[Disclaimer: 25+ years in <thebiz:://>. Started with Oak and ECMA-(-1)]
Lol. You are very diplomatic compared to the full-stack professionals I interact with every day.
- I agree with everything you say. We don’t agree in context.
- My title was not click-bait but perhaps inflammatory.
- I admit I was vague. I should have stated out-right the article documented the journey of an ML guy. I wanted to show hope that they could put out a URL GUI for their ML predictor and did not have to wait for the full-stack team to put it out.
- I stated (it maybe was buried) that Facebook could not use Streamlit.
- By the 3rd installment, Streamlit can not yet replace Flask; I stated that security (aka encryption and secure login) was lacking. That is, don’t use it outside the firewall.
- After suffering as a pre-mature adopter in the 2nd half of the ’90s, I was enthusiastic about Streamlit.
- Having programmed in assembler, that is what is felt like to me. — the simile where I compare apples to tanks.
- We only use streamlit internally to put out beta-level dashboards. Production is in the full-stack over-sight.
- Wow! I hope I did not imply that you should use SttreamLit to build microservices. If so, that is wrong. What I meant to say was that we combined several Flask-based microservices into a StreamLit app. I find the designer appreciates microservice building blocks, but the impatient user does not like the sub-second slowdown. If anyone has a work-around for this, please let me know.
- (IMHO) I think (hope) that tools will appear to relieve the Data Engineer and full-stack engineer of the low-level boiler-plate and enable them to spend more time on solutions.
— — -