Hi all,
Please try to be respectful with each other. I think perhaps Gonzalo
tried to be a bit pushy about his opinions, which didn't provoke the
best reaction among the others.
My opinions and decisions about this thread are the following:
1. Anaconda
Perhaps this was the most controversial of all of Gonzalo's proposals.
My decision (as the current maintainer) is to remain vendor neutral, and
not to support any scientific distribution above the others. This is my
rationale:
* We don't have any hope to push something like this on Linux
distributions. If we'd try to, Linux packagers would simply patch Spyder
and remove the code related to Anaconda.
* I think this would be true too for source-based Mac distributions like
MacPorts and Fink.
* There is still a lot of effort and work going on on Windows
distributions like WinPython and PythonXY. I've heard they are used on
several workplaces, and not just by individual programmers. So, we don't
have any right to hinder them by trying to impose what we think is the
"best solution". At the end they would probably patch Spyder too to
remove the "problematic" code.
However, I want to state the following:
* I'm not opposed to improve Spyder integration with Anaconda, when
Spyder is running *inside* Anaconda. That's why I accepted the conda
package manager plugin, that some of you have seen already. But let me
assure you that this plugin will only be shown when we are 100% sure
Spyder version is an Anaconda one, and not in any other circumstance. We
won't promote it in the "Optional dependencies" dialog now anywhere else.
* To be consistent, I'll add more references to Anaconda in our "About
Spyder" dialog and Help menu. At the same time, I'll remove the actions
to use the WinPython package manager inside Spyder, and (I think) others
related to PythonXY. I had left them until now out of respect for
Pierre, but now that Spyder has grown beyond these distributions, I
think it's ok to remove those references.
2. Templates
I agree with adding more code templates (e.g. in the Source menu). We
could have scientific, data science and other kind of templates, which
would be really cool!
But as Adrian and I explained clearly, we can't impose any of those
templates by default. That would give the false impression that what's
in the template is *necessary* to program in Spyder (and we know that's
not true).
I don't have time to work on this, but as always, pull requests are
welcome :-)
3. PEP8
I'm really opposed to enable it by default. Look for example at this code:
1: # -*- coding: utf-8 -*-
2: """
3: Created on Sat Feb 7 19:35:42 2015
4:
5: @author: carlos
6: """
7:
8: def foo():
9: print("Hello, world!!")
10:
On line 8, PEP8 reports this incomprehensible error: "E382 expected two
blank lines, found 1". And if you forget line 10, then you'll see
another error. It's easy to understand what those errors mean when you
know what PEP8 is, but at first sight they seem like syntactical errors,
when they're really not! :-) That's why I mentioned this would confuse
beginners a lot.
However, I'm not opposed to nudge people in the PEP8 direction by
promoting it more. We already mention it in our tutorial, and we could
also mention it in one of our (soon to come) interactive tutorials.
There we could explain people what PEP8 is and why it's really useful.
We can also make the option more accessible, by putting it in the Source
menu, and in a more visible place in the Editor Preferences page.
4. Object Inspector
I already explained my rationale about not to automatically connect any
plugin to it by default. So far, I haven't seen complaints about it
(well, some users say automatic connections are not working correctly,
but I don't know why :-). So I'll take that as a sign that I didn't go
wrong with my decision.
About its speed: I made some tests and found that the first rendering
(of any page) can be quite slow, but after that the Object Inspector is
as fast as it always has been. If you are seeing a different behavior,
please use "hg bisect" to tell me when and where a speed regression was
introduced.
Well, I think that's all :-)
Cheers,
Carlos
El 02/02/15 a las 10:27, Adrian Klaver escribió:
On 01/28/2015 12:06 PM, Gonzalo Peña-Castellanos wrote:
Hi Adrian
Thanks for the answers, sincerely the idea of these posts is to explore
opinions, because that is at the end all we have, from these opinions
interesting ideas come up, even if you (or others) disagree with my
opinions.
/"Seems to me that fixing these is bigger priority then starting on a
campaign that will only add to that list."/
I am not starting any campaign, I am giving an opinion and asking for
more.
I do not see the difference. The initial post had specific solutions
outlined, with a question tagged on the end. If this was really about
opinions the post would have only asked, what do you the end user want
from Spyder?. Instead a series of things where presented on which work
had already started and for which forgiveness was being asked. This
does not represent community to me.
/
/
I am well aware of the number of issues and I am taking the lead in
solving
some of them.
And it is appreciated, but lets stick to that before creating a new
set of issues, for instance
https://groups.google.com/forum/#!topic/spyderlib/4m2XJNVLYWQ
/
/
/"If you want to help Spyder, then I would suggest working on the
Project/
/manager."/
This is what I have been working on since the past months (besides the
other minor PRs I helped with). A project manager would work best
if you could handle different python versions and different package
versions
as needed. virtual env and pip could do for python only packages, and I
have some work in this direction. For nonpython things, well conda+pip
work pretty well and I also have work in this direction.
As long as these are kept as strictly optional plugins. I do not
relish the thought of an IDE I am using in Linux, carrying Windows
baggage to enable it to work in that developer unfriendly environment.
Thanks again :)
--
You received this message because you are subscribed to the Google Groups
"spyder" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/spyderlib.
For more options, visit https://groups.google.com/d/optout.
