Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.lang.python > #39137 > unrolled thread

Re: Python Warts: The where, when, how, and why of a PyWart.

Back to article view | Back to comp.lang.python

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.

Contents

  Re: Python Warts: The where, when, how, and why of a PyWart. Terry Reedy <tjreedy@udel.edu> - 2013-02-18 17:31 -0500
    Re: Python Warts: The where, when, how, and why of a PyWart. Rick Johnson <rantingrickjohnson@gmail.com> - 2013-02-18 15:55 -0800
      Re: Python Warts: The where, when, how, and why of a PyWart. Terry Reedy <tjreedy@udel.edu> - 2013-02-18 22:22 -0500

#39137 — Re: Python Warts: The where, when, how, and why of a PyWart.

On 2/18/2013 1:04 PM, Rick Johnson wrote:

>    This thread is actually an extension (of sorts) to a thread
>    started by "Anatoly tecktonik" back in December 2012; posted on
>    the python-ideas mailing list; titled: "Documenting Python
>    warts on Stack Overflow".

This was a threat to abuse StackOverflow with off-topic posts if 'we' 
did not pay him more attention.

>    I don't feel that python-ideas was the proper location to
>    discuss this issue (and neither did the BDFL)

Correct. I said so first, though Guido saying so later had more effect ;=(.
http://mail.python.org/pipermail/python-ideas/2012-December/018414.html
I also said that the post was obnoxious and best ignored. You dirty 
yourself by associating yourself with that post.

I will just briefly repeat two points:
1. we already have a searchable list of Python problems -- the tracker; 
2. everything that Anatoly complained about has already been discussed, 
sometimes ad nauseum.

His real complaint is that he tends not accept and respect that other 
reasonable and intelligent people disagree with him and that we develop 
Python for all users, not just him.

> Anatoly, I've seen your posts on python-ideas regarding this issue
 > and wanted to inquire as to your progress so far.

Anatoly has made progress in restraining his most obnoxious behaviors, 
such as repeatedly re-opening closed tracker issues, and otherwise 
trying to divert developer attention from issues we can and will fix* to 
his pet dead issues. You would do him a disservice if you encouraged him 
to regress in that respect.

However, moving his diatribes to python-list from the tracker, pydev, 
and python-ideas *would* be progress, and I hope you succeed in 
encouraging him to do that. I have so far failed.

*The mostly volunteer developers have closed an average of perhaps 5 
issues a day for the last 2 years that I have been watching. Quite 
amazing. A few are rejections of proposed enhancements or behavior 
changes, but most are fixes and improvements.

 > Also, you should not be surprised that many
> in the Python community do not want a searchable list of warts available;

Such a lie. Issues on the tracker are *not* removed. Neither are posts 
on the mailing lists.

Rick, it is one thing to say you don't like some feature of Python -- I 
am sure you don't -- and wish it to be changed -- I am sure you do. It 
is really quite another to lie about people who volunteer their efforts 
to improve Python. The more you do this, the most you encourage 
developers to ignore python-list.

> But let's not fool ourselves, of /course/ the word "wart" is subjective!

As someone who has suffered from aggressive, objectively real warts for 
at least 2 decades and probably will for the rest of my life, I do not 
find the bad metaphor very amusing. I wish people would stop using it.

-- 
Terry Jan Reedy

[toc] | [next] | [standalone]

#39141

On Monday, February 18, 2013 4:31:01 PM UTC-6, Terry Reedy wrote:
> This was a threat to abuse StackOverflow with off-topic posts if 'we' 
> did not pay him more attention.

If that is in-fact true then i am going to be as upset with Anatoly as you seem to be. Pointing out problems on appropriate list is fine, but making spam threats is totally wrong.

> [...]
> I will just briefly repeat two points:
> 1. we already have a searchable list of Python problems -- the tracker; 

But i don't believe a "Bug Tracker" and a "Language Wart" listing are the same. My belief is that a Tracker is for posting patches for "bugs"; meaning either the kind that throw errors:

[Warning: Naively hypothetical examples ahead!]
    
    py> 1+2 
    Exception: Seg Fault!
    
...or the kind that produce illogical results:
    
    py> 1+2
    12

I also believe a Bug Tracker is a serious place where stict etiquette and posting rules should be maintained:

 * only post if you have a working solution!
 * express the problem as succinctly as possible in text.
 * provide example code that exposes the error in a succinctly manner.
 * check any emotional baggage at the door.
 * only present "real bugs" and not illusions or preconceived notions of how a certain chunk of code, or syntax, or whatever, should behave.
 
As you can see, a bug tracker is in contrast to a PyWart listing that i propose. You can think of the PyWart listing as the first stop to solving all problems in the language. Only the true bugs, for which a patch has been written, should continue on to the Bug Tracker.  

In the PyWart list we want to get feedback not only for true bugs, but for un-intuitiable API semantics, inconsistencies, multiplicities, and asininities. 

We also want people to express themselves without applying any filters. If a certain feature "bugs" them, then rant about it a little bit, let people know how you feel about the wart, and THEN try to offer a solution if you have the capacity to do so. If not, hope that someone more talented will be influenced enough by your arguments to submit a patch on the tracker.

So, you could conclude that a PyWart listing is an informal manner by which /all/ levels of the community can participate in the evolution of the language. Anyone who has a mouth and a problem will now have an outlet. 

Sure, some people will consider ridiculous things to be warts that need removing, and they might rant and rave and nobody will pay them any attention, but hey, at least with my proposed PyWart listing they have a safe outlet for these emotions. And, "maybe" a fix will be the result, who knows.

All groups need an area where you can go and misbehave without being ostracized from the community. Some groups call this area "the bar" or what ever. With a PyWart listing we can maintain a high level of professionalism at py-ideas and at the bug tracker whilst simultaneously providing a synergy of evolutionary ideas to be expressed without censorship.

[toc] | [prev] | [next] | [standalone]

#39179

On 2/18/2013 6:55 PM, Rick Johnson wrote:
> Pointing out problems on appropriate list is fine,

I agree. Python-list is the most free among python.org lists.

> But i don't believe a "Bug Tracker" and a "Language Wart" listing are the same.

I agree. Anyone is free to make their own site or blog, call it what 
they want, and run it however they want.

 > My belief is that a Tracker is for posting patches for "bugs";
 > meaning either the kind that throw errors:
>
> [Warning: Naively hypothetical examples ahead!]
>      py> 1+2
>      Exception: Seg Fault!

We take these seriously if they involve core python. There is even, 
somewhere, a separate list of 'crashers' that we think should be fixed. 
The one left are the ones we do not know how to fix. Some reports are 
closed as "won't fix" when the crash is a result of freedom we will not 
take away. Two examples are ctypes and the possibility of hacking the 
bytecode and other fields in CPython code objects.

> ...or the kind that produce illogical results:
>      py> 1+2
>      12

Would that all bug issue were so clear. The hard one involve behavior 
that is not clearly defined in the docs.

The tracker is also for improvements, though I think idea are best 
discussed off the tracker first to discover if there is really any 
support and possibly to improve the idea. There are currently 1170 open 
'enhancement' requests on the tracker. Most are likely forgotten about 
and will never go anywhere without off-tracker discussion.

The tracker is definitely not the place for 'How do I use Python?' 
questions, or "How do I fix my code?" questions. They belong here.

> I also believe a Bug Tracker is a serious place where stict etiquette
 > and posting rules should be maintained:

Absolutely. Thank you for saying that.

>   * only post if you have a working solution!

That is ideal, but some people find problems they do not know how to 
fix. What is annoying however, is people how say "I could contribute 
code, but I am too busy, so I want one of you volunteer to do the work I 
will not do."

>   * express the problem as succinctly as possible in text.
>   * provide example code that exposes the error in a succinctly manner.

These two apply to posts here too. I am still amazed at 'My code doesn't 
work. How do I fix it?" posts.

>   * check any emotional baggage at the door.
>   * only present "real bugs" and not illusions or preconceived notions  > of how a certain chunk of code, or syntax, or whatever, should behave.

I would add that when people do post the last kind of issue, they should 
respect the developers enough to accept an issue being closed and not 
reopen it to drag on the discussion. We have too limited time for too 
many issues and are not quite keeping up as it is.

I have little or no argument with the other points you made.

-- 
Terry Jan Reedy

[toc] | [prev] | [standalone]

Back to top | Article view | comp.lang.python

csiph-web