i'm thinking of among the following during twitter's hackweek {sep 23-27}: (1) instant reload (e.g. ~1 second to recreate a pex and re-run it via

--loop

) for python run + jupyter notebooks running on pants code (which our data science teams use a lot, and it takes a long time to reload the notebook each time you change any python source code). this is definitely going to happen even if not during hackweek, and i'd really like for it to be something our team supports internally, so fingers crossed that it gets adoption! (2) improved rule graph visualization e.g. spreading the nodes more vertically, showing different relations between nodes (see https://github.com/pantsbuild/pants/pull/7024). i think that if i want to do this during hackweek that i want to finally go in on @happy-kitchen-89482's suggestion in that PR to incorporate something like d3.js or another existing way to interactively explore graphs -- worst case i will at least ensure that PR as it stands is merged. - also on the table is making error messages like the above message from @hundreds-father-404 a little easier to read (3)

--query

(see https://github.com/pantsbuild/pants/pull/7350). i think it will save twitter a ton of CI time if we can finish up the query pipelining work from https://github.com/pantsbuild/pants/pull/7356, so i might bump up the priority of this one to do during actual hackweek (then ask nicely for a promotion???). @hundreds-breakfast-49010 has been doing some incredibly cool work on the v2 engine recently so if any thoughts you might have (or anyone else!) on what might make developing v2 rules easier are appreciated! hackweek last year was a resounding success and I'm not planning to break the streak

better rule graph error messages is something I'd personally like to see