Currently I am working on a design that has been my best learning experience ever as a knitter, no, as a pattern writer, no, as a learner. Fannie Fouche has been written and rewritten so many times I've lost track. I'm knitting the body for the third time. The Providence yarn (Ball and Skein) has been incredibly patient and resilient. I have reskeined it and soaked it in some Eucalan to remove the ramen noodle effect, and it is good as new.
I am deep, deep in the rabbit hole of armscye math : the construction, shape, grading and proportion. It is embarrassing to admit how fascinating it is. I had the pleasure of chatting with Amy Herzog recently (she was a guest presenter at our guild meeting), and she seemed to understand my excitement, that or she's a brilliant (and polite) actress.
I have written a pattern algorithm for myself, a series of questions that determine how many stitches and rows are needed at any particular location on a set-in sleeve sweater. The algorithm talks to the pattern template, which is written in CorelDraw : I can change the gauge in the algorithm, and all the numbers in the finished pattern change accordingly. See why I'm so thrilled? Tons of work up front, but the next time I want to write a pattern with this construction method it should go smoothly (she innocently and whistfully typed as the Pattern Gods chuckled heartily).
Fannie Fouche is constructed from the top down. It has set-in sleeves that are knit at the same time as the yoke. Why? Because we can. I am confident there are knitters up to the task. I am continually amazed at today's knitter : their Ravelry page states "just learned to knit last year", their project pages have fourteen pairs of lace socks, and they're on Sweater Seven of their Twelve-Sweaters-in-a-Year-Challenge ... in April.