Our model of cancer keeps evolving and the history of the model itself might be important

(written by lawrence krubner, however indented passages are often quotes). You can contact lawrence at: lawrence@krubner.com, or follow me on Twitter.

This is an interesting talk:

“Four years of Datomic powered ETL in anger with CANDEL” by Ben Kamphaus and Marshall Thompson

They emphasize that as they learned more they had to adjust their model several times. One great strength of their implementation was that all of the behavior was based on “reflection” of the model, so when the model changed they only had to change a few lines of code, they did not need to re-write the whole codebase. This is an excellent example of where non-typed dynamic programming has an advantage over strictly typed programming.

Also, it strikes me that the history of the evolution of the model is itself important, as it reveals trends in the research. Understanding what surprised before might help us develop intuitions about what will surprise us in the future.

Post external references

  1. 1