## 5 Badges

11 years, 102 days

## Maple's very **unusual** sign convention...

Yes, I saw a few comments by well-known relativist Matt Visser on the *very unusual* choices of signs in Maple's Tensor Package & I'm getting used to those 'peculiarities'. Now I can foresee the correct results of any computation. Thank you much for your response.

Best,

Isaac

## Maple's very **unusual** sign convention...

Yes, I saw a few comments by well-known relativist Matt Visser on the *very unusual* choices of signs in Maple's Tensor Package & I'm getting used to those 'peculiarities'. Now I can foresee the correct results of any computation. Thank you much for your response.

Best,

Isaac

## Hi Lark 49, Thank you for your help. I c...

Hi Lark 49,

Thank you for your help. I copied your short code in a blank wrksheet and it worked fine with me too.

My former wrksheet read:

>with(tensor):
>coord:=[z,w,u]:g_compts:=array(symmetric,sparse,1..3,1..3):
>g_compts[1,1]:=1:
g_compts[2,2]:=cosh(z/a)^2: g_compts[3,3]:=cosh(z/a)^2*cosh(w/a)^2:
>with(tensor):g:=create([-1,-1], eval(g_compts)):
>ginv:=invert(g,'detg'):D1g:=d1metric(g, coord):
>Cf1:=Christoffel1 (D1g): Cf2:= Christoffel2 (ginv, Cf1):D2g:= d2metric (D1g, coord):
>RMN := Riemann( ginv, D2g, Cf1 ):RMNc:=get_compts(RMN):
>map(proc(x) if RMNc[op(x)]<>0 then x=RMNc[op(x)] else NULL end if end proc,
[ indices(RMNc)] );
>RICCI := Ricci( ginv, RMN ):
>RS := Ricciscalar( ginv, RICCI );

and it keeps returning the same errors as before.

Cheers,

Isaac

## a Maple-13 COLOSSAL bug in built-in *ten...

Further to my previous post here of a few days ago, namely "wrong sign in scalar Riemann curvature using Maple 13", so far uncommented, I should add the remark that the 'tensor package' built-in in Maple 13 has got what one could call a *COLOSSAL bug* that comes at least from Maple 9.5 and has apparently never been detected by any user so far.

Hard to believe, Maple-13 calculates & returns the components of the Riemann tensor, the ones of the Ricci tensor and the Riemann curvature scalar, all with the WRONG SIGN (!!) even for the current 3-dimensional hyperbolic space (in Lobachevsky coordinates):

ds^2=dz^2+cosh(z/a)^2*dw^2+cosh(z/a)^2*cosh(w/a)^2*du^2

The Riemann scalar curvature returned by Maple-13 is *R=6/a^2* in stead of the correct value *R=-6/a^2* as appropriate for an hyperbolic variety whose curvature is NEGATIVE as is known to every student of non-Euclidean (hyperbolic) geometry: (the returned values of the Christoffels of both kinds are correct). The BUG is in the computation of the components of the Riemann tensor and all sign errors come from here.

So, please be warned about the errors that can turn out with the careless use of the said software. This should be taken as a service to Science rather than a simple malevolent comment.

Best,

Isaac

 Page 1 of 1
﻿