It isn't really secret. We went back more than a few steps earlier in the year, which is one reason it would

not be useful at all in its current state. At the moment, it is not, by any means, a series of incremental improvements

that could simply be checked in to hg or published on a 9p server. At times it can go many days

before it can even be recompiled, let alone run, because old fundamental structures have

been re-organised. Large chunks are in flux. Although Nemo was working steadily

away, after some initial work, I was bound up with another project (that has nearly finished),

which meant I couldn't even review and consider some of the changes until early August.

Then there was an August break, although I did manage to do some work on it of interest to me.

Meanwhile, the other project I mentioned has changed my priorities, so I anticipate more churn.