[pydotorg-www] web frameworks page feedback

Paul Boddie paul at boddie.org.uk
Sun Dec 21 14:28:58 CET 2014


On Sunday 21. December 2014 00.14.01 Martijn Faassen wrote:
> Hi there,
> 
> [I'm not sure where discussion about the wiki content should go,
> perhaps not here, in which I apologize. please skip to read the
> bottom]

I think this list is as good as any for discussion of the content.

> I noticed a recent change to the
> https://wiki.python.org/moin/WebFrameworks that moved Grok (I
> framework I helped found) down from the popular frameworks. I agree
> with that move; it's not been popular for a while now and was never as
> popular as, say, Pyramid or Flask are now.
> 
> Here are some concerns I have with the current page:
> 
> * Frameworks that I consider are big and get a lot of attention such
> as Flask and Pyramid do not come under any "popular" heading. But they
> are very popular, and right now they're ranked with obscure things
> like BlueBream or Aquarium. I don't think this reflects the state of
> the Python community. If you're going to do popularity stuff at all,
> why not list some of the popular non-full stack frameworks?

The problem with popularity is that it is fluid. Originally, the notion was 
introduced into the page because people were unhappy that the "cool" 
frameworks (my choice of terminology, here) weren't being emphasised enough 
and that people were being shown too many options. One has to look even 
further back to see that the page and its predecessor were only ever intended 
to catalogue the available solutions and not offer advice.

At some stage, someone even wanted to remove all but the most popular 
frameworks, citing the meme going round the Python community at the time that 
people were overwhelmed with choice and that the way to deal with this was to 
remove evidence of choice. Perhaps they overlooked the fact that people who 
perceive there to be insufficient choice may then go and create their own Web 
framework in response. ;-)

> * Why are some frameworks in a table format (which gives them more
> attention, I think) and some not? I can see doing so for popular
> frameworks, but why do it for some other full stack frameworks and
> some not?

It looks like someone started converting the lists to table format and didn't 
finish:

https://wiki.python.org/moin/WebFrameworks?action=recall&rev=423

It would have been a lot of dull work to do that, and I don't think it's worth 
the effort, really.

> * A framework like wheezy.web is considered "full stack" and is
> highlighted in a table as such. I appreciate wheezy.web, but I really
> think this is unfair to frameworks like Pyramid and Flask, which offer
> more or less the same facilities. You can argue they offer some of
> them through extensions, but so does wheezy.web. If you're going to
> say wheezy.web is full stack, then so are Pyramid and Flask.
> 
> For a traditional Python full-stack framework I'd look for something
> like tight database integration and a form generation system, like
> Django has, or Grok has. wheezy.web is far more "roll your own". (For
> a more modern full-stack framework a lot of it might be done
> client-side instead)

I think you've just described some of the problems with categorising things 
like Web frameworks. Unless a thorough job is done in presenting the 
characteristics of each solution, however, any categorisation will seem 
arbitrary. And again, people tend to get upset if you put any detail on a page 
like this because "it confuses people". (Personally, I think that the people 
who get upset should stick to their 140 character "tweets".)

> Finally, I'm the developer of a non-big non-full-stack framework
> called Morepath. I'd like to see it added to the page somewhere too.
> (http://morepath.readthedocs.org).
> 
> If someone gives me editor access I can see about making some edits
> myself, but I'd be happy if someone else took care of these concerns
> too.
> 
> My account name is MartijnFaassen

You sound like someone with sufficient enthusiasm for, and understanding of, 
the topic to be remedying some of these issues. :-)

Marc-André mentioned the editorial notes on the page. I added these a long 
time ago, and they were really only intended to rein in some of the more 
extreme editing exercises going on at various points to "improve" the page. 
Some of the notes seem rather specific because they actually reference fairly 
destructive behaviour on the part of one contributor:

https://wiki.python.org/moin/WebFrameworks?action=diff&rev1=91&rev2=92

(Of course, one gets described as a censor if one tries to prevent people 
wiping away other people's work because "they know best". There's another page 
I can think of where people have whined about choice and "confusion" before 
now, and I find that it's often the case that the people with the problem just 
happen to have a pet solution that they think everybody else should be using. 
That's quite a coincidence!)

Nevertheless, a new perspective is welcome for this page, and I encourage you 
to edit away. :-)

Paul


More information about the pydotorg-www mailing list