A collaboration between Agnes Cameron and Tom Price.
The goal is to encode knitting patterns as .stp files compatible with
recent versions of Designaknit (DAK) so that
patterns written by other means can be used by people who have access
to that software.
The priority is making patterns usable for people with domestic-market knitting machines from brands like Brother and Silver Reed.
Progress has already been made on decoding the proprietary format (see DAKimport). There is an initial header of 0xF8 bytes whose important content is known. Following that are blocks of run-length encoded colour data whose contents reference a panel of colour data. This is followed by blocks of stitch data, encoded in a similar way. The raw data are obfuscated to produce a compact binary file.
Tom has written a Python module that can perform and reverse the obfuscation. Also, the stitch codes have been largely documented with the exception of cable stitches.
One thing that has not been investigated yet is how memo information is encoded.
For this to be useful it needs to be able to work with various different
input formats. I don't have firm suggestions how best to go about this.
The absence of a clear alternative is one reason why .stp files are a
useful format for publication and exchange.
There are various projects that have attempted to design open source pattern formats. The most useful are Knitspeak and Knitout because there are already people developing applications for them (see for example https://github.com/mhofmann-uw/599-Knitting-Complete).
I have come up with a format called Knitscheme
which is essentially a generalisation of Knitspeak to multiple yarns. It is
currently possible to input multicolour .png files. Both import and export
methods are available for .stp files. I am planning to write
a parser for Knitspeak in the near future.
My tentative suggestion is to use Knitscheme as an intermediate format, at least
initially, so that we can test pipelines from .png files and Knitspeak to
DAK. This would take advantage of existing pattern collections
(e.g. https://github.com/AllYarnsAreBeautiful/ayab-patterns, https://stitch-maps.com/patterns/).