[Idle-dev] hi, newbie questions

[phil jones]

On 2/9/07, Tal Einat <taleinat at gmail.com> wrote:

> Not quite - sorry to have been vague. I think IDLE should -NOT- be a
> full-blown complex IDE, with support for such things as version control and
> projects.

OK, gotcha

> I do think IDLE could be far better at what it is meant to be - a good
> interpreter with a nice editor, simple to learn but powerful enough to use.
> You might say that this is more-or-less how Python is often described, which
> is precisely why IDLE should be that way too.

> > Is there any tutorial
> > documentation explaining how IDLE is put together, and showing how to
> > extend it?
>
> Not that I know of. The code itself is also quite poorly documented in many
> places.

I noticed  :-)

> (There is a short guide to writing IDLE extensions ( i.e. plug-ins), but
> that's not what you meant...)

 > What's MVC again? ;)
>
> IDLE is very, very dependent on Tk, especially on the event binding
> mechanism and on the workings of the Text widget. MVC was not used when
> writing IDLE.

OK.

> I don't want to take IDLE to new places. Like I said - I want it to be a
> truly awesome Python shell, with a usable editor, easy to learn and use for
> newbies. Perhaps in the future I'll try to make a variant with a more
> powerful shell, or a new shell altogether - but that's for the future.

What do you think of PyCrust etc? Do you think the shell should be
more like these?

> > Yeah, seems like it shouldn't be that difficult to make more of it.
> > And I think being bundled with Python is a great opportunity.
>
> Mostly, I mean IDLE being a great Python shell, with potential to be greater
> being built with a GUI toolkit (unlike IPython), and being 100%
> cross-platform.

re: fragmentation

> You could say the same about Linux and the much larger fragmentation there.
> But fragmentation is natural, and the selective forces in the Open Source
> world are much weaker, so the fragments take a long time to die out / merge
> together. The free Python IDEs just don't have enough of a driving force
> behind them.

yep, and OK, I grant that pluralism is a good thing

> One reason for this, worth pointing out, is that people don't know what's
> good for them. Programmers don't know what to ask for in a good IDE. And
> even worse, some experienced programmers do know what's good for them - and
> it's vastly different for each one. Outcome - you don't get a strong driving
> force behind a single IDE.

yes.

> I think WingIDE is in the right direction, it's supposed to be very good -
> but it's not open-source and costs money, which is a problem for most
> Pythoneers. But I think making a really, really good Python IDE, which costs
> a bit of money, is a good way to get the resources needed to build a good
> IDE.

suppose so. I tried WingIDE demo and didn't like it all that much. I
prefer Stani's - and not just because it's free.

My issue with plurality is not *choosing* what I want. It's installing
and managing. I use four different PCs in different places, and carry
the source-code around for my projects on a pen-drive. I want to stick
as much as possible with out-of-the-box python because having to
install and upgrade various external libraries on multiple machines
gets boring pretty quickly. And especially as my laptop is not always
connected to the net,

> As for Rails, it makes me understand how important popularity and "buzz" is.
> Popular frameworks gain momentum quickly since there is more support for
> them, more job opportunities for someone who knows them, more potential
> employees for a company with a product based on them, etc. Not necessarily
> because they get the job done better/faster than the alternatives.

Sure. My point was not that Ruby or Rails is cool. My point was that
having a development framework that lets you get on and produce stuff
quickly is a great attraction, and if you pull in a larger community,
more people drive things forward faster.

BTW : never really used Ruby, but one thing that impressed me a couple
of years ago was how I managed to download and install Instiki (a Ruby
wiki) on my office intranet very easily because of the Gems thing.
Python Eggs seems like it should make searching and managing other
packages easier. But I don't see any support for them in IDLE.

Do you think this kind of thing : eg. a menu option for downloading
and installing eggs via IDLE is considered outside the scope of what a
tool like IDLE should provide?

> Yea, it would, wouldn't it? They way it is now, you actually have to decide
> for yourself what's best for -your- needs, and everyone hates that! *wink*

(See above ... also, I want to make software for people who don't,
themselves, know how to download the appropriate supporting libraries
and packages)

> I do agree that there aren't good and easy enough tools for developing
> desktop applications with Python right now. Heck, it's what I've been doing
> for the past 2 months, and I can say that wxGlade doesn't meet the grade.
> This really is why there are so few apps written in Python, compared to VB.

But Python had such *potential* here ... :-(

Tkinter wrapping other lib.

> I don't think so - Tkinter is a thin wrapper around Tk. There is no way to
> wrap wxWidgets, for example, with a Tkinter-like API.

Ah well

> I think wxPython is slowly gaining momentum, with the lack of a WYSIWYG GUI
> builder holding it back. As it is, wxPython is great if you like to write
> your GUI by hand. But I don't see it entering the main CPython distribution
> any time soon.

Any reason why not? Too heavy? Political reasons? Too much investment in Tk?

> What I think is more plausible than having wxPython distributed with
> vanilla-CPython, is that someone will create a Python distro that comes with
> a GUI toolkit (e.g. wxPython) and a sound application development
> environment.

yeah, some kind of one-stop solution would be nice.

phil