell with you
>own definitions. This was a problem with an older
>Mathematica version, that the parser scan the cell
>before the kernel evaluate it and so the parser
>create the symbols before it got the definition
>for it from the evaluation of Get[]/Needs[]
>
>Regards
>  Jens
>
>"leigh pascoe" <leigh at cephb.fr> schrieb im 
>Newsbeitrag news:dkfcjb$eo5$1 at smc.vnet.net...
>| Dear Mathgroup,
>|
>| I tried to post this question a few days ago, 
>but it didn't appear. I
>| assume this was because I pasted in error 
>messages from Mathematica in
>| non text form, so I'll try again.
>|
>| I have a notebook with an initial cell that 
>calls Combinatorica and
>| contains a number of function definitions of the 
>set delayed type. Each
>| time I open the notebook I seem to get a 
>different behaviour when I
>| evaluate the first cell. Two different sets of 
>messages resulting from
>| an evaluation of the first cell are pasted in 
>below this message.
>|
>|
>| Most recently I did the following:
>|
>| 1. Start Ma and open the file. Evaluate the 
>first cell. This gives me
>| the long list of error messages pasted in below.
>| 2. Then close Ma without saving anything. 
>Restart Ma and reopen the
>| notebook, without evaluating the first cell. I 
>then opened a new
>| notebook and pasted the definitions from the 
>original notebook and
>| evaluated them one at a time. No error messages.
>|
>| Other times I get a single warning about using 
>the variable "list"
>| (defined inside a module) that is similar to 
>"List". Does Ma remember
>| data from previous sessions even though I have 
>closed the program? What
>| am I not understanding here??
>|
>| Ma 5.0 Windows XP Pro
>|
>| Thanks
>|
>| LP
>  
>
 >Clip

Dear Jens,

Thank you for your response and suggestions. I am used to debugging code 
with errors, and the point of my question was not to find a particular 
error in the initialisation cell. What I found astonishing was that I 
could have different behaviour from what seemed to be identical inputs.

I think my problem may have been related to the questions regarding 
"whitespace" and "cell execution dependence" discussed in this forum. In 
any case I seem to have cured it by spacing out the cell a bit and 
making sure that comments are on separate lines from code. I had also 
been using the forms (* comment *), (*comment*), ( * comment * ), ( 
*comment* ), more or less indiscriminately wrt whitespaces? I have 
corrected this so that all comments are in the first form with no space 
between the left bracket and the asterisk.

I will also put the package calls in a separate first cell - thanks for 
the hint.

LP

• References:
  • Re: Inconsistent evaluation
    • From: "Jens-Peer Kuska" <kuska@informatik.uni-leipzig.de>