Agreed! We’ll never have error-proof methods so transparency is the best approach. To the extent that analysis using the flawed code was done reproducibly, it’s that much easier for an investigator to find and recompute flawed results.
-
-
-
That last line of the article,tho...
- Još 1 odgovor
Novi razgovor -
-
-
But publishing the code led to the problem! IMHO publishing of theory, equations and algorithm descriptions are key to reproducibility, not codes (which always have bugs).
-
Not to mention that as soon as the next Python comes out all research codes will break since the Python ecosystem does not respect backwards compatibility. So use a published research code 5+ years from now? (An instant in the lifetime of ideas...)
- Još 1 odgovor
Novi razgovor -
-
-
OPEN ACCESS is the only way. Don't publish unless you publish everything.
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi
-
-
-
Also, for programmers experienced in writing portable code, this particular glitch (if it is what it sounds like) is not very obscure (I've seen it more than once).
- Još 9 drugih odgovora
Novi razgovor -
Čini se da učitavanje traje već neko vrijeme.
Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.