> If you cannot the release the code because the code is too ugly or untested or has bugs, how do you expect anyone with the right expertise to assess your findings?
Yes, that should be this way.
Also all cases where some company research team goes to a scientific conference and presents a nifty solution for problem X without telling how it was purportedly done, it should be absolutely required to publish code and data for this.
*And that's also something which is broken about software patents - patents are about open knowledge, software which uses such patents is not open - this combination should not be allowed at all).
With the caveat that while in some cases, like computational science, numerical analysis, machine learning algorithms, computer-assisted proofs, and so on, details of the code could be crucial, in other cases, they should not matter that much. I too have the impression that the HN public tends to over-value the importance of code in these cases when it is mostly a tool for evaluating a scientific result.
Yes, that should be this way.
Also all cases where some company research team goes to a scientific conference and presents a nifty solution for problem X without telling how it was purportedly done, it should be absolutely required to publish code and data for this.
*And that's also something which is broken about software patents - patents are about open knowledge, software which uses such patents is not open - this combination should not be allowed at all).