Leave it for compatibility and then make a new linear variation called LinearMo or something like that.

--
My Favorites: [link]
My Gallery: [link]

The forward compatibility (Apo->Meta) will be preserved in all cases. The backward compatibility as well - as long as you don't change the variables

We can also have overall variables for all variations: before_a, ..., before_f, after_a, ..., after_f

But I'm not sure whether they are needed so I made this poll

--
Home of Apophysis 7X - get it here
You can of course also take a free tour through a gallery of mind catching abstracts
And never forget: The is a lie!

Hi,

If you do make it into mobius that would mean we could use the other mobius at the same time

--
[link]

'All you need is love' - John Lennon

[link]

I think the more variables, options, and control you can give the user of the software the better.

I love the idea that Meta will have layers much like Ultrafractal.

--
My Favorites: [link]
My Gallery: [link]

I assume Meta won't be flam3-compatible, so it would make sense to combine linear and mobius Regarding the julian2 a-f variables, aren't these merely duplicating the transform coefficients? I can see that 2 distinct operating modes would be possible, graphical or numerical according to preference. One way would be to put the xform coeffs under the variables to avoid duplication.

--
ID
~~~~~~~~~~~~~~~~~~~
UltraGnosis Fractal Art

Well, Meta should be mostly flam3 compatible. The new "linar variables" will be ignored in flam tho.
The julian variables are applied to the polar-like coordinates which are calculated in julian. So its a step within the julian calculation itself and neither equal to the "normal" coefficients nor to the post coefficients Thats the main sense of this stuff in julian2
But you right, those work exactly like the triangular coefficients:
x' = a * x + b * y + c
y' = d * x + e * y + f
OR
x' = s * (cosA * x - sinA * y) + x0
y' = s * (sinA * x + cosA * y) + y0

--
Home of Apophysis 7X - get it here
You can of course also take a free tour through a gallery of mind catching abstracts
And never forget: The is a lie!

I make it >like< mobius. Not exactly the same as mobius. You handle it the same way but these are extra coefficients you can also express as triangle. I think about giving the variations a "has custom coefficients" flag which is useful for mobius tho. You can control the mobius variables also with a triangle then. But this is only a concept and it will not be in the coming alpha

--
Home of Apophysis 7X - get it here
You can of course also take a free tour through a gallery of mind catching abstracts
And never forget: The is a lie!

If you combine an affine transformation, spherical and post-transformation you get a mobius transformation. Why complicate linear() with something that isn't a linear transformation?

Whether to do this or not is the question of this poll
I start to agree that it would not make much sense with linear but what about the "extra coefficient set" flag for Meta variations? Julian2 is the best example to explain this. You can control it using a triangle if this is implemented. Or we take mobius and let the user control the variables with a triangle.
Maybe we can also map triangle points to six variables?

--
Home of Apophysis 7X - get it here
You can of course also take a free tour through a gallery of mind catching abstracts
And never forget: The is a lie!