Re: few easy questions
- To: mathgroup@smc.vnet.net
- Subject: [mg11500] Re: [mg11451] few easy questions
- From: "C. Woll" <carlw@u.washington.edu>
- Date: Fri, 13 Mar 1998 12:21:52 -0500
Hi Arturas,
Concerning your first question, the reason Dt[x_]:>1 doesn't work can be
found by examining the FullForm of the rule
FullForm[Dt[x_]:>1]
RuleDelayed[Times[Dt[x],Derivative[1,0][Pattern][Blank[]]],1]
We see that the derivative is acting on the function x_ (which in
FullForm is given by Pattern[x,Blank[]]). I can see two ways around
this. One is to use the rule
Dt[a] /. HoldPattern[Dt[x_]]:>1
The other is to use the rule
Dt[a] /. x_Dt :> 1
as you basically did. The problem with the latter approach is if you
want the rule to depend on the value of a. Suppose you want Dt[a] to go
to some function of a, f[a]. Then you could do the following
Dt[a] /. x_Dt :> f @@ Unevaluated[x]
Is this the explanation you were looking for?
Carl Woll
Dept of Physics
U of Washington
On Thu, 12 Mar 1998, Arturas Acus
wrote:
> Hello, Group
>
> 1)
> Why this two substitutions act differently In[9]:=
> {Dt[a], Sin[a]} //. {Dt[x_] :> 1, Sin[x_] :> 1} Out[9]=
> {Dt[a], 1}
>
> Of course the dirty trick bellow works, but I am still interesting
> what's
> going.
>
> In[10]:=
> Dt[a] //. {x_Dt :> mydt[x], mydt[x_] :> 1}
>
> 2)
> Thanks to P. Abbot I know about number of packages for tensor
> manipulation. I was rather interesting in "EinS" and planned to use it
> as a starting point of my own purposes. Unfortunately, the package is
> encoded, (like STensor, and MathTensor) and the author do not intended
> to put the actual
> code for free use. So I have few questions about that format a)Is the
> format simple encoding or it includes some precompilation, compression,
> etc.?
> b) How the encoding acts on general performance of the code(speed up?)?
> c) Do algorithms on manipulation of dummy indices are such a huge
> problem ?
> d)How to make this encoding myself ? ( I could not hope that anybody
> could answer me how to decode them :-)
> ).What is WRI opinion on the topic.
>
>
> Sincerely,
> Arturas Acus
> Institute of Theoretical
> Physics and Astronomy
> Gostauto 12, 2600,Vilnius
> Lithuania
>
>
> E-mail: acus@itpa.lt
> Fax: 370-2-225361
> Tel: 370-2-612906
>
>