Meta: Lifetime of an idea. 1, 2 & 3

A lot of core developers have publicly stated that they ignore the ideas category because the signal to noise ratio is so bad. Given that any core Python change needs the support of at least one core dev, the net result is that the ideas category is getting progressively less relevant over time.

That is the problem here.

9 Likes

Makes sense. Maybe there can be a separate category (simply named Proposals perhaps) for >50%-baked ideas where new topics are required to be written in a template format similar to that of a PEP. If a poster can’t fill out most if not all sections on the template, the topic should belong in Ideas until the poster feels ready to rewrite the post as a new topic in the template format in Proposals.

4 Likes

Or maybe that can be called “Ideas”?

1 Like

How about PEP proposals? Or drafts?

I feel that there is a clear distinction between posts asking for feedbacks on ideas to make changes to Python, and posts asking for help in using Python as it is, and lumping <=50%-baked ideas into the Help category makes it hard for some of us who are actually interested in reading about potential Python changes to find such posts (I personally have different notification settings for them).

2 Likes

But the poor quality ideas make terrible Ideas posts. They are FAR more useful as Help posts, asking for explanation of why Python is the way it is.

You are, of course, always welcome to post ideas on a completely unrelated forum and discuss them with whoever you want to discuss them with. Why do you want to discuss them here, specifically? Is it not because you hope for them to be seen and adopted by the core devs? You need to get at least one core dev to approve your idea before it can even move forward. But you won’t get that when the idea discussion is more of nailing jelly to a tree than discussing a real idea.

Jelly ideas are better in Help.

All the “what if python did XXX” posts are not help requests, just quick (and dirty?) ideas. People writing these kinds of “proposals” won’t ever think about posting those in the Help category, because they are not help requests (even if they maybe should).

Adding a “<50%” category feels like a graveyard where bad (or bad written) ideas go die. In the Packaging category, there are multiple tags, maybe they could be used here also ? I’m not good at finding names, but something like an “incomplete” tag could help filtering out lower quality ideas. I don’t know if it’s possible to filter ou some tags, if it’s not maybe add a “to be discussed” tag for >50% ideas, so that peaople can concentrate on those ? Only some privileged users (core devs?) could add this tag.

My 2 cents.

1 Like

An “Incomplete” tag would also be a place for badly written ideas to die.

The solution is not to give people another way to post and get ignored. The solution is for people to rethink the pitch: if it’s not a properly written idea, it IS a request for information. Don’t post it as “hey, here’s an idea, I expect you to take it seriously even though I haven’t put in half an hour researching it”. Post it as “Hey, I was exploring this, and I’m curious about this aspect of Python”.

Otherwise, it’s just going to be ignored by everyone who matters.

That’s why a separate category makes sense. I (and quite a few others I think) am interested in any discussions about potential changes to Python even if they are <50%-baked, which is why I have a higher notification setting for Ideas. Help is noisier by a long shot so I have a lower notification setting for it and I would only visit it when permitted by my spare time, often missing many Help posts between my visits.

By having a separate Proposals category, where only quality ideas are allowed, people like many core devs who are only interested in more fleshed-out ideas can have a place to subscribe to, while still leaving people like me who are interested in any ideas a channel to more easily tune in. That’s what categories are for after all.

Well, try to convince one of the Discourse admins that it’s worth it. Otherwise, the Help channel will work fine.

Of course, but they would still be on the Ideas category, still visible for anyone willing to educate newcomers but less intrusive for people with more limited time. I tend to think that it’s a good middle ground between “do nothing” and “compeltely remove from the Ideas category”.

I agree on the principle, but how do you plan to force people to think about it before writing? We have to find ways to manage what will inevitably happen anyway.

Any unthoughtout posts get moved to Help, where they belong. You can help with this by flagging them as “Off Topic”.

This interpretation of the problem prompted me to start thinking of potential category names that would distinguish a new category from both “Python Help” and “Ideas”. They’re all in the vein of open-ended questions starting with “Why”:

  • “Why?”
  • “Why not …?”
  • “Why is it so?”

The idea would be to distinguish it from both the “How?” questions that are the main focus of “Python Help” and the more developed proposals for change that are better suited to “Ideas”.

Here’s the way Ideas is currently summarised: “Would you like to change something in Python? This might be your feedback forum.”

A new “Why?” category might be summarised as: “Wondering why something in the Python ecosystem is the way it is, or wondering about the absence of features you’ve seen in other languages? This is your forum to find out more.”

Proposals for change necessarily start with questioning the “Why” of the status quo, so I genuinely like the idea(!) of framing things that way. (It would also make sense to have a pinned & locked thread referencing Design and History FAQ — Python 3.12.3 documentation )

6 Likes

By having a separate Proposals category, where only quality ideas
are allowed, people like many core devs who are only interested in
more fleshed-out ideas can have a place to subscribe to, while
still leaving people like me who are interested in any ideas a
channel to more easily tune in. That’s what categories are for
after all.

Dunning-Kruger type cognitive bias plays a huge role in this,
unfortunately. People with limited experience in these matters are
often quite certain that their ill-considered notions are “quality
ideas” because they lack sufficient context/awareness to see the
difference for themselves. I’m starting to agree that reclassifying
and moving them to a different category whenever they appear is the
best available option, even though that doesn’t do much to prevent
the initial posts from landing in the inboxes of people who rely on
E-mail for this forum.

2 Likes

I’m starting to agree that reclassifying and moving them to a
different category whenever they appear is the best available
option, even though that doesn’t do much to prevent the initial
posts from landing in the inboxes of people who rely on E-mail for
this forum.

I suppose something else that might reduce the nascent post volume
in Ideas and also soften the blow of inevitable reclassification to
Help could be to expand the pinned post in Ideas with guidance along
the lines of:

“IMPORTANT: The Ideas category is for serious proposals about
improving the Python language, CPython interpreter or its standard
library. Thoroughly research prior proposals and consider the
feedback they’ve received rather than repeating the same or similar
ones. If you lack the time and skill to (perhaps with some
assistance) follow your idea up with a formal PEP and supply the
bulk of any eventual implementation yourself, please discuss it
elsewhere instead. Your post will get flagged off-topic and
moderators will move your post to the Help category when this does
not appear to be the case, so it may be best to just start there.”

3 Likes

This is a good way of separating the topics, but it doesn’t solve the problem which is “enforcing” this. I mean that new users will probably look at their fancy new operator or syntax as an idea, and not as a “why doesn’t this exist?” question.

I think the best way is actually setting the bar higher for one of the forums. A mandatory template for ideas, which will actually be more specific and tough than PEPs, since in PEPs you have a sponsor for feedback.

Of course, people can still post low-quality ideas with a template, but I think demanding a lot of answers before posting will usually make people understand they shouldn’t post there if they didn’t think about seriously, and should instead ask why.

An example question in the template could be: 'Has this idea or similar ideas been discussed before in a PEP, the forum, or the mailing list? If so, please provide an answer to all substantial reasons this idea was rejected for", or questions and notes about specific types of changes, like syntax changes.

A lot of these ideas are nice in theory but I’m worried they put too much burden on the moderators to make things work properly.

A much less egalitarian but very practical solution is to split Ideas into open and closed sections…a (proto-)Ideas category where everyone can propose and discuss ideas at any stage of baking, and a pre-PEP category where starting a thread is restricted to established contributors[1]. People who are getting burnt out on the SNR of Ideas can fully mute that category without missing out on serious proposals.

I know that this will be controversial because it imposes an explicit hierarchy on participation, but really I think that hierarchy is already present in this forum and this is just an honest acknowledgement of the different ways people are participating in discussions here.


  1. i.e. core devs, people who have at least written a PEP in the past, etc. I’m certainly not in this category, FWIW ↩︎

4 Likes

I don’t think there is anything wrong with the actual categories. The problem lies with users not willing to listen (for whatever reason) and being stubborn about moving their ideas to Help, or just dropping them when they realize they don’t really have a clear idea yet.

Some idea threads are being created from an XY Problem situation.

I don’t think that adding the problem of categorizing threads is going to ease the status quo.

2 Likes

I’m not against a hierarchy, but the separation you made probably won’t help much, since it excludes 99% of the users here. You have more than 700 posts and 800 likes, and you say you wouldn’t be in this category, most of the ideas aren’t from PEP authors. The point is to separate the ideas “1% baked” and “50% baked”, not 98%.

2 Likes

The existing hierarchy is based purely on quality of posting, though. Anyone can become respected in the Ideas category just as a result of their behaviour. If we had a “closed” invite-only subcategory, that would be exclusive to people who managed to get an invite, which would be a much harder barrier to achieve.

2 Likes