2x37: Amidst the turmoil and darkness

Jeremy Garcia, Jono Bacon, and Stuart Langridge present Bad Voltage, in which there is more GDPR but not as much as there is on every website everywhere, there is discussion of rolling pins which is probably taking the retro-technology trend a bit too far, and:

  • [00:01:49] News: Python community generally feel that having no leader is gonna be OK, at least for now; Bad Voltage team say, this is our sceptical face ... Facebook to start reviewing people on their "trustworthiness"; also debates changing logo to a black pot surrounded by 2.2 billion kettles ... and London "fake news is not our friend" FB ads get "adapted" by Protest Stencil to say "it's our business model" ... the FBI decide they can solve crimes by just demanding Google tell them everyone who was nearby at the time; Google say: pull the other one, J. Edgar, it's got bells on ... and according to the Verge, $1000 phones have become normal, although it's not clear anyone knows what normal is, these days ...
  • [00:40:25] Slack. What is it good for? There are companies with a three-person team all in the same room who communicate through a Slack with six channels, and there are some trying to run a 10,000+ people community through it (or deliberately stopping doing so). What's it supposed to be for? And is it succeeding? Has it stopped being trendy now, or are these merely growing pains?

Come chat with us and the community in our Slack channel via https://badvoltage-slack.herokuapp.com/!

Download from https://badvoltage.org

Did you record this in BANGBANGBANG the middle of an industrial BANGBANGCRASH building site, by BANGBANG any chance…?

Facebook is getting really good at missing the point, isn’t it? “The solution to fake news is to have software decide who we should ignore when they report abuse” seems…well, typical. I wonder where the line is where people realize they can walk away and, if they walk away, are there better alternatives? Maybe more importantly, is there a (maybe even open source) solution that doesn’t require mass re-education or trust in an effective monopoly? I have some PageRank-like thoughts, but don’t know that I’ve thought it through and don’t know if it’d work in a network that allows privacy.

Slack, likewise, definitely needs to take a step back. They seem to really want to be Facebook, in many ways, designing the interface to “drive engagement,” walling off the gardens, and trying to convince users that they’re providing an innovative service and not something that’s been available forever. And it’s…not a great experience, beyond being really easy to set up.

I dunno, maybe there’s a world where we all migrate all communications to Scuttlebutt or something…

Also, what’s wrong with open source (and, heck, corporate) governance, such that everything is either totalitarianism or anarchy, with the occasional plutocracy thrown in? Is democracy, not even necessarily direct democracy, really that foreign or abhorrent an idea to these groups? It seems like a fairly straightforward move to me, but maybe the feudal/imperial nature of the Internet (see Facebook, Slack, Google…) means that we need an Age of Enlightenment before the equivalent of the Atlantic Revolutions can happen.

erm… it sounds fine to me? Did I miss some place where there was a loud noise?

Most of the Slack discussion has thumps and clangs in background, I think especially while Jono is talking (e.g. 41:50 onwards, but I’m sure there are worse bits after). Maybe an artifact or echo? I’ve not noticed it in previous shows.

I don’t hear any issues myself.

I definitely can. It’s like heavy rain on metal, or someone typing on a typewriter? Or maybe just a weird mic. Definitely there, right as you say “I think Slack is a medium that is marketed…”, after you say “build a community in it”, after Stuart says “let’s set up a slack to do it” and again for a bit. I tried with different (Sennheiser) headphones just to make sure it wasn’t an artifact of crappy Apple ones, but it’s definitely there at that point. Reappears here and there (there are thuds when Jeremy says he’s the most fatigued), then goes away when the Slack segment ends. Ah well, maybe the moon is making my ears hypersensitive or something :slight_smile:

No issues with sound quality for me, another interesting show folks!

@toyg there was an issue in the mind control signal. Evidently, your equipment was sensitive enough to pick up small portions of the, what is suppose to be, inaudible messaging. This shows subliminal message was “Don’t disrespect small hands. We need smaller phones”

Sorry for the inconvenience.

1 Like

I don’t hear loud noises, but I while Jono is talking just after the timestamp toyg mentions, there are noises which sound like a neighbour is playing the drums, and the ducking between sentences is quite aggressive, meaning any background noise while someone is talking becomes very noticable.

1 Like

@newky the sound issue was around eleven minutes.

18:35 So I searched briteabart, chroebook and 14.04 - and realized that Stuart had asked the Bennet question from 2015 later in Open cooks (not the Louise one)

When comparing Guido and LInus, an important distinction to make is that the Linux Foundation pays Linus to work on LInux full time while Guido has a full time job at Dropbox (and previously at Google) and has never wanted to work on the Python language full time (I believe the Python Software Foundation could have created such a position but I don’t have a source to cite for that claim). So the role he was playing was something he was doing because he enjoyed it and was not a full time commitment. The Python community has a fair amount of governance overall. LIke Guido said in his resignation email, the parts that he was responsible for were approving new core developers and signing on language enhancement proposals. There are enough competent core developers that I am not worried about them coming up with a way for them to approve new core developers from the pool of contributors.

The tricky part is guiding the future direction of Python overall, as you alluded to with the comments about people coming of the shadows to push all the features that Guido nixed over the years. Most developers are not too worried at the moment because there is a backlog of approved but not implemented language features so there is no rush to approve more language features, but a new way of guiding language development does need to be decided upon, and a dictator with a strong vision often produces a more consistent product than a committee representing different interests. Python is really popular now but it does need to maintain a vision and a focus to maintain its status as it still has many challenges (application deployment is messy, performance vs. compiled languages, async api’s, competition with javascript for role of first teaching language, etc.).

Agreed, but the word “often” is clearly doing all the work there. That’s not a law of the universe; I’m not sure the correlation between “strong single-person vision” and “gets it right” is so strong that one should a priori reject all other methods for running a community! For a start, there are quite a few dictators with strong visions who end up getting the consistent product they wanted but it’s consistently rubbish. Also, JavaScript, which you explicitly mention, is designed by committee, and via a tortuous standards process at that. I’m sure people will leap here to say “haha it shows, JS is terrible, why would anyone use it, lol!!!1!!!”, but those people then need to explain why it is that Python is, as you note, battling with JS for the role of first teaching language.

Anyway, I personally think you’re probably right in general about dictators, but I don’t think there’s an obvious dictatorial candidate for Python (I think Barry would be good, but that’s mostly 'cos I know him and know he’s smart, which isn’t the right reason :)) and I feel like forcing someone to become one is a good way for it to go wrong, which is why I think that a loose committee process among the more senior core devs will probably be best. But I don’t know the dynamics of the core dev community particularly well, so I’m quite happy to hear alternatives there!

I’m not sure the correlation between “strong single-person vision” and “gets it right” is so strong that one should a priori reject all other methods for running a community!

Yes, I agree with you. I think I wrote my last post too late at night. I was partly responding to jcolag’s question about why democracy was rare in open source and partly responding to the attitude of skepticism in the show regarding Python’s current status (without a dictator). The sense I meant to convey was that I’d prefer some more communal form of governance, but that I have to concede that a dictatorial approach is often effective (with Guido’s leadership of Python and Linus’ of Linux being two good examples).

Regarding Javascript, I think it has been getting a lot better in recent years. I don’t do web development so my main experience with it is in writing web browser extensions, mainly for personal use. In the past, it often felt like a chore switching to Javascript for this, but I recently wrote a WebExtension for Firefox and used a lot of ES6 and ECMAScript 2017/2018 features and found I could write much closer to the style I would use in Python (e.g. using async/await instead of convoluted callback/promise chains). Still I think Javascript’s main advantage over Python for a teaching language is just that it is readily available in all web browsers, whereas deployment remains a sticking point for Python. Maybe that will change if WebAssembly makes it easy to run Python in the browser.

I hope Python does adopt a multi-person council of some kind because it seems more sustainable than relying on a single person. Perhaps it would help to have a new PEP that expands on PEP20 (the Zen of Python) to state the guiding principles that should inform decisions about changes to the language. I find PEP20’s aphorisms contradictory and a little opaque (e.g. PEP572, the one that made Guido resign, seems to violate several lines of PEP20 to me). I don’t know the core devs well but Brett Cannon seemed like a likely BDFL candidate from what I could tell.

Does it, though? I know it’s a common assumption and a dictatorship can certainly produce a product faster than alternatives, but there are plenty of single-programmer shops out there producing incoherent messes and more than a couple of large-team products with goofy quirks because the boss insisted on it. And that doesn’t even get into the problems of getting the right product out the door…

That’s obviously not to pick a fight, since you point out you’re not supporting dictatorial projects. But Guido stepping down seems like a perfect excuse for at least the open source community to start having the discussion.

Well, like Stuart pointed out, I am leaning on “often” there. It is certainly not necessarily the case that a dictator produces a more consistent product than a committee. Perhaps it is more efficient for a single creator to produce something new with a consistent vision than a committee since that person can internalize the vision and cut out a lot of communication required for a committee. However, it may be that a committee is better suited to maintain an established product since the committee can represent different segments of a community and have different areas of expertise. So Guido might have been well suited to guide early Python, but a committee might be better able to oversee its further evolution.

Also, it is probably worth mentioning that Guido was not that active of a dictator. He had certain areas of interest that he contributed to (like type annotations) but, for the most part, his role was to say yes or no to PEPs. The actual proposals came from community members, so the community has already been involved in guiding the direction of Python. Also, he would defer to judgment on certain PEPs to other developers with more expertise. If he is succeeded by a committee, there will need to be more explicit communication about decisions which could be good. If that is the case, I hope the committee takes the time to generate a design document to guide future decisions.

1 Like

@wsha excellent comment with the information about Guido. I did not know that.

@jcolag a democracy in software development seems a bit unnecessary to me. Democracy is good as a government model because every citizen is equal there. For software development do you include the users in the decision making process? That’s kind of the case in commercial software already; if it’s not got a big market share already then the customers’ needs dictate the product.
In the case of open source software that means the developer of nano has to listen to some random nutter on the internet for the most part. Maybe then there will be no more nano development.
Perhaps then the idea is to have only developers be given a vote? But is every developer equal? Does that one person with that single pull request last year get to vote? Does the full time developer get 40 more votes than the one who does an hour of QA in the weekend? Then the art contributors to projects might have to use a webcam to log all their hours on working on art for the project. The net result would be the model of KDE as talked about on the show, with the respect towards people, only a lot more annoying.
With an indirect democracy you might as well just have a committee.

@Klover, my concern is how much similar lines of reasoning occur in political systems. Why should legislators be required to speak with their “nutter” constitutents? Why shouldn’t rich people (using money as an indirect proxy of productivity) have more say over what laws get passed than some poor person? Well, because pick from (A) it’s the right thing to do when you’re governing and (B) most of the kingdoms and empires died as soon as there was an alternative and those that didn’t cost an arm and a leg to maintain.

Mostly, it just fascinates me that the business world and the Internet at large basically acts like feudalism is the latest and greatest thing, when the outside world seems to have accomplished a heck of a lot with Enlightenment thinking. And because of that, I have to wonder what the results of an “Age of Internet Enlightenment” (or corporate) would be.