# Re: patternTransform -- and toroidal wraps

• From: Dr. Olaf Hoffmann <Dr.O.Hoffmann@gmx.de>
• Date: Tue, 4 Oct 2011 18:30:19 +0200
• Message-Id: <201110041830.19213.Dr.O.Hoffmann@gmx.de>
```Ken Stacey:

>Successive tiles can paint over the overhanging content of previous
>tiles.

A method to conserve the symmetry intended by the author is typically another,
because this method obfucates typically non rectangular symmetries and will
often ofuscate some rectangular symmetries as well, replacing them with a
lower symmetry group, the author has typically no interest in for his use
case, therefore it does not really solve a problem for the author at all ;o)

As a typical example if you have pattern of p3... or p6... type, you will run
into a symmetry problems, if you do the painting for each tile separately,
what basically means, that such an approach excludes the usage for a
larger amount of pattern groups ...

What should work:
If the 'master template' consists of the elements A, B to X (in this order)
one can use the following  method:
1) do the complete pattern for A fill and apply fill-opacity to the result
2) do the complete pattern for A stroke and apply stroke-opacity to the result
3) apply opacity to the result
4) continue first with B, then C up to X, do correspondingly 1)  to 3) for
each of them and put the B result on top of the A result, C on B and so
on.

I think, this will conserve the intended symmetry for all groups and
typically will avoid major problems with clipping gaps artefacts.

But there are other methods possible too, for example combine the
tiles for each element of a master to some kind of super path. And paint
the super paths in the painting order of the 'master template'  elements.

In general one has to avoid, that the creation of the pattern with
tiles breaks the intented symmetry. Else the method is not meaningful -
sure, one can define it, but then pattern is only useful for a subset of all
existing pattern types.

Olaf
```

Received on Tuesday, 4 October 2011 16:30:50 UTC