[Tutor] Learning Objectives?

[Mats Wichmann]

On 02/27/2017 07:57 AM, leam hall wrote:

> When I was coming up as a Linux guy I took the old SAGE guidelines and
> studied each "level" in turn. It was useful for making me a well-rounded
> admin and helped me put off some higher end stuff I wasn't really ready
> for.
> 
> Things like Testing and documentation are useful, but only learning what
> seems to bee needed for this one project seems harder for the new coder.
> Most of us didn't know TDD was useful until we started doing it. Same for
> documentation. It's sort of the "if we hired a junior or senior coder, what
> basics would we want them to know?"
> 
> Thanks!
> 
> Leam

Just as a suggestion,

I think if you look at the curriculum of various training courses,
books, online tutorial series you will get each author's view of what a
logical progression is. There are probably as many different
progressions as there are people expressing an opinion, but it seems to
me the flow is at least roughly aligned when I glance at these from time
to time.

For example (this is no endorsement, just a site I do peer at now and
then and so remember it), look here:

http://www.python-course.eu/python3_course.php

the left bar has a kind of order that's not insane (except I'd be
looking at files in the first 10 minutes, but that's just me); then
there's a tab called "advanced topics": forking, threads, etc. and then
some stuff you look at if you need it: wsgi, mod_python, sql connectors.
And then another tab for a great big "external" topics like NumPy,
Machine Learning and there are lots more in that category like Django,
Flask, SciPy, and so on - all can be considered "advanced" because
they're not the language itself, but stuff some people might expect you
to be proficient in for a given job that's billed as a "Python Job".

If you're looking at abstract base classes in your first few days,
that's probably not the right order ;)

I even have a bit of a bone to pick with the TDD movement, not that it's
a bad idea by itself, but when treated as a "religion" it seems to lead
to a mindset that unit tests alone are sufficient to prove a codebase,
when integration testing is just as important.