Mathematica 9 is now available
Services & Resources / Wolfram Forums
-----
 /
MathGroup Archive
2007
*January
*February
*March
*April
*May
*June
*July
*August
*September
*October
*Archive Index
*Ask about this page
*Print this page
*Give us feedback
*Sign up for the Wolfram Insider

MathGroup Archive 2007

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

Search the Archive

Re: [TS 35470]--Re:bouncing from error (bug?) to error (bug?)

  • To: mathgroup at smc.vnet.net
  • Subject: [mg72545] Re: [TS 35470]--Re:bouncing from error (bug?) to error (bug?)
  • From: "Chris Chiasson" <chris at chiasson.name>
  • Date: Sat, 6 Jan 2007 03:42:31 -0500 (EST)
  • References: <200701051500.l05F09NT004482@localhost.localdomain>

On Fri, 5 Jan 2007 15:00:09 UT, Tom Zeller wrote:
> Chris,
>
> You have probably looked at the documentation for the ssle message.  It
> is a general message:  General::ssle which is generated when the indicated
> function argument does not have the form indicated in the message.  The
> example from the documentation is Context[{Integrate}] where the error
> occurs in this example because the argument in Context is a list rather
> than a symbol.
>
> Since Attributes works on the Head and the Head of your function is blah[{1}] it is seeing  the list as a problem even though it is not, in this
> case, an argument.  I'll send your code in to Quality Control.
>
> I did try Trace on the NMinimize example you gave to see where Attributes
> was called.  The result was quite a few more messages and 42 MB give or take of output,amongst which was quite a few things like this (in part):
> for a variety of Optimization`NMinimizeDump`*:
> ------
> Optimization`NMinimizeDump`dbPrint[
>                                         3, {"   Ok Elements"}];
>                                         If[FreeQ[
>                                         Optimization`NMinimizeDump`tmp,
>
>                                         Integers] &&
> \((MemberQ[Attributes[
>
> Evaluate[Head[\(blah[{1}]\)[a]]]],
>                                         NumericFunction]
>
> ---------------
> One that had the ssle message associated with it was (in part):
>
> Optimization`NMinimizeDump`vars]\),
>                                         HoldForm], ",",
>                                         RowBox[{"{",
>                                         RowBox[{
>                                         RowBox[{"{",
>                                         RowBox[{
>                                         RowBox[{"{",
>                                         RowBox[{
>                                         TagBox[\(Head[\(blah[{1}]\)[a]]\),
>                                         HoldForm], ",",
>                                         TagBox[\(blah[{1}]\),
>                                         HoldForm]}], "}"}], ",",
>                                         TagBox[\(Attributes[blah[{1}]]\),
>                                         HoldForm], ",",
>                                         RowBox[{"{",
>                                         RowBox[{
>                                         TagBox[
>                                         RowBox[{"Message", "[",
>                                         RowBox[{\(Attributes::"ssle"\), ",",
>                                         TagBox[\(Attributes[blah[{1}]]\),
>                                         HoldForm], ",",
>                                         TagBox["1",
>                                         HoldForm]}], "]"}],
>                                         HoldForm], ",",
>                                         RowBox[{"{",
>                                         RowBox[{
>                                         TagBox[
>                                         RowBox[{"MakeBoxes", "[",
>                                         RowBox[{
>
>                                         RowBox[{\(Attributes::"ssle"\), \(\(:\
> \)\(\ \)\), "\<\"Symbol, string, or HoldPattern[symbol] expected at position \
> \\!\\(1\\) in \\!\\(Attributes[\\(\\(blah[\\(\\({1}\\)\\)]\\)\\)]\\). \
> \\!\\(\\*ButtonBox[\\\"More\[Ellipsis]\\\", ButtonStyle->\\\"RefGuideLinkText\
> \\\", ButtonFrame->None, ButtonData:>\\\"General::ssle\\\"]\\)\"\>"}], ",",
>                                         "StandardForm"}], "]"}],
>                                         HoldForm], ",",
>
>                                         TagBox[\(RowBox[{RowBox[{"Attributes",\
>  "::", "\"ssle\""}], ": ",
>                                         "\"Symbol, string, or \
> HoldPattern[symbol] expected at position \\!\\(1\\) in \
> \\!\\(Attributes[\\(\\(blah[\\(\\({1}\\)\\)]\\)\\)]\\). \\!\\(\\*ButtonBox[\\\
> \"More\[Ellipsis]\\\", ButtonStyle->\\\"RefGuideLinkText\\\", \
> ButtonFrame->None, ButtonData:>\\\"General::ssle\\\"]\\)\""}]\),
>
> ---------------
>
> Two changes to the function definition can avoid the error.  Drop the
> NumberQ which from your preamble does not seem optimum or drop
> the "indexing" of the function, e.g. go with blah[a_?NumberQ] := a^2 or
> blah1, blah2, etc, which may or may not fit the design of your code.
>
> You can also just turn off the message as it seems the the NMinimize does
> evaluate once it gets numeric values.
>
> Tom Zeller
> Wolfram Research Technical Support

The indexing was only important because I needed to filter (hide)
certain Parts of a complicated  argument (a function and some
constraints) to NMinimize. I decided to make the index of the
"screening symbol" match the Part of the original expression it
screened.

One workaround is to simply generate a unique symbol for each part
instead of indexing one symbol (which is the same as your "drop the
indexing" workaround).

Thanks for sending the code to QC.

-- 
http://chris.chiasson.name/


  • Prev by Date: Re: integrate
  • Next by Date: Re: Re: A pattern matching problem
  • Previous by thread: Re: integrate
  • Next by thread: conversion from DensityGraphics to Graphics ignores ColorFunctionScaling->False in Mathematica 5.2