Python education survey

[Rick Johnson]

[...]

Conversely, why write an IDE into IDLE when perfectly-good IDEs
already exist? I don't use IDLE for development per se; it's for
interactive Python execution, but not editing of source files.

I believe the answer is two fold:

1. Including an IDE like IDLE into the Python distro helps noobs to
get started quickly without needing to traverse a gauntlet of unknown
IDEs on their own. If later they find something that they feel is more
appropriate; so be it.

2. (and most important to me... IDLE is written in Python using the
Tkinter GUI (which ships with python also). Therefore, the source code
for IDLE can be a GREAT teaching resource for how to write
professional Tkinter applications. I KNOW THE CURRENT SOURCE SUCKS!
However, we could change that.

So, with those points being covered, i believe IDLE is very important
to the Python community and could be useful to more people IF we clean
it up a bit. It's really a great little IDE with even greater
potential. If Guido would just say something (or at least some of the
top Pythionistas (Hettinger i am looking at you!)) this community
might work together to fix this problem.

 

[Nathan Rice]

[...]

Conversely, why write an IDE into IDLE when perfectly-good IDEs
already exist? I don't use IDLE for development per se; it's for
interactive Python execution, but not editing of source files.

I believe the answer is two fold:

1. Including an IDE like IDLE into the Python distro helps noobs to
get started quickly without needing to traverse a gauntlet of unknown
IDEs on their own. If later they find something that they feel is more
appropriate; so be it.

2. (and most important to me... IDLE is written in Python using the
Tkinter GUI (which ships with python also). Therefore, the source code
for IDLE can be a GREAT teaching resource for how to write
professional Tkinter applications. I KNOW THE CURRENT SOURCE SUCKS!
However, we could change that.

So, with those points being covered, i believe IDLE is very important
to the Python community and could be useful to more people IF we clean
it up a bit. It's really a great little IDE with even greater
potential. If Guido would just say something (or at least some of the
top Pythionistas (Hettinger i am looking at you!)) this community
might work together to fix this problem.

Not everyone who has Python installed wants to learn the language. I
do think that a "learning" distro that has a lot of core tools
pre-installed, and ships with some tutorials, would be a decent idea.
Sort of like Enthought for new users I don't feel IDLE is worth
salvaging though.

Nathan

 

[Rick Johnson]

]
Conversely, why write an IDE into IDLE when perfectly-good IDEs
already exist? I don't use IDLE for development per se; it's for
interactive Python execution, but not editing of source files.

I believe the answer is two fold:

1. Including an IDE like IDLE into the Python distro helps noobs to
get started quickly without needing to traverse a gauntlet of unknown
IDEs on their own. If later they find something that they feel is more
appropriate; so be it.

2. (and most important to me... IDLE is written in Python using the
Tkinter GUI (which ships with python also). Therefore, the source code
for IDLE can be a GREAT teaching resource for how to write
professional Tkinter applications. I KNOW THE CURRENT SOURCE SUCKS!
However, we could change that.

So, with those points being covered, i believe IDLE is very important
to the Python community and could be useful to more people IF we clean
it up a bit. It's really a great little IDE with even greater
potential. If Guido would just say something (or at least some of the
top Pythionistas (Hettinger i am looking at you!)) this community
might work together to fix this problem.

Not everyone who has Python installed wants to learn the language.

I'll admit you make a very valid point. But since "they" are not
familiar with the language, what they don't know cannot hurt them.
Although i do see a need to keep the core distro small. If we remove
IDLE then we face the next big question... Should we remove Tkinter?

 I
do think that a "learning" distro that has a lot of core tools
pre-installed, and ships with some tutorials, would be a decent idea.
Sort of like Enthought for new users

This is the kind of inventiveness i have been looking for in this
group. I think this is a great idea!

 I don't feel IDLE is worth salvaging though.

Agreed. Not in it's current state anyway.

 

[Steven D'Aprano]

If Guido would just say something (or at least some of the top
Pythionistas (Hettinger i am looking at you!)) this community might work
together to fix this problem.

The sheer cluelessness displayed here about open source is painful.

If Guido would "just say something", this community would yawn. He's not
the boss of us. Python development is an all-volunteer effort. "The
community" works on projects that we find interesting, or that we need
ourselves, not just because Guido "says something" about it -- and
certainly not projects that you demand we work on.

Instead of demanding that GvR order us to work on IDLE, put your money
where your mouth is, Rick, and stop being a blow-hard. You've been
nagging us about IDLE for long enough -- what is it now, two years?
Three? I haven't been counting, but it feels like geological ages.

Start a project to fix IDLE. There is plenty of free hosting on the
Internet for open source projects, like GitHub and Google Code Hosting.
Track some bugs. Plan some concrete enhancements. Write some actual code.
Demonstrate actual effort, and you may attract others who care.
Otherwise, you're wasting our time.

 

[Rick Johnson]

If Guido would "just say something", this community would yawn. He's not
the boss of us. Python development is an all-volunteer effort. "The
community" works on projects that we find interesting, or that we need
ourselves, not just because Guido "says something" about it

Believe it or not Steven, all communities need leadership. Just
because you are a rebel and don't like anyone trumping your opinions
does mean that the majority of people here don't need (or want for
that mater) some direction.Communities work best when leaders lead and
workers work. And "leader" does not always mean a single person. NO.

In the best communities leadership roles are dynamic. For example: In
such a diverse field as programming language design, no ONE person can
lead all the time! We need people to step forth who are competent in
the various areas of expertise. Someone who is really knowledgeable
writing code for Linux may not be as knowledgeable writing code for
windows. Some will be better at writing efficient algorithmic code and
others may be better at documentation. etc,etc...

HOWEVER

Even IF we adopt the intelligent paradigm of dynamic leadership, we
still need a small group (or at least) one very intelligent person to
bind the process together. Guido van Rossum IS that person. That is
why i call for him to get more involved. Of course i don't expect him
to write code. But at least he could be more involved in this
community. AND PYDEV IS NOT THE ONLY MAILING LIST IN THIS FREAKING
COMMUNITY!

He needs to step out from behind the coattails of his dev group and be
a real leader. He needs to make his opinion known about Tkinter, IDLE,
and also, the drastic changes that will need to be implemented in
Python 4000. How can HE expect a true FLOSS community to move forward
when SO many are just blatantly ignored? HOW?

-- and
certainly not projects that you demand we work on.

I am NOT demanding that people WORK on anything Steven. I am ASKING
for people to get more involved! Just open your mouth and state your
opinion (and not just troll all the time!). But as long as Guido and
the other members of pydev refuse to acknowledge the existence of
c.l.p, we will forever be a stagnate community.

Instead of demanding that GvR order us to work on IDLE, put your money
where your mouth is, Rick, and stop being a blow-hard. You've been
nagging us about IDLE for long enough -- what is it now, two years?
Three? I haven't been counting, but it feels like geological ages.

Start a project to fix IDLE. There is plenty of free hosting on the
Internet for open source projects, like GitHub and Google Code Hosting.
Track some bugs. Plan some concrete enhancements. Write some actual code.
Demonstrate actual effort, and you may attract others who care.
Otherwise, you're wasting our time.

I will do this, as soon as the "leaders" start to lead. Only a fool
would invest time to make a module or package "stdlib ready" without
getting the input of his fellow brethren first. Sounds like a recipe
for disaster to me. Why has Guido not, at the very least, contacted me
privately? He could remain anonymous.

 

[Chris Angelico]

Why has Guido not, at the very least, contacted me
privately? He could remain anonymous.

And how would you know if he did contact you anonymously?

As to your demand that one of the "top Pythionistas" [sic] say
something? I declare hereby that I am, in fact, a top Pythonista. [1]

Rick, go fork Python onto Bitbucket. Call it Python 4000 if you like,
or just "Python with a better IDLE" or whatever. Go! Shoo! Code!

There. A top Pythonista has spoken. And if the above satisfy you not,
I actually said it out loud too, just to be sure. That's how important
you are to The Python Community (capital letters and all).

ChrisA


[1] Start here and follow the thread. I've done all of them. (Actually
I probably haven't, and I haven't even read the whole thread, which
just came up when I googled 'pythonista', but I defy you to prove me
lying.) http://mail.python.org/pipermail/python-list/2003-June/207114.html

 

[Carl Smith]

I think ipython and a good editor gives a much nicer experience
than IDLE, which I actually almost never used, and
for everything else there is python and python-mode.

New users however can be pointed to something like PyCharm
or Eclipse+PyDev if they are more familiar to IDEs..

I agree; IPython is a excellent choice. You have a much more powerful
interactive Python experience, with all the features you need from an
IDE. You can use any editor (VIM) and you can also readily hack
IPython to death.

I think the fact that anyone with basic programming skills can
substantially enhance their console is a big winner in CS education.
It gives students something they personally value to work on, it's a
place to store all their little bits of code and actually benefit from
them in real life.

I've never met a programmer that got familiar with IPython and then
went on to stop using it. It should be included in the standard
library and used as the default Python interactive environment.

The last line of my .bashrc file:

ipython3

 

[Carl Smith]

Do you use IDLE when teaching Python?
If not, what is the tool of choice?

I believe IDLE has the potential to be a very useful teaching tool and
even in it's current abysmal state, i find it to be quite useful.

Students may not be experienced with the command-line and may be
running Windows, Linux, or Macs.  Ideally, the tool or IDE will be
easy to install and configure (startup directory, path, associated
with a particular version of Python etc).

Why install an IDE when IDLE is already there? Oh, yes, IDLE SUCKS. I
know that already. But this revelation begs the question... Why has
this community allowed IDLE to rot? Why has guido NOT started a public
discussion on the matter?

Though an Emacs user myself, I've been teaching with IDLE because it's
free; it runs on multiple OSes, it has tooltips and code colorization
and easy indent/dedent/comment/uncomment commands, it has tab
completion; it allows easy editing at the interactive prompt; it has
an easy run-script command (F5); it has direct access to source code
(File OpenModule) and a class browser (Cntl+B).

Yes, IDLE has all the basic tools anyone would need. Some people
complain about a debugger, but i never use a debugger anyway. I feel
debuggers just wreaken your debugging skills.

On the downside, some python distros aren't built with the requisite
Tcl/Tk support;

And who's fault is that?

some distros like the Mac OS ship with a broken Tcl/Tk
so users have to install a fix to that as well; and IDLE sometimes
just freezes for no reason.

And who's fault is that?

 [IDLE] also doesn't have an easy way to
specify the startup directory.

Are you kidding me? That could be fixed so easily!

If your goal is to quickly get new users up and running in Python,
what IDE or editor do you recommend?

IDLE, of course. But NOT in its current state.

Why would myself (or anyone) go to the trouble of downloading third
party IDEs when IDLE is just waiting there for us to use? I for one,
like to use tools that have open source code.  And what is a better
Python IDE than a Python IDE written in PYTHON? I ask ya?

Also, what is the purpose of this thread Raymond? Are you (and others)
considering removing IDLE from the source distro?

You know. Many folks in this community have known for a long time how
much i love IDLE, but at the same time how much i loath it's atrocious
code base. I also know for a fact, that many "movers and shakers"
within this community simultaneously use IDLE, and want to see IDLE
code improved. However. None of these fine folks have taken the time
to contact me privately so we can discuss such an evolution. Why is
that? It boggles the mind really.

Do people seriously use IDLE? I thought it was just there for
scratchers, like turtle.

 

[Eelco]

I agree; IPython is a excellent choice. You have a much more powerful
interactive Python experience, with all the features you need from an
IDE. You can use any editor (VIM) and you can also readily hack
IPython to death.

I think the fact that anyone with basic programming skills can
substantially enhance their console is a big winner in CS education.
It gives students something they personally value to work on, it's a
place to store all their little bits of code and actually benefit from
them in real life.

I've never met a programmer that got familiar with IPython and then
went on to stop using it. It should be included in the standard
library and used as the default Python interactive environment.

The last line of my .bashrc file:

ipython3

Youve got one here. I like IPython a lot, but it quite rarely enters
into my workflow.

While I agree that a good interactive python console is a good way to
get your feet wet with programming, I also strongly feel that a more
comprehensive programming environment should be introduced to
students. That includes opening and editing files, syntax
highlighting, and code completion. And painless installation. There
are no lightweight editors that provide all this functionality in
conjuction with Ipython*. So I prefer to work the other way around;
use something like pycharm, and open an IPython interactive session
within it.

*Your suggestion of VIM is especially objectionable. Though I am sure
it is a great tool to you, the subject here is beginner education.
Just because it is a good tool for you, does not make it a good tool
for a beginner.

IPython bundled with a lightweight but function-rich and non-hacker-
but-WYSIWYG editor would be a great choice. But until that comes
around, pycharm it is for me.

 

[Rick Johnson]

Do people seriously use IDLE? I thought it was just there for
scratchers, like turtle.

I know for a fact that many folks use IDLE, even some rather "well
known" folks around here. The fact is, more people use IDLE than admit
to using IDLE. Of course, out of respect, i will not mention their
names myself. Hopefully they will chime in..?

There is a stigma in this community towards IDLE. I am not sure
exactly how it started, or on what logic (or lack thereof) it is
based, but the stigma exists no doubt. I believe it may be just an
extension of the TCL/TK stigma (since IDLE is coded using Tkinter).

My logic is this:
""" Including an IDE in the stdlib may have been a bad idea (although
i understand and support Guido's original vision for IDLE). But since
we do have it, we need to either MAINTAIN the package or REMOVE it. We
cannot just stick our heads in the sand and ignore the elephant in the
chicken coop. It's bad enough to bloat ANY stdlib with seldom used
modules, but i dare say, it's far worse to bloat a library with seldom
used modules that are poorly maintained! Every module in Python's
stdlib is a testament to the skill and professionalism of this
community as a whole. When a module looks or works badly, we ALL look
and work badly."""

It's no big secret to anyone in this community that both Tkinter and
IDLE are the red headed step children of Pythons stdlib. Very few
people speak kindly of either package. I believe GvR still believes in
the merits of both packages (psst: batteries included!) but he finds
himself at odds with the elite at pydev. Hmm, Maybe he DOES want to
remove them but fears loosing face...FYI: Guido IS is the original
author of both Tkinter and IDLE... it was HIS idea after all. However,
it is high time to re-ignite the original vision and form these
packages into something we can be proud of -- OR -- cut our losses and
remove them forever.

This indecisiveness, collective bickering, and lack of true leadership
is shameful.

 

[Lie Ryan]

My logic is this:
""" Including an IDE in the stdlib may have been a bad idea (although
i understand and support Guido's original vision for IDLE). But since
we do have it, we need to either MAINTAIN the package or REMOVE it. We
cannot just stick our heads in the sand and ignore the elephant in the
chicken coop. It's bad enough to bloat ANY stdlib with seldom used
modules, but i dare say, it's far worse to bloat a library with seldom
used modules that are poorly maintained! Every module in Python's
stdlib is a testament to the skill and professionalism of this
community as a whole. When a module looks or works badly, we ALL look
and work badly."""

AFAICS, Python has a pretty good reputation even outside Python
community. I haven't seen anyone looks at Python badly because of IDLE.

It's no big secret to anyone in this community that both Tkinter and
IDLE are the red headed step children of Pythons stdlib. Very few
people speak kindly of either package. I believe GvR still believes in
the merits of both packages (psst: batteries included!) but he finds
himself at odds with the elite at pydev. Hmm, Maybe he DOES want to
remove them but fears loosing face...FYI: Guido IS is the original
author of both Tkinter and IDLE... it was HIS idea after all. However,
it is high time to re-ignite the original vision and form these
packages into something we can be proud of -- OR -- cut our losses and
remove them forever.

I hope you're not attempting to put words into his mouth, what you think
of Guido's ideas is not necessarily Guido's ideas.

In any case, IDLE is open source -- I don't know exactly what license
it's under but I'd assume it's the same as Python -- and anyone in this
list and outside this list -- including you -- can freely fork it and
work on it to improve it. In case you haven't realised it, it is pretty
much impossible for a large open source project to "die"; even if Guido
decided to remove IDLE from the standard library, it's not unlikely that
someone will fork it and maintain it as a third party application.

 

[Lie Ryan]

*Your suggestion of VIM is especially objectionable. Though I am sure
it is a great tool to you, the subject here is beginner education.
Just because it is a good tool for you, does not make it a good tool
for a beginner.

Before using VIM, I used to use gedit (and still do, though not as often
now); I don't think I've ever had any problem with not using a full
blown IDE with Python. I generally don't miss not using an IDE since
Python doesn't have the tradition of using overly verbose names like in
Java.

I'm personally of the opinion that beginners generally should start with
a simple programmer text editors (gedit is a good example). Firstly, you
don't want to confuse beginners with IDE vs language features; secondly,
at the size of the problem beginners typically had, they don't **need**
95% of those features; and if you teach beginners powerful IDE features
too early, by the time their problem gets big enough that the IDE
features would actually help, they'd already forgotten about them.

 

[K Richard Pixley]

Do you use IDLE when teaching Python?
If not, what is the tool of choice?

If your goal is to quickly get new users up and running in Python,
what IDE or editor do you recommend?

I would:

a) let the students pick their own editor.
b) encourage emacs and use emacs as a reference editor.

The problem is that IDLE is hard to set up. (I've never managed it and
I'm a well seasoned veteran). And pretty much only good for python, I'd
expect. You'd do better to encourage eclipse, but setting that up isn't
trivial either. You could create your own distribution of eclipse, but
then you have that "only useful for python" problem again.

If students are going to go anywhere else after this class, they're
going to need to either be able to learn to switch editors or find an
editor they can use more generally. Everyone ends up writing some html
eventually, for instance. Either way requires climbing a learning curve
that would be difficult to justify for a single class.

OTOH, there are binary emacs distributions for all systems you've
mentioned. And they work.

I'm an antimicrosoft bigot, but I think my answer is probably the same
regardless of whether we know the OS the students will be using or not.

--rich

 

[Rick Johnson]

In case you haven't realised it, it is pretty
much impossible for a large open source project to "die"; even if Guido
decided to remove IDLE from the standard library

I don't remember stating that Python would die if IDLE was removed
(not sure if you misunderstood me or you're just making a general
statement???). My belief is that the atrocious state of IDLE and
Tkinter's code bases are making us look bad as a community. And i can
go either way on the issue; remove them both, or enrich them both.
Either way we improve on the current situation. My point is that we
CANNOT just ignore the issue.

 

[Rick Johnson]

The problem is that IDLE is hard to set up.  (I've never managed it and
I'm a well seasoned veteran).

Can you qualify that statement? Do you mean "difficult to set up on
certain OS's"? Because for windows there is no difficulty.

And [IDLE is] pretty much only good for python,

Yes, i will agree on that! We could create snap-ins for other
languages, but, there are many good multi-language editors out there
already.

I'd
expect.  You'd do better to encourage eclipse, but setting that up isn't
trivial either.  You could create your own distribution of eclipse, but
then you have that "only useful for python" problem again.

Same boat, different lake.

If students are going to go anywhere else after this class, they're
going to need to either be able to learn to switch editors or find an
editor they can use more generally.
Agreed!

 Everyone ends up writing some html
eventually, for instance.

most folks write more than just HTML! Ruby, Lisp, Perl, C, Java, etc.

 Either way requires climbing a learning curve
that would be difficult to justify for a single class.

Yes, i must agree that IDLE does not scale. It's true that IDLE is
only good for Python. Could we make the IDLE usable for other
languages? Yes, but what good is that going to do? IDLE is already
slow due to Tkinter being slow, due to Python being slow... etc. I
must admit that IDLE is only useful for complete beginners, those who
code only in Python, or those who don't care about using a specific
Python IDE.

But that damn "batteries included" and CPFE keeps creeping up! What to
do, what to do?

 

[Andrew Berg]

You'd do better to encourage eclipse, but setting that up isn't
trivial either.

IIRC, all I had to do to set up PyDev was copy a URL to Eclipse's
"Install New Software" wizard, and have Eclipse download and install it.
Extra steps are needed if a different implementation of Python (e.g.
Jython) is needed, but other than that, the user only needs to specify a
couple options (e.g. Python grammar version) at project creation time.
This assumes that Python is already installed, but why wouldn't it be?

You could create your own distribution of eclipse, but
then you have that "only useful for python" problem again.

AFAIK, Eclipse should always be good for Java unless you do some serious
hacking.

If students are going to go anywhere else after this class, they're
going to need to either be able to learn to switch editors or find an
editor they can use more generally.

There are a ton of editors that have syntax highlighting and other
little features for many languages.

 

[Eelco]

Before using VIM, I used to use gedit (and still do, though not as often
now); I don't think I've ever had any problem with not using a full
blown IDE with Python. I generally don't miss not using an IDE since
Python doesn't have the tradition of using overly verbose names like in
Java.

As for my personal use, I very much prefer an IDE. I hate having only
crappy code completion, for starters, and I like a good integrated
debugger. But then again, im spoiled I suppose coming from C#. On the
other hand, ive worked for many years using a very minimal notepad
+command line compilation setup as well.

But I can very well imagine that people are perfectly happy with more
hackerish tools. That is, once they have gotten past the learning
curve.

I'm personally of the opinion that beginners generally should start with
a simple programmer text editors (gedit is a good example). Firstly, you
don't want to confuse beginners with IDE vs language features; secondly,
at the size of the problem beginners typically had, they don't **need**
95% of those features; and if you teach beginners powerful IDE features
too early, by the time their problem gets big enough that the IDE
features would actually help, they'd already forgotten about them.

A good IDE should get out of your way if you want it to. I like
pycharm in this regard; click an x or two, and you are facing just
your text editor and console/output window.

Generally, I think a non-cluttered IDE is ideal for a beginner. You
dont have to explain to them how to open a file, and if you tell them
to hit the 'play' button to start running their code (not a hard
concept to grasp or remember either) they are good to start hacking.
Also, I think code completion is a blessing to beginning programmers.
IPython is good in that regard; until you switch to editing files,
where your choice is between notepad, or configuring an other editor
you dont want your class to spend any time on, and youll end up with
something that still doesnt do much more than syntax highlighting.

 

[Rick Johnson]

Eelco, please don't get offended, but can you (and everyone else) stop
using silly verbage like "used to", "use to", "suppose to", "hard"
when you "difficult", and "pretty" when you mean "very". I find this
verbiage to be quite ridiculous. In this case you could have simply
said...

"""Before using VIM, I USED gedit."""

or if you want to stress that you don't use gedit anymore you could
say...

""" Previously i used gedit, but have since moved on to VIM."""

Thanks

As for my personal use, I very much prefer an IDE. I hate having only
crappy code completion

Syntax highlight is important ESPECIALLY if you're multi-lingual.
Sometimes when i switch between the many languages i know, the
highlight of the syntax is the only blindingly apparent clue. Take
Python and Ruby for example.

A good IDE should get out of your way if you want it to. I like
pycharm in this regard; click an x or two, and you are facing just
your text editor and console/output window.

A "good IDE" is nothing more than an extension of a base text editor
with options to switch on as little or as much IDE machinery as you
like. Why would you have a text editor AND an IDE? When an IDE is just
an intelligent texteditor? That is one thing that i find missing in
IDLE; the ability to turn off certain things. With IDLE, it's all or
nothing.

Generally, I think a non-cluttered IDE is ideal for a beginner.

Agreed, see last response ^^^

You
dont have to explain to them how to open a file, and if you tell them
to hit the 'play' button to start running their code (not a hard
concept to grasp or remember either) they are good to start hacking.

I always though "run" was a perfect verb for "running" code... but who
knows

 

[K Richard Pixley]

Can you qualify that statement? Do you mean "difficult to set up on
certain OS's"? Because for windows there is no difficulty.

The distributed binaries haven't worked when I tried on any version of
linux or macosx that I used. Attempting to build from source also
failed. (I'm pretty much an anti-microsoft bigot).

most folks write more than just HTML! Ruby, Lisp, Perl, C, Java, etc.

Dedicated computer folks and experimental tinkerers, yes, I concur.

But even casual computer users eventually want to write html, or compose
for a wiki and get annoyed when their 3 hours of work in a web browser
test field are lost because they changed pages.

--rich

 

[Lie Ryan]

I don't remember stating that Python would die if IDLE was removed
(not sure if you misunderstood me or you're just making a general
statement???). My belief is that the atrocious state of IDLE and
Tkinter's code bases are making us look bad as a community. And i can
go either way on the issue; remove them both, or enrich them both.
Either way we improve on the current situation. My point is that we
CANNOT just ignore the issue.

The point is, I didn't think it's such a pressing issue. I haven't seen
anyone in or outside Python communities making their conclusion about
Python based on IDLE.

In any case, removing IDLE without a much better replacement is pretty
much out of the question. If people installed Python in vanilla Windows
install, they would only have Notepad to edit their code.