There were no set literals in Python 2, historically curly braces were only used for dictionaries. Sets could be produced from lists (or any iterables):
set([1, 2, 3])
set([i for i in range(1, 3)])
Python 3 introduced set literals and comprehensions (see PEP-3100) which allowed us to avoid intermediate lists:
{1, 2, 3}
{i for i in range(1, 3)}
The empty set form, however, was reserved for dictionaries due to backwards compatibility. References from [Python-3000] sets in P3K? states:
I’m sure we can work something out — I agree,
{}for empty set and{:}
for empty dict would be ideal, were it not for backward compatibility. I
liked the “special empty object” idea when I first wrote the PEP (i.e.,
have{}be something that could turn into either a set or dict), but one
of the instructors here convinced me that it would just lead to confusion
in newcomers’ minds (as well as being a pain to implement).
The following message describes these rules better:
I think Guido had the best solution. Use
set()for empty sets, use{}
for empty dicts, use{genexp}for set comprehensions/displays, use
{1,2,3}for explicit set literals, and use{k1:v1, k2:v2}for dict
literals. We can always add{/}later if demand exceeds distaste.