On Sat, 28 Feb 2015 17:36:44 +1100, Steven D'Aprano wrote:
> Dan Sommers wrote:
>> And thank goodness for that! I've been writing Python code since
>> 1997 and version 1.5.<something>,¹ and I still do a double take when
>> emacs colors all my ids that faint blue that means "builtin."
> Although it is not helpful for people using screen-readers, and may be
> of limited use to the colour-blind, I am in favour of colourising
> built-ins so they stand out.
Now if only emacs were clever enough *not* to colorize "id" when it's
one of my names and not the builtin... ;-)
> On the other hand, I recall seeing an editor which rejected the idea
> of colour-coding built-ins, keywords etc., instead it coloured your
> own variables. So given:
>
> spam = 23
> eggs += cheese*len(sausage)
>
> spam, eggs, cheese and sausage would be different colours. The idea
> being, when scanning a large code base, all the places that use a
> specific variable would stand out ("just look for the dark green
> word").
As a mostly visual person, I can see (pun intented) the logic and the
value in that. I wonder how many variables could be easily
distinguished, though, before running out of easily distinguishable
colors. Then again, a clever underlying algorithm might choose colors
based on *dissimilarity* of the identifiers, so that "i" and "j" would
be very diffent colors, but "spam" and "throat_warbler_mangrove" could
be the same color because they look so different anyway.
>> I don't think I've ever used the builtin function id in a program.
>> Ever. Not even once. Honestly, what is a valid use case?
>
> Here's one. I think it's the only time I have seen id() used apart from
> interactive experimentation:
>
> https://code.activestate.com/recipes/577504
Hah. Very nice. That sort of thing is probably useful for detecting
self-referential objects, too (e.g., to prevent infinite output for a
circular list).
>> ¹ No, not continuously. I have eaten and slept since then.
>
> Slacker!
Sorry. I'll make up the hours later, I promise!
--
https://mail.python.org/mailman/listinfo/python-list
