> Salgado remembered [adding '@'] when “we actually changed the language to help these communities… that was mainly for the data science people,” where this operation is very commonly used.<p>For a bit of ancient history, the first time this happened was back in the 1990s for 1.4beta2. From Misc/HISTORY:<p><pre><code> - Changes for use with Numerical Python: built-in function slice() and
Ellipses object, and corresponding syntax:
x[lo:hi:stride] == x[slice(lo, hi, stride)]
x[a, ..., z] == x[(a, Ellipses, z)]
</code></pre>
That one change, IMO, made Python significantly more useful for numerical computing and data science people.<p>I think "@" is the second time this happened. Did I miss one?<p>(Perhaps third, if we count adding the __index__ method from PEP 357.)