Most of the python packages are honestly... kind of terrible to maintain. Sure, it is easy to whip up a python script to answer your computational question, but maintaining that is extremely hard - as evidenced by the shear number of abandoned python programs written by graduate students that can no longer be run. I used to care about building things quickly, but now I just care about maintenance of things. Maintenance in computational biology is awful right now.
Are any end users willing to pay for comp-bio software tools? Or the professional support of open source tools? I understand academic labs preference for free/open-source software, but there are lots of biotech companies out there as well.
Seems like there is some funded software in this space, and lots of academic research code of varying quality - sometimes very useful and I've certainly appreciated it. But also common are many shortcomings: usability, performance, integration with other tools, packaging & distribution to users, docs & training material, abandoned tools, etc.
Maybe the use cases are too diverse, with the common needs having evolved good open source solutions, leaving a constant uneven frothing of other bits of software being born and then declining for the all the other specialist needs. Or something. Still, I wonder if it could be better.
I feel this in my bones. Not only that but discoverability is terrible partially due to all the dead ends you can go down when looking for some library or solution, only to find out they don't work anymore.
Most of the python packages are honestly... kind of terrible to maintain. Sure, it is easy to whip up a python script to answer your computational question, but maintaining that is extremely hard - as evidenced by the shear number of abandoned python programs written by graduate students that can no longer be run. I used to care about building things quickly, but now I just care about maintenance of things. Maintenance in computational biology is awful right now.