On 18/04/20 4:18 AM, dcwhat...@gmail.com wrote:
Yes,personal preference is definitely a factor and a bias, in these matters.
But aside from trying to make it easier for future A.I. to figure out what the
heck we humans were doing, it also makes a difference in how the IDE interpets
the code.
Maybe it isn't true for all IDE's or all languages. (I know SOMEONE will
interject here, to argue for the sake of arguing). But when I worked with
Groovy in Intellij about 5 years ago, there were times when the IDE was
confused, during a debugging sessions. I don't remember the exact details, but
that anomaly happened only with DEFed variables ; it didn't happen when the
data type was specified.
I am going to interject here - and because there is arguing!
1 the above description seems to suggest that the coder's objectives are
now to be split: firstly to make Python understand his/her instruction,
and secondly to ensure that the IDE knows what's going-on.
Does the IDE serve you, or do you serve the tool?
- collaboration notwithstanding
- similarly, without considering UX (users what users?), etc!
2 AI's creative processes may initially, in a macro sense, be modeled on
human cognition; but when it comes to operation, their methodologies are
quite different. Why would a computer, programming a(nother) computer,
use a language/tool developed for human-computer communication?
3 The core-point seems to be, that without increased detail in the
Python code, an IDE or AI will have difficulty understanding (my/our)
human-intent. This is a feature/failing* of all "communication" (a
component of "transmission" and/or "noise").
For example, please review the original post. Can you show me where it
even hints that a response should consider anything other than *human*
understanding? (ie your own)
* 'trickery' aside, how many students/trainees have you heard utter
phrases such as, "you didn't tell [me|us] that"?
4 As previously mentioned: [typing, hints, PEP, Python, ...]. If you'd
care to build a tool which makes use of typing, that is the current
philosophy/invitation (per earlier comment and refs). Pending such a
tool, should we (all) start writing Python-with-typing in the
expectation of the requirements of some future hopes, or 'vapor-ware'?
(see also mypy)
5 I like the idea (any idea) of something that will make (my) life
easier (I've had to work very hard to be this lazy!). Thus, a tool (IDE
or whatever) that will do 'stuff' for me, or save me time, eg by saving
me from my own coding-mistakes; is bound to attract my attention.
Similarly: speed of coding, improved readability for others, ...
One of the joys of Python is "for". We (still) tend to call it a
"for...loop", but to be correct it should be something like a
"for-all...loop"! The traditional for...loop goes right back to early
"high-level programming languages", eg FORTRAN's DO-loop. The reason
Python has its (different) approach is because so many
programming-errors relate to "pointers" and their oft-related
"out-by-one errors", ie
for pointer from 1 to 10 # this is NOT Python!
print( array[ pointer ], array[ pointer } ^ 2 )
What if the "1" or the "10" is a variable, and the programmer fails to
keep them related/limited to the size of the array? What if the array
was populated from a zero-th element? ...
Instead, Python says:
for element in listNM:
print( ...
Easy - doesn't matter how long/short the list may be!
What's the point here (for you)?
Much analysis has been performed on programming, algorithms, code
structures, the psychology of computer programming, etc. Certainly, our
BDEVIL did when designing Python! Rather than (apparently!) starting at
the syntactic-level (with data-typing), might your research benefit from
looking at such meta-observations, and working from there? .
Alternately, perhaps Python does not lend itself to this sort of
research? Might another programming language might be a better choice?
The designers of the Rust programming language appeared to take a quite
different approach from Guido Van Rossum. A (rather shabby) comparison
between the two languages shows one taking the line 'we are all adults
here', but the other attempts to be 'secure' (in several definitions of
the word).
Web.Refs:
http://mypy-lang.org/
https://www.python.org/dev/peps/pep-0401/
--
Regards =dn
--
https://mail.python.org/mailman/listinfo/python-list
