Many of us have also been thinking about what we would do when this day arrived for years (Guido has hinted at retiring previously), so there's a lot of pent-up thought to get out there while we all start to think about what comes next to a project some of us have dedicated decades to. The conversation has been thoughtful and not a single person has said "I should be the next BDFL". All names put forward by anyone for any position has been by another.
IOW the insinuation that any of us who are trying to grapple with this are trying to vie for power while ignoring why this occurred does not seem like a fair assessment to me. This kind of unnecessary negativity and accusation is why Guido is retiring early.
I don't think it's at all unseemly that people involved in the Python project respond to GvR's LOA announcement by working out continuity. As someone who has to interact with a lot of Python code professionally, that's exactly the response I'd hope for.
The market gave us the absolute mess that is HTML/CSS/Javascript today, so I'm sincerely hoping the Python community will keep agreeing on some greater design principles instead of leaving everything to market forces and pragmatism.
When the market did take over, the problem was that those poor foundations weren't thrown out completely. You can only do so much by strapping turbines on a camel (no offense to camels). At least we can actually write (mediocre) applications with HTML/CSS/JS now.
The other part of the mess is caused by the ever-growing amount of trend-hopping junior developers that want to try out new things - and their superiors letting them do it. If the foundation wasn't so bad, there would be less incentive to try and re-invent anything. Other platforms are fully market-driven, they didn't produce such a mess, because the market rewards stability (hence the low initial adoption rates of Python 3).
Yes, but a committee (or a community if that term has too many bad connotations) can at least argue about the design principles and on occasion decide to change them using a well-defined process (e.g. rough consensus or voting).
> When the market did take over, the problem was that those poor foundations weren't thrown out completely.
The original plan of the W3C TAG was to pull the foundations and make a fresh start with XHTML2. The browser vendors objected to that and chose to instead evolve the original HTML into what we have today.
> The other part of the mess is caused by the ever-growing amount of trend-hopping junior developers that want to try out new things...
But that's the point. Those trend-hopping junior developers and their bosses are the market: If you view programming languages as products, those are the early-adopters, one of the generally most sought-after part of the customer base. If you want to be 100% market-driven, you have to listen to them.
> ...because the market rewards stability
The market apparently didn't bother much that HTML is now a "living standard", browser release cycles are measured in weeks and generally a software is considered dead if it doesn't receive any more updates.
The underlying assumption here must be that this somehow leads to better results overall, but where's the evidence for that? "Design by committee" has a negative connotation for a reason.
> The original plan of the W3C TAG was to pull the foundations and make a fresh start with XHTML2. The browser vendors objected to that and chose to instead evolve the original HTML into what we have today.
That's because "the market" doesn't want things to break. Like I said, it should've been thrown out, but for obvious reasons it wasn't. The point is, you can't blame "the market" for having created the mess in the first place.
> But that's the point. Those trend-hopping junior developers and their bosses are the market...
They are a force within that market and it just so happens that so many new people come into the industry because "web stuff" is needed now, but it won't be growing like that forever. These young programmers will grow old and tired (and so will their bosses) and at that point conservatism will settle in, like it has in many other areas as well. It's a market fluctuation.
> The market apparently didn't bother much that HTML is now a "living standard", browser release cycles are measured in weeks and generally a software is considered dead if it doesn't receive any more updates.
The pace at which new browser features are adopted is actually rather slow, but more importantly, old stuff usually doesn't break. For example, jQuery might be outdated from a developer perspective, but it still powers a lot of stuff.
TCP/IP development happened quietly for the most part in the 1960s-70s. There wasn't a lot of pressure, and they had a decent amount of time to get the protocols right. There wasn't an economic demand for Arpanet.
And it wasn't until about 1993/1994 that the web exploded in use and popularity. That was only 4 years after TBL created HTML. That's when you saw the explosion of JS/Java Applets/CSS/Browser Plugins/etc.
In some sense, the same thing is happening in the python world. While python has been around for a while, there were maybe 800 python devs at Pycon 2010. In Pycon 2018 there were 3000ish(?).
I do agree with you that the Python 3 update wasn't done well, I think it is because they didn't predict the language's explosion during the 2010's.
Also, an apology for the way this turned out would be seemly.
So again: why, as a professional who interacts with the Python ecosystem, am I interested in anything more than what is already happening on the thread?
Even if infuriating jerks are 0.1% of the population, when your language has hundreds of thousands of users, you are going to deal with a lot of jerks. Frankly, I'm surprised he lasted as long as he did without going mad or something. I would not wish for any of my friends to be subjected to that kind of attention.
We were very lucky to have him leading the project for so long. Python will survive without a BDFL and I hope he enjoys the vacation.
It's kind of cold to me, but I know this field is not filled with the emotional types...
Completely tangential, but the density of negatives in that sentence was nothing short of majestic.
I think you'd say the same, if you agreed that culture was a solvable contributor to GvR's exit (even if you, like me, knew he was leaving for market reasons eventually). By analogy, if 'dang said "I'm leaving yall, this sucks, elect a replacement" and we were like "cool who's it gonna be". That would be an error, and I think you would be at the top of the comment page saying "Also, let's all make some changes so that the next 'dang doesn't have a miserable life."
A huge thank you is more than deserved, but not an apology. Nobody did anything wrong, GvR's position just stopped being fun because he was too successful.
The market didn't give the world any language nearly as readable and approachable for non-programmers.
I do hope he knows how his contribution changed the world for better for so many people. I have played around with many languages in my career but Python is the most productive and fun of them all. Thank you GvR, and may you have fun whatever you do!
[0] https://legacy.python.org/search/hypermail/python-1994q2/104...
And I'm not sure who would be the right person or people to apologise. A number expressed regret at what happened, but I don't really think the core devs are at fault here.
Does being more empathetic when dealing with contributors and their suggestions extract a greater price on one's emotional state, compared to the more brash approach Linus has?
Linus shoots first, that keeps the bullying down -- of course everyone else thinks he is the bully, however he's just protecting himself. I've spent many years victim of bullying, and I know that the only way to beat a bully is to out-bully him.
A lot of people actively enjoy and seek conflict.
Think apple did not fall too far from the tree :)
There's certainly a friendly nod to the CoC and an invitation for people not willing to adhere to it to leave —and I think that alone would help chill things out— but it's a sustained level of often-thankless hard work. Especially if you have a day job where they're pushing for more and more performance.
This missive is about helping the project move on, not himself. Bollocking the people who have made things hard isn't a productive next step. It's not what's required now; the core devs need to work out how decisions are made.
This is that prompt.
It's kind of funny to see someone worrying about sparing Linus Torvalds' feelings.
This is perfectly consistent with 99% of the management behavior I've witnessed for years in the enterprise. We understand that we live in a world of cause and effect, but almost everyone is so busy poking at the effect that there's no energy left to discover and treat the cause.
It's a shame that so many of us workers continue to have so much passion for the work at hand and so little for everything else. It seems like it almost always comes down to this.
Respect, Guido.
- What was fun 3 years ago isn't anymore
- That feature that was useful to your customers is no longer serving purposes
- It was about new feature, and now it's about maintaining compatibility
- ...
There is no real cause to "treat", you just need to adapt to the new environment you are facing.
> There is no real cause to "treat", you just need to adapt to the new environment you are facing.
Maybe, maybe not. The purpose of a deeper analysis should be to find out whether or not it might be like this - but you can't just assume it's like this from the start.
Exactly. I can't imagine keeping the same job/role for 20 years. Guido even says it : he is tired. He has had the energy, drive and will to steer the python ship for 20 years. He now feels that it is time for him to take a well deserved holiday/vactaion/time off.
As other have said, this has probably already been discussed within the core devs already.
But people need to find a "reason" to explain what is just a very natural process. And will spend countless hours looking for it instead for just moving to the next adventure.
True, although sometimes there isn't much to be done. He mentions he is tired of fighting and, with such an enormous project, how do you solve that?
This does not bode well for the future of Python, though. The ecosystem is fragmented as it is, I can only imagine how things will look like without the well known leader.
The lack of introspection that causes one to, in the heat of the moment, behave like a tedious buffoon on a public mailing list is probably associated with the lack of introspection that causes one to not reflect on having behaved like a tedious buffoon on a public mailing list.
On the other hand, I had never looked at the Python mailing lists until today and my main impression was that a real dictator on a mailing list without a Code of Conduct could have stopped some of that endless chatter in its tracks if he'd really wanted to. Absolutely no one springs to mind...
Inflammatory generalizations with no evidence like this count as flamebait, so please eschew them.
You're making an astoundingly general claim and chimpanzees are not going to back it up. I understand that you might disagree, but there are community standards of discourse here that are not being met and we need you to work on that if you're going to comment.
In any case I've acted in good faith to contribute positively to the discourse. I haven't antagonized anyone and I've provided relevant new information that might shine new light on the discussion at hand. If anyone isn't meeting the standards of civil discourse, it's you and whoever flagged my comment. You'd think people on HN would be curious about new ideas. Instead all potential positive discourse is shut down due to a single individual's subjective judgment of what is inflammatory, despite it being scientific consensus. It's clear this isn't a space for civil discourse amongst reasonable people.
When we moderate HN like this, we don't mean to imply that you weren't posting in good faith. We're just being vigilant about what is known to cause flamewars. Flamewars are the #1 problem here because they consume everything they touch and can easily lead to the death-by-fire of the forum. We're basically just being Smokey Bear, or Smoky Bear's ranger friend.
It's interesting to observe that the above holds true even if you're right in everything you said. The old question "would you rather be right or alive?" applies here.
It's not hard to figure out the intended use of the site. If you don't want to use it that way, please don't post here and please don't make new accounts to get around the restrictions.
I was moderator for more than 10 years on linux.org.ua and nobody told me that I was bad moderator. I also regular user on other forums/sites, so I have view on the problem from both sides.
HN is badly moderated. However, most problems with moderation/flames/etc., can be fixed using technology or by writing better rules. If you are interesting in fixing of HN problems, contact me: [email protected] .
IMHO, HN can be improved to automatically filter out discussions. Imagine, we have a magic system, which automagically labeled all comments as a) on topic or not on topic b) improvement, critics, opinion, correction, controversy, alternative view, support, trolling, joke, fun, suggestion, discussion, flame, etc. c) history, politics, physics, mathematics, engineering, computer science, programming, etc. Then we can just place some labels on top of each submission with number of comments for that label. By default, only on-topic comments with good labels should be enabled. But user should be free to enable other topics as well.
Now, we need to imagine how to implement such automagic labeling system with minimum of up-keeping cost. IMHO, first we should allow user to label his comments himself. If comment is labeled properly, then user will not punished. If comment labeled incorrectly, then other users can vote to change label with cumulative score (if user has higher rating, his vote worth more), and user will be automatically punished if label changed from good to bad.
On-topic/off-topic should be implemented as checkbox on submit form. Other labels can be implemented as hashtags or collapsible boxes. Users should be able to add or remove labels to other posts, if necessary, when they have high enough rating for such action and label. Users should be able to chose which labels they want or don't want to see with reasonable defaults.
IMHO, it's much better to be constructive ans ask user to label his post as "#politics #flame" instead of forbidding him to post.
Example:
A post title. yyyy.mm.dd hh:mm on-topic(12) off-topic(21)
vlisivka x hours ago | off-topic(up dn) #hn-site #suggestion (add label) I have a suggestion about how to improve HN.
Just want to say: It has nothing to do with sex - women do this also. Please refrain from characterising whats going on with Python right now as being a consequence of some sort of masculinity out of control, because I fear that is a falsehood and does you - and anyone who tries to agree with you - a major disservice.
There is no wrongdoing mathematically, but it does show that selective vision is a force for damage if wielded improperly or with agenda.
[citation needed]
Maybe an environment that is more respectful and professional?
Guido van Rossum has given his life for this language and besides the obligatory 'thanks for all the fish' there isn't even a single person who stops the clock to evaluate what went wrong that they pushed out the person that started this all.
Instead it's 'kthxbye' and they're already dividing up the cake to see who gets to rule.
Not the nicest moment in the history of FOSS, I wonder what kind of a mess will ensue when Linus steps down.