Re: BioConductor MOFA2 needs basilisk - which installs a conda environment
On 20 September 2021 at 07:50, Andreas Tille wrote:
| On Sun, Sep 19, 2021 at 03:45:57PM -0500, Dirk Eddelbuettel wrote:
| >
| > | I'm thinking about a "fake-basilisk" like r-cran-bh or r-bioc-zlib.
| >
| > I had a similar thought. I am also friends with the author of it.
| >
| > They are doing 'The Right Thing' there for _their_ purposes at BioC in
| > abstracting OSs away and offering 'Python as a Service' for all setups.
| >
| > We have a much more controlled setup and _maybe_ we could splice in a
| > 'basilisk-lite' not requiring conda and all that crap. Would be worth a
| > shot, but is almost a new research project. I don't have time to dive into
| > this though. I could ask him if it ever came up before.
|
| It would be great if you could involve upstream into the discussion
| here.
Tricky. I think I take a pass. I said I am friendly with Aaron, I did not say
I would promise to pull him into a discussion he did not sign up for. What
they have does solve their problem -- we have to learn to accept that.
I still have a problem with us pestering upstream _for packages that are
scoring perfect at CRAN_ and have no errors there. Our dirty laundry (often
due to us mismatching packages) is not their problem. Anyway, I said it
before, nobody listened to it then so no hope today either ...
'basilisk-lite' is a really nice (research) idea and would be a seller on
cloud services. Maybe Aaron would even be interested in working on it. You
could ask him, his email is public.
Dirk
--
https://dirk.eddelbuettel.com | @eddelbuettel | edd@debian.org
Reply to: