The only reason we are avoiding notebooks are that we can’t test them in a unit test fashion, or we don’t know how. For this reason we are now leaning towards native Scala. I would have liked to have notebooks for reasons mentioned in the article but how do we have automated tests to ensure the notebook is actually correct?
Brilliant article. I have just over 15 years of experience and have had the privilege to work with some really smart developers that has taught me a lot. I’ve also been a mentor to junior colleagues as well as a teacher to school kids between 10–11 years old through www.codeclub.org. What I’ve found with a lot if junior devs are that they don’t make…
This sounds like an architecture/orm issue, not a .NET problem. You could have issued the same SQL from the c# code and avoided whatever was used that caused table scans.
I would love to do this with my team and have tried with a previous team (once) but the biggest resistance/question is around productivity. I understand why this should be as productive or in some cases more productive than the conventional approach. Do you have any statistics or insights from your own experience that can be used to convince management to try this?