MathGroup Archive 2007

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

Search the Archive

Re: Question

  • To: mathgroup at
  • Subject: [mg74892] Re: Question
  • From: dh <dh at>
  • Date: Tue, 10 Apr 2007 05:17:25 -0400 (EDT)
  • References: <ev51pb$ai7$>

Hi Peter,

you are right , there is something very surprising concerning "Part". I 

beg you to mention this to Wolfram, e.g. to Daniel Lichtblau.

Consider the following expression that has one level less than yours, 

but will do to make the point:

expr = a[

b1[c11[d111, d112, d113], c12[d121, d122, d123], c13[d131, d132, d133]],

b2[c21[d211, d212, d213], c22[d221, d222, d223], c23[d231, d232, d233]], 

b3[c31[d311, d312, d313], c32[d321, d322, d323], c33[d331, d332, d333]]


Now expr[[1,1,1]] gives d111 what is expected. expr[[1,1,{1,2}]] does 

not give {d111,d112}, but c11[d111,d112]. Therefore, as soon as we use 

more than one index on the same level, the head of the level above is 

applied.  I wonder if this is a bug or intention, I just can't imagine 

an application for this. Further, expr[[2,1,{1,2}]] gives: 

c21[d111,d112] and expr[[{1,2},1,{1,2}]] does the apply "trick" once 

again: a[c11[d111,d112],c21[d111,d112]]

Hope this will help, Daniel

> Hello Daniel,

> thanks. It is quite a decision to apply the head to a list when talking about parts.

> I would expect just a list of parts.

> Still I am not able to predict what happens to:

> Lambda0[[{1, 2, 3}, 1, {1, 2, 3}, 1]]

> which gives: Lambda0[Lambda2[1, 1, 1], Lambda2[2, 2, 2], Lambda2[3, 3, 3]]

> What confuses me is:

> Lambda0[[{1, 2, 3}, 1, {1, 2, 3}]]

> which gives: Lambda0[Lambda2[Lambda3[1, 1, 1], Lambda3[1, 1, 2], Lambda3[1, 1, 3]], Lambda2[Lambda3[2, 1, 1],

>     Lambda3[2, 1, 2], Lambda3[2, 1, 3]], Lambda2[Lambda3[3, 1, 1], Lambda3[3, 1, 2], Lambda3[3, 1, 3]]]

> So, I do not understand anything about these manipulations.

> Can anyone explain these results? In one way or another I am wrongly focussed.

> with frienldy greetings,

> Peter van Summeren


  • Prev by Date: Re: numerical inversion of laplace transform
  • Next by Date: ColorFunction help?
  • Previous by thread: Re: question
  • Next by thread: Re: Re: Question