Ian Holmes

Yes, I was dragging on the location bar. Apparently I do not know how jbrowse works. Closing in shame.

Thanks @rulatir and thanks for the clear test case. I have added a note to the README that this feature is missing. Not sure at the moment when we'll get...

It shouldn't be blocked by that, per se. The Makefile parser could bring all that stuff together. It's more just a question of introducing these new assignments into the Prolog...

it shouldn't be that hard to record metadata, just look for where the MD5 hash is updated and stick another hook in there.

@cmungall I like this, especially how clean the mapping to PROV-O is: I think most/all of those things in the diagram are already being calculated at some point in biomake.

The Python example would give almost-free SnakeMake compatibility (modulo one parser). This, with a couple other features like provenance, seems noteworthy enough for publication (to me at least), which is...

yes, biomake currently leverages the filesystem far too intimately; it assumes that the subdirectory is there, to stash its internal files in the hidden directory `.biomake` This is certainly ugly...

hmmm, now i am wondering if that example I gave _would_ work... actually it might still run into the same issue. going to think about this some more...

Perhaps the best thing is just to do the `mkdir -p` implicitly after all.... hmmmmm..... any thoughts @cmungall @sjackman ?

@sjackman The problem with that is that the biomake directory for target X would then depend on which directory biomake was invoked in, as well as the location of X....