@arjangash 

Y:=numer(y)*denom(y):
CodeTools:-Usage(
SolveTools:-SemiAlgebraic([Y>0, t>0,lambda>0,c>0])
):
 memory used=484.52MiB, alloc change=113.04MiB, cpu time=4.37s, real time=4.28s, gc time=312.00ms

nops(%);
                               8

You cannot hope for nanoseconds!

@tomleslie 

It is a linear problem. simplex[maximize] finds the exact solution.

@AmusingYeti 

My remark refers to eigenvectors, not Cholesky which anyway does not apply directly to Av = aBv.

It would be interesting to see an example where eigenvalues works only after the mentioned transform.

@MDD 
{3, 4, 11}  contains the indices of the facets in the list L.

@acer 

Thanks, I forgot that simplex is still table-based, but I wonder why ":-" works in Maple 2017 and not in Maple 18 (or maybe <18 ?). I changed this in the code.

@MDD 

OK, I shall include a direct solution in your initial question https://www.mapleprimes.com/questions/223770-How-Can-Remove-The-Redundant-Inequalities

@MDD 

You may try to use the Convex package http://www-home.math.uwo.ca/~mfranz/convex/

This is just like computing m*n using the cartesian product {1,2,...,m} x {1,2,...,n}.

@carriewong 

@carriewong 

I mean that no prerequisites are needed to obtain the desired one-line procedure. You asked for num1dsubspaces and you have it in my answer.

@mmcdara 

1. Forget the square roots.
You have a 3 digits FP unit and two numbers x = 1.41 and y = 1.73.
What is x*y using this machine?

2. On the same machine compute x = sqrt(2) and y = sqrt(3).
GO TO 1.

@Giulianot 

Plotting procedures has some advantages and it's the simplest fix. It does not work for you?

@_Maxim_ 

This has a single index and is of course very useful (and consistent with rtables). My opinion was about two indices L[u,v].

@Alex Bowden 

You refuse to try to understand the problem and see the general picture.
It is far beyond your 4(1+x) and (x+1)(x-1) and you do not have a constructive solution.
There are other solutions to remove the ambiguities. For example, Mathematica accepts 4(x+1) but requires square brackets for functions ( Sin[x],  f[x] ). I prefer Maple's approach. So, let's stop here.

@acer 

This is a parametrized solution too, the difference is that it is not global (it has 4 maps) and uses cartesian coordinates.
The corresponding solution with 2 maps  and spherical coordinates (curves included):

el:=[x=9/2*sin(phi)*cos(theta),y=6*sin(phi)*sin(theta), z=3*cos(phi)]:
sp:=[x=4*sin(phi)*cos(theta),y=4*sin(phi)*sin(theta), z=4*cos(phi)]:
phi1 := arccos(sqrt((63*cos(theta)^2-80)/(63*cos(theta)^2-108))): phi2:=Pi-phi1:
phi3 := arcsin(21/(2*sqrt(140+49*sin(theta)^2))): phi4:=Pi-phi3:
p1:=plot3d(rhs~(el), theta=0..2*Pi,phi=phi1..phi2,color=blue,style=surface):
p2:=plot3d(rhs~(sp), theta=0..2*Pi,phi=phi3..phi4,color=red,style=surface):
p3:=plots:-spacecurve( eval(1.02*rhs~(el),phi=phi1), theta=0..2*Pi ,color=yellow, thickness=2):
p4:=plots:-spacecurve( eval(1.02*rhs~(el),phi=phi2), theta=0..2*Pi ,color=yellow, thickness=2):
plots:-display(p1,p2,p3,p4,scaling=constrained);

@Alex Bowden 

If you really hate the explicit multiplication operator then stick to Math 2D notation and put a space instead of *.

a (b)  or  a b  or (a)  (b)  will be interpreted as a*b.

Note that the "mathematical notation" is (or can be) ambiguous without the context  and a language (such as Maple, Mathematica, C etc) cannot cope with ambiguity.