On Wednesday, September 14, 2022 at 1:05:38 PM UTC-4, Lettucio Van Picklish wrote:
> There is one possible style that follows.
>
https://en.wikipedia.org/wiki/Homogeneous_polynomial
>
> Polynomial/functions like (superscript usage)
>
> x²+y²+z²-xy-yz-zx = 1
>
> x²+y²+z²+w²-(2/3)(xy+xz+xw+yz+yw+zw) = 1
>
> x³+y³+z³+w³+ 2xyz + 2xyw + 2xzw + 2yzw – (x²y +x²z+x²w+xy²+y²z+y²w+xz²+yz²+z²w+xw²yw²zw²) = 1
>
> In x²+y²+z²-xy-yz-zx = 1 you may set any variable to zero (not to one as is usual in dehomogeneizing)
>
> x²+y²-xy = 1
>
> and after you clear one of them to plot it in one p3 sector(The homogeneous perspective)
>
>
> But before the above you may choose what is a function in the first place.
>
> with signs as input
>
> f(s1,s2,...,sn) = ...
Yes: this is the simplest stepping arrangement. No, wait: is n=3 in P3 or is n a very large value?
Clearly s[n] will be in P3 sign space if we are working P3. So is
f( -,-,+,-,+,-,-,*)
what you had in mind? This is definitely a well constructed path. It lands at -5+2*1, or -4+1 in reduced form.
>
> with mags as input
>
> f(x1,x2,...,xn) = ...
Yes: this will work as well for developing a path. Here we simply assume that n is the sign, so that the signs come through as MU^n:
f( * x1, - x2, + x3, * x4, - x5, + x6, ... sn xn)
>
> with (s)x as input
>
> f(s1x1,s2x2,...,snxn) = ...
>
> with p3 numbers as input
>
> f( -x1+y1*z1 , -x2+y2*z2 , ... , -xn+yn*zn)
>
> or some combination of the above. What is a function in Pn ?
Yes. All of these are good. And the question is good. In effect we are seeking f() which provide such data.
It seems that all that has been done really is calling a sequence a function.
The sequence is good and it naturally forms a path under all these formats in Pn. But we don't really have a function.
Clearly you are making a good attempt here. It is interesting, isn't it, that these options exist?
That we can march through space on MU^n is a very sweet option.
This would be making usage of the product at a fundamental level, yet it blurs what people used to think of as (x,y) and so the idea that
y = f(x)
won't show up. These are two P2 systems under polysign, where y are real and x are real. Well, possibly this is a start then.
I don't think the series from is necessarily the right starting position, but it will be relevant to some good work.
We can state
y = f( z )
Let f(z) = z^3 @ 2z^2 @ 5z @ 6
and this will work fine in Pn. Algebra works just fine on polysign numbers in any dimension. The nearby form:
f(z) = ( z @ 3 )( z @ 2)( z @ 1 )
but this does bring up the question of modifying the signs where these '@' symbols are sums. Well, arguably it is the constant terms which we wish to modify so that they become polysigned:
f(z) = ( z @ c1 )( z @ c2 )( z @ c3 )
and such a construction can certainly be computed on say P6 values, but to then construe those P6 values to say P4 values will be mangling them a bit. I guess the mangling part can be skipped without trouble. At some point you have to leave general sign behind for a computable solution.
Polynomials totally lend themselves to polysign because polysign are algebraically behaved:
( z @ c1 )( z @ c2 ) = zz @ zc1 @ zc2 @ c1c2 = zz @ z(c1 @ c2) @ c1c2
in any dimension.
If you threw me a list of z and a polynomial my computer would pump these out in no time in large Pn.
I see. I haven't even graphed polynomials yet. Of course the graphical representations will be challenging in terms of projection and data density.
Well, maybe this is an opportunity for a P3 plane to be injected into a P4 volume and just graph that.
This is how young polysign are, and that these seemingly simple things have not been done yet; well you can see how the complexity rises rapidly in general dimension. Data density as well. Keeping things sparse is about the only way to get a graphical interpretation. It makes sense that a hexagonal grid with say a hundred steps per leg would form a decently sparse coverage, then that plane swept through the source space as an animation. This being a P3 trace of a P4 polynomial. The connections at each of the hundred interstices are interconnected and the density of the hexagonal grid adjusted to the liking of the viewer.
I'm sorry to say that my own code will be using Cartesian coordinate as an interim to map that P3 plane onto P4 as I have no native projection algorithms yet. It was all done in Cartesian with conversions from polysign to cart and back.
>
> When you decide what a function may be, then from that a trig, or a calculus can flow from that.
>
> For example the "total space" of a p3 function is a 4d space ?
This is a new term to me.
https://mathworld.wolfram.com/TotalSpace.html
Unfortunately Wolfram circularly defines base space and total space so that they each rely upon each other.
To actually have some data is helpful I would think. It's an interesting question whether a function f(z in P3) could spit out a P4 or a P5 result.
If f is composed of sums and products in P3, which seems traditional, then the result will be in P3. Somehow I guess you'd try to go back to one of those earlier sequence forms and get something convincing going. I don't think simply mangling signs is that convincing.
Possibly you'd have five P5 functions yielding magnitudes per sign?
I'm not feeling very strong about getting an interesting result here. It sounds like more mangling. Something two dimensional does not magically turn into something interesting in higher dimension. It's flat or its extruded through to the higher space. I don't even think the extruded option is on most people's radar. That's like dividing by zero and yielding everything. If for some reason your functions were convincing you might have something.
Here is a strange option: from P3 to P6 via the hexagonal form. A P3 hexagonal lattice goes through a polynomial and is morphed by the polynomial in strange ways. There are no hundred interconnects on the final interpretation; instead those morphed hexagonal points map into P6? Probably too complicated.
>
> For example if you feed certain function with -2+3*4, then you feed it with -3+4*5, and
> the output is different. Will you dismiss that function for not comply with the homogeneous standards.
> Then, will this disobedient function necessarily corrupt the definition of some derivative(or
> some other calculus artifact) ? Not necesarily. Things that do not work on the macro may work in the small.
Very good analysis. I suspect that it will work out without conflict. It almost seems as though the simplest of things like y=f(x) have been habituated into us in a way that is not analytical. Looking at the early works like Descartes as you mention does expose how complicated their analyses were.
Newton's too.