3 ms·
An Empirical Study of Implicit Type Conversions in JavaScript
- jkaptur 11y agoUnder "Threats to Validity" the authors note that "some" of the code studied is generated, and that they tried to deobfuscate this code so they could understand it. I think this understates the danger to their results. I would expect most of the code on the top 100 websites to be generated/minified. I would further hypothesize that minifiers would only make transformations that are correct, but do not necessarily make code more readable, thus skewing these results towards correct type conversions that humans would not tend to write. It would be interesting to see these tests run over the minified and unminified jQuery source to see if there's any difference.