Transcript: DjangoCon US 2022 Recap

I'm Will Vincent, joined by Carlton Gibson.

Hello, Carlton.

Hello, Will.

Austin said hello, Carlton.

Hello, Will.

How are you?

We don't have a guest for a change, but we're going to do a recap of DjangoCon US, which we both attended uh last week.

This will come out a week or two later.

But lots to talk about.

The message boards are blowing up.

This this week's Django newsletter is super, super long.

And what are your quick takeaways from the event and seeing Boston for the first time?

Well, so there's two things.

Um, first, the event is in this applies equally to Django New York.

It's just been so good in both the conferences to finally get back together after so long.

Um, I think people were really struggling um with the pandemic.

I think the community was sort of feeling a bit like, ah, is it still what gonna work?

Is it still gonna work?

And it's just as soon as you're in the same room, it's like, yes, it's you know, here we are.

And you know, and then you know, the vibrant in-person thing, that's just amazing.

And then the second thought I had to say, uh just about Jan Con US particularly is I just have to praise their COVID policy.

Um, because they had this very strict COVID policy about vaccines and about pre-testing before you arrived and about masks wearing masks inside the venue.

Outside, no, it was okay, no mask, but inside the masks.

And whilst it's a bit like, oh, this is a bit of a pain, it's just good, it turns out that it's 100% successful.

I don't know if nobody got COVID there, but I've seen so many um tweets and reports that you know, negative tests when you get home.

I'm sure people are tired, they might have a sore throat or whatever, but they haven't gone and caught COVID at the conference.

Whereas I've seen other in-person conferences where it's perhaps been not so strict, and it's like, oh, I got COVID, oh I got COVID, oh, I got COVID.

So honestly, I just have to salute that 100%.

Um, it was a tough choice to make, you know, there was a lot, some controversy about it.

I think they've absolutely made the right decision and they've implemented it really well, and it's got a proven positive result.

So well done, DjangoCon, for that.

So we can you should sprinkle that in.

With the the videos are up, we're promoting them on the newsletter, but you know, there was a bit actually, I think twice the size DjangoCon Europe this year, which is

It's like uh the force it's been off and then all of a sudden it's open and just yeah.

It typically DjangoCon US is like five, six hundred and Europe, I don't know, a little less.

I don't know how that translates into future years.

Um who knows?

Um, but yeah, I mean, what can I say?

It was it was lovely.

I mean, the venue in Porto was very big and they didn't have such a strict COVID policy.

But I mean, I didn't get COVID though, I don't know what other people do, but it you know, it was fine, there was lots of outdoor space, there was um big meeting rooms, you know.

If you've got ventilation, but the the main sort of worry area, I guess, is in the um the seating bit where you've got um you know the talks and you're all crowded in a space and ah, what should we do, especially after two years of not seeing anybody and not doing these events, and then it's an in-person event and there's a bit of ah I don't know what the correct protocol is.

And no, as you as you say, like um some past conferences everyone's been testing positive after, or you know, either on Twitter or anecdotally, and um, so anyway.

I I have no problem at all with with that.

Um, so that was my view um on it.

How was Boston?

Boston was nice because on the way I stopped off to see you, and uh Will's always telling me that uh he lives in Brookline and should he move to the country?

I'm always like, yeah, move to the country.

But I went to see Will and his family in Brookline, and that was very nice.

Um so now I'm like, you know, you should stay in Brookline, it's lovely.

So no, you know, but I have to say, fall, fall in particular is a pretty I mean you got to see foliage, right?

You hadn't seen foliage in a while.

What's that in New Money?

60, 70 degrees.

So um yeah.

I like I like I like to hear that.

Yeah, good, good, good.

You'll have to come and see us in Spain.

Well, uh next Django in Europe is May in Edinburgh, which I would love to attend.

I hope I can do that.

Um, but certainly I'll

Certainly you're gonna come.

So I was gonna cut over you then with um we both gave talks.

Let's talk about your talks, uh, your talk, and then we'll talk about my talk, then we cut back to you know round up about the the conference.

Um yours, the your European version is out, so we'll put it in the notes.

Um yeah, my talk, 45 minutes, um, the Django Jigsaw puzzle, aligning all the pieces, trying to provide a comprehensive look about what is Django, what is the entirety of the framework, and do some high-level code examples.

So I think the talk went okay.

I wasn't as prepared as I have been for past ones, partly because there's no Django Boston meetup anymore where I would normally run through it.

So you have

to stop.

Well, someone needs to.

Well, and there were there were people, um uh one of the sponsors, Path AI, of the conference, is based in Boston and and would provide hosting space before.

So yeah, it's doable.

Um but yeah, I was I was pleased that the talk happened.

Um I think it'll be I was actually talking to some people from Cactus Group yesterday about mentorship because they they have that's an agency with about 24 developers in North Carolina, which is likely where DjangoCon US will be next year.

And they have some mentorship programs, and it's hard for there's this imposter syndrome when you're learning web development that you just have no idea what you know and don't know.

And so the goal of the talk was to say, here's as much as I can show you a roadmap of of Django and all the pieces, you know, the beginning pieces, intermediate, advanced pieces, uh, so you can check things off and know, oh yeah, there's signals or middleware.

Um, and along the way, the biggest thing for me is I wanted to have a detailed drawing of the request response cycle so someone can see the file system, WISGI, again, middleware, like just how these things all connect.

I'd never seen that really uh when I was starting out.

And there's some talks now that are quite good that go in detail on this, but this was a high-level overview where we built up all the pieces.

So that's the goal.

A beginner level talk.

I want to do a deep dive talk, I think, next time, but uh I hope this talk will be helpful for people.

I thought it would be um very handy in terms of like just that uh that orientation.

Like when you come in, it's like, oh wow, there's this kind of uh flood of information that you get.

It's like, well, you know, here's the auth system, and like that's great, but like to appreciate it, you have to know what sessions and cookies are, which if you're new to web development, you have no idea what they are.

So, you know, I introduced the auth system and was like, here's a very very basic configuration, and then here's quickly how sessions and cookies work, because you should know that.

And then all auth and yeah, just trying to go.

I mean, I ran out of time, I went like 3x speed for the last the last couple minutes because I was so nervous I didn't look at the person giving me cue cards.

But I was trying to, you know, basically how do I distill 10 years into just here's 45 minutes of all I can give you.

You know, because there's thing is even the command line, like I have a little bit on the command line because like just this morning someone emailed me for my API's book where there's a section where I say open up another uh console because you're running the background server, the Django server, and then you spin up a React one.

And this person was like, I don't know how to open up another console.

You know, I mean it seems trivial, but stuff like that happens all the time, and it's just like a full stop.

And it's it's not the doc's job to do that, but I'm fairly attuned to those pain points.

So I try to make the ramp up as smooth as I can and take 10 seconds to say, hey, this is how you do that.

Or, you know, I mean, I think a talk I could do off the top of my head that I mentioned was like choose your own adventure Django.

Like, there's so many things where it doesn't really, you know, do I have a period at the start project?

Do I not?

Do I how do I configure my URLs?

Do I use function-based views?

There's so many things that for us, it's like, oh, here we go again with these religious debates, but it it's a full stop for a beginner because they lose confidence.

So trying to stay,

But for each one, you're like, and of course, there's this question or this question.

If you're a beginner, that might stump you, just pick one, doesn't matter, you know, and you move on.

But yeah, you know, I remember um beginning, and I don't know, I had some I don't know, version conflict with some PHP thing, and literally I was just right, that's it, I can't progress anymore um yeah yeah no until I was able to find a workaround.

So yeah I've been stuck for I've been stuck for weeks on stuff, you know, which is such a waste of time.

Um but I didn't know any better the Django forum didn't exist.

I didn't know how to ask for help.

You know I think hopefully this podcast one of the things we do is and one of the things I'm thinking about this with the mentorship conversation I had is beginners don't know like when they're stuck stuck.

Right?

It's like you know you gotta put right so if someone comes to us like I saw this at the conference for you.

I mean you're you're quite the celebrity people come up with questions and you know if somebody has made a go of it and you can tell and they say okay I'm trying to do this thing and I research this this and this and they're just stuck like that's great.

Like put your hand up.

But if sometimes people are just being lazy right so like you know don't be lazy but don't don't spend days on something that you could get help for.

Like you know ask ask in the forum ask in Stack Overflow.

There are places for people to get you unstuck because that's I've been there and it's it's you know you shouldn't be that stuck when you're starting out.

That's not you're not getting the value out of it.

But it you know sometimes you do need the answers in the back of the book to help you forward.

And the the tr the challenge I guess for beginning is how to know which is which is this just something I've got to work out or am I really stuck.

So I try to resist that but I see the appeal at a certain point.

But anyways enough about me let's let's talk about your talk because you updated it a little bit from the European version.

I don't know why I

So I I again same as you I was rushing right yeah and I was flicking through but I think if you watch both talks you get a full you know

rite about build a talk around that.

Yeah um but the idea was that I mean async's really complicated um and it it just as a matter of plain fact that once you adopt async in your application it's like twice as complicated or three times as complicated you know you're just making it harder by doing that.

So okay well what's the benefit and when would we use that?

And so I just introduced a few base um a basic async IO example to show a couple of things about the event loop and then you know can you use that for background tasks in Django?

Because if you're running under Asgate you sort of kind of could but you don't have um like all the error handling and retries and scheduling and all these nice things that a proper queue is going to give you so you're still probably better off using a proper queue implementation than just a sort of quick and dirty create task with async IO.

And then what oh my favorite example is aggregating views so um for years I've been uh building REST APIs and then uh have you have a client app that wants to fetch I don't know a list a detail view plus a related list view all in one go so it only has to make one HTTP request rather than you know multiple because on a mobile application it's going to be slow and battery training and and painful.

And so you create a little proxy view that sits inside sits in front of those two REST views makes the the the the rest views and then aggregates them and returns them and we always used to do that with NodeS because you know that was what you did.

But now with um async dev views in Django even if you're running under WISG you can write that view in in just in Django and it keeps your your code much simpler.

So you've got the example I gave is a hotel detail with a room list um so you've got a hotel with a list of runes and prices and you want to combine them to return one.

And instead of spinning up a separate server instead of having you know a different technology stack you just got a Django view with you know perhaps HTTPX as your async um HTTP client to fetch those and combine them.

And I just think that's such a beautiful pattern and it it I've been literally been using it for years.

I never got into GraphQL because I had this pattern and I was established with it and I never felt the need for the complexity of GraphQL.

I've I've never um architected an application that used GraphQL I've worked on applications that use them but not they weren't my creations.

And it kind of feels like okay maybe I've maybe I can just never have to learn that technology it's like

Like the fundamentals don't change and you know you don't need to have the burden of feeling like being left behind especially with front end stuff.

So yeah I know I really liked your your talking

to go well both times even though I was rushed uh nearly Katie uh McGrocken who was chairing she nearly had to go and get Kojo who was running the the lightning talks to come and time me out um but fortunately she let me go on uh for a few minutes extra

but that was it was the sprints were the sprints were fantastic I mean I it's just it's much smaller group and it's sort of like the you know the core core core people who really want to be there.

So you know that morning I was gonna go see your um uh you know contributing to Django bit and I got there right at nine o'clock and you weren't there because we'd bike down to the ocean and stuff and so I went and had breakfast with like Jeff Triplett and Jacob Kaplan Moss and ended up talking about the board for an hour.

And then I came in and that whole morning, you know there were the how many were there were at least a half dozen or more, you know, to be clapping when someone else did a new commit somewhere to Django based on, you know, you and Marius were running around.

And it's just like in a couple hours, there's so much progress made.

It's really amazing to see.

Um, and so the last few years I've well last few years.

In 2019, I ran it once.

So I went to um Copenhagen uh in 2019, Gen Con Europe, and I was trying to lead new contributors, but I was doing it on a sort of one-by-one basis, and so I ran round and did the same sort of talk 50 times, and I was exhausted and my throat was dry, and I, you know, I couldn't be in all the places I want.

And there were still people who were sort of lost as to, well, what can we do?

And um, someone suggested to me that we'd have a uh like some there needs to be some kind of structure.

So in Django Con U and Django US that year, 2019, I did a workshop and it went really well, and we had a like four tables or so, and that was just to try and get all the people who who perhaps haven't got a project to work on or want to get tribute into Django together, and then they can sort of help each other.

Um, and then I ran it again this time this year in Porto and it went okay, and again here in um San Diego, and it went, I thought it went really well.

Uh it was really successful.

And the the the the all it is is get set up with the test suite and have a go at find a ticket, have a go, browse through the track.

Because there's the thing is we've got in Django, we've got a thousand open tickets just under.

And so if I give you a list of a thousand tickets, you're like totally overwhelmed by what could I possibly contribute to.

And then there are no easy tickets either, like because it's 17 years old, right?

So there's nothing where you just come in, snip something, and there's your commit, and it only took five seconds, right?

Every well, no, not every single ticket, but almost every ticket requires you to sit down, read it, open the code, have an explore, become to understand it.

And I have this idea, I have this little spiel to people.

People say to me, Well, how can I contribute?

I'm new, I don't, how can I, how can I suggest anything?

It's like, well, because by the time you've spent that hour or two hours really digging into it, you are literally, literally are the world expert on that ticket at that moment.

And the goal of the workshop is literally to get people to have that experience so that they realize how it is that just by giving it a bit of time and love, they are legitimately able to contribute on an equal footing with anybody who's been you know who's been around 10 years and you know, millions of commits and all the rest of it.

Because, you know, they they someone will come to me, they've been looking at a ticket for two hours, and they're like, Oh, Carl, what do you think about this?

And I'm like, literally, I don't know, you have to explain it to me.

I may or may not have some memory of that ticket from whenever I've come across it previously, but right now you know you know 10 times more about that ticket than I do.

And then you can have a really good conversation, and then they could they go, okay, well, I could do this, and they open a PR, and it's like, yes, there's your first PR.

People were, you know, even though there aren't super low-hanging fruit, people were finding stuff and they were submitting, you know, it's just great.

It was like almost like a gong, like a sprint at work or something.

I mean, there was every half hour there was another round of applause, and someone else had done their, you know, first or another commit, usually a first commit.

I think so.

From that extent, um, I think the workshop was really successful.

My goal for that workshop is to have that freestanding.

Um, I don't want that to depend on me.

Uh, it has depended on me over the last few conferences, that's fine.

It can depend on me for a couple more conferences, that's not a problem.

But I would like to write that up and have that as a freestanding resource that then, you know, other sprint leaders could choose to use or not, or adapt or whatever, you know, in future Django cons.

Well, and and there are so many good ideas that came out of the conference and the sprints.

I mean, one is um Paolo just wrote wrote up thoughts on the fact that uh core sprints.

So saying, you know, do we have can we just have two, three days for people to come together um and do stuff on Django?

Because his experience, he'd uh the Plone framework, he'd use that.

Um and so he discussed that with a number of people.

He just wrote up a blog post on that.

You know, that would be really exciting to do, especially if we did one, you know, we could do one in the your side of the world and one in the US side of the world, and just a long weekend, and then someone, I think it was Pamela Fox, said, Well, you know, can we do uh virtual or hybrid?

Some of us can't, you know, can't leave for whatever reason.

Um, and yeah, we could, you know, I think just it's just like a modicum of structure uh allows for these things to happen.

And that's something that, you know, we can talk about the Django Software Foundation board, but I think that was the kind of big takeaway is there's all these ideas, and we just need to get the organization right to have them happen, right?

Because people on their own need a little bit of that.

We need um so Simon Shirk gave a keynote on the state of the ORM, which just fantastic talk.

You've you know you really must watch that when it comes out.

Um, but one thing he talks about is the contributors to the OEM ORM, and I've sort of over time the pool of it, the pool of people who can review PRs to the ORM has never been big, but we need to make sure that it doesn't ever narrow to you know two, one, zero, oh dear, what happens, right?

We need to we need a way of bringing on people and over a release cycle, because let's say you add a feature to the ORM in, I don't know, and I think he did mean a full release cycle like a 4.x or a 5.x, so that would be 4.0, 4.1, 4.2, not just 4.1.

Because let's say you introduce a feature in 4.0, guaranteed that there are regressions in that that don't show up until 4.2 when people upgrade from LTS to LTS.

And so if you've introduced the feature, you kind of still need to be around like that couple of years later.

Um and but that Pete, it's not realistic, I think, for people to just come along and uh you know open the open the source code to Django.db and start reading.

I mean you can do that, and that's a great thing, and I do recommend doing that because it's not that scary, but you know, to to get a comprehension of the ORM from that is is asking

And unless you have a direct need that overlaps with work, it's you know, I think like Mark, like you know, Marius um Django Fellow had that with Oracle with some of his work.

I think Simon at Zapier had there was some overlap and then he continued it.

But um not a lot of people have jobs where hacking on the ORM is you know, they're doing things that require that.

And as well, I think over the last kind of year I've come to realise that um it's no good us just asking for more contributors because it like if you're if you're basically economically privileged, then sure you can find time and we get the same the same sort of um demographic, like you know, essentially white blokes from North North America, North America, Europe, you know, the the the global north, so to speak, um, who've got the economic capacity to to give that kind of time.

But we won't get um people from a wider uh demographic or you know more representative of the community unless we're able to say, hey, you know what, there's a there's an opportunity grant here that goes with um with this mentorship so that you've got the capacity to actually do this.

So we're not asking for free labour from people who are already struggling to do it.

I think you know, I've previously done talks on your web framework needs you and your web framework still needs you, and I'd get contributing to Django and how that does benefit that benefits you and it benefits your career, and it's a great learning opportunity and all these things, but that doesn't that doesn't remove the economic barriers to people contributing.

So I think as part of thinking about mentorship, thinking about the future sustainability of um the of the maintainership of um Django Django, of the Django projects, the Django Code, we also need to be thinking about opportunity grants and funding for that so that we can widen access.

I mean, not for everyone, but for many of those people, I think it would add an additional layer of um it would feel good, right?

I mean, yeah, you know, because you all aren't you all aren't gonna be there for forever, but if there was a sense of handing off the torch in good hands, I think that would either if if not prolong how long you'd stay, just make it more rewarding to keep doing it again and again and again.

I mean one thing contributing directly, like just fi fixing tickets, it's it's great, it's good fun.

And I always in in a way it's kind of like the feeling I have about freelancing after freelancing for years and years, was that great, it's brilliant, freelancing is brilliant.

You pay, you know, you get a good paycheck and you you know get to work on interesting stuff.

But the end of it, you're kind of no closer to your long-term goals, and when the paycheck's gone, you're back sort of where you were, and so you have to carry on.

It's a bit like a treadmill, it's a bit like, well, hang on, how can I how long can I do this?

I mean, till I'm 80, perhaps not.

Um, that's an interesting question.

But the same with fixing tickets, it's like, well, I can fix tickets, I can fix tickets, I've been fixing tickets for years now, but there comes a point where it's gonna be like, oh, you know, I haven't quite got the energy to fix tickets.

Where if I can help people come along and and you know, learn to fix tickets, it's like a false false multiplier.

Well, and it's also this sense of like I was joking with you about something the other day, like, you know, self-deprecation.

You know, it's like you don't want to just leap out, you know, call to their

Yeah, he will he will be removed in Persian.

So that's like I wrote up some personal thoughts on the Django uh Software Foundation board because I've been having the basically this whole week has been me having discussions with people about all this because I really and I I've seen a lot of people are stepping forward, I think, who weren't fully aware of the need, but also weren't aware of the the possibilities, you know, of what like what you know.

I s I sat down with uh very prominent people and they're like, I have no idea what the board does, like I have no idea what the budget is, you know, and and so it's not a scary thing, it's like a you know, here's what it does now, but here's what it could do, right?

Like, so I would at the moment, um having been on the board three years, it feels like whenever we come out of a conference and things and we say Django should do all these things, the only way to move it forward is the board.

And the board already has its hands full with stuff.

So I think the board, I would like to see it get in a position where it's managing more than doing.

Because right now, we in the board do a the great deal of all these other things.

So I I'll link to it.

I wrote up a whole long thing of thoughts on how we can do it.

I think it'd be you know, there's a number of people stepping down.

I think it's great to have new people come in and come in with at least one thing they want to push.

Um at the same time, like Anna's stepping down, I'm probably stepping down, a couple other people are stepping down, I still want to be involved.

You know, I want to be on a working group.

You know, the mentorship thing.

Like to me, I want to make that happen.

So uh maybe you can't do that being the treasurer and all these other things, but if you're in a working group or if you're on the side, you know, you can push that forward.

So that's what I would and people have said again, people have said, hey, is there is there a way to contribute?

And you know, no, we don't really have that structure.

So we need to, you know, working groups, whatever it's called, we need these areas where people will say, a couple people will report to the board every so often, ask for um approval or funding as they need.

That's what like Paolo did with his course prints.

I mean, that's what I I said to him.

I was like, you know, please make an ask because I the board should exist to approve and help things, but it can't like we already are doing a lot of stuff.

So, anyways, offloading things and people, new people coming in with things they want to push, and just yeah, wintership, DSF grants, we can do that.

Django project.com site.

Sorry, go ahead.

You know, okay, he goes to he's been to every conference and he speaks a lot, and he's Andrew, but it was just like, oh, like what am I even doing here?

But yeah, he gave a great talk on the state of Django, among other things.

Um one one thing amongst those was um he wants to uh not just him putting forward ideas, he wants, you know, if you've got an idea, come forward and let's you know let's propose it.

But um, you know, two or three big projects that if they could be funded projects, um, that would make them much more realistic.

And the the whole point, I think the point with the Django grants is that if there are concrete ideas, it's easy to it's easier to put together a fundraising mech um drive, you know, and this money would go towards that than oh, you know, if there's nothing specific on the agenda that we want to address.

No, absolutely.

So, how would we do Django Software Foundation grants?

Well, we have a technical board, they could in one meeting agree on five, ten things that can be done in a couple month period of time by the right person.

They could review who that person is, and we on the board could provide funding as needed to say, okay, like bring serialization into Django, okay?

Like the technical board has scoped it, like this is what we're thinking, and then we can review, you know, over the next year who potentially could have the time, okay, and then what are the funding needs based on your personal situation?

Like, yeah, boom, like we can do it.

Like, I but I think again, that's what it requires, and that's what um the board or working groups like we just need that extra thing of someone to, you know, and I'm planning to push this, like, but I can't push everything.

No, you

But like it's that's to me, that's the only bit in this whole ecosystem that's missing.

Like, we can get the funding, we have the people, we have the ideas, we just need that organizational structure.

And that's you know, I I I was able to talk f multiple times with um Jacob Cap Jacob Kaplan Moss, who has said publicly he's gonna run for the board again, which would be great to have you know a creator on there.

But his a lot of his focus last few years has been around organizations and trying to and he he writes wonderful pieces almost every week about this.

So people who can bring knowledge from the PSF from private companies into rethinking, you know, how Django is structured, because we haven't changed it in a while.

Um and we we can and should.

And so, okay, we've got a chance as we start to come out of the pandemic to re-energize.

And you know, that was the thing, to see everybody.

That's the the opening point was that seeing everybody is really re-energizing.

And it's like, yeah, okay, we can do this.

We we there's still this heart of a community here that really wants this and is really passionate about the framework and the community and the ecosystem.

Let's let's make it happen.

Um, there was a panel.

We were on a panel with uh the end.

Yeah, there was a a state of Django panel that um Don Wages ran, which was I I I I wish it could have been a lot longer.

Um I'm sure the audience may not may have felt that way because it was the last thing on Wednesday.

But it was uh people

Like what do we say in the podcast all the time?

So it was it was you, me, Marius, um uh was it Rachel from uh Django Group.

Rachel Cahun.

Yep, yep.

Um Daphna PSF Revsis.

Um we we covered a lot of yeah.

You know, watch the panel when the videos come out.

But Jeff made, I think, a good point is that you know, uh perhaps uh uh historically Python sort of looked to Django as, oh what's Django doing?

Let's do that.

But in many ways, in terms of um particularly diversity in the steering council, uh Python's pushed on in front.

And you know, this ties back to your point that you were talking about, Jacob.

Uh, we haven't updated the Django structures in quite a long time.

And it's perhaps now time for Django to look at what's being done uh in Python and see if we can bring some ideas back and you know update.

We did we had Dep 10, which was to widen participation, but then the structure of the rules haven't quite resulted in uh like the technical board as being the same sort of group of contributors and not as pluralistic as it could be.

And so can we can we adjust the the election rules there?

Can we loosen them?

Can we can we allow for people to tag out of um of contributing and then say, yeah, do you know what?

After my break, I'm ready to come back and give back to the community again and have a mechanism when they they can be on the steering board rather than oh no, you haven't been contributing, therefore you're not you know you're not um what's it called, eligible.

Well, there's uh pretty active discussion on um Django Developers, uh Google group and also among DSF members about about this.

So that's

Well, and I think it's just we haven't been together, so like even just on the board, I was the only board member there, but like you know, we could we the board could just do a lot of stuff, but if we don't talk to the community, we don't feel like empowered to do that stuff, you know.

Like these in-person events in DjangoCon Europe, I think there are three board members there, so

And Heim made this point about in his um talk at DjangoCon Europe um on the state of Django.

I can't remember the exact topic, but it was one of the last ones on Friday.

But um, he talked about well, okay, if you don't listen to the podcast or if you don't subscribe to Django News, or you're not a met an individual member of the DSF, we've got no way of contacting you.

Um we've got no way of outreaching.

So, okay, there's the blog, but you know, we

And it's like, well, uh, you know, have we got other can we think of other channels so that we can reach out?

Again on Django Developers um just this week.

So uh I think prompted by the discussions coming out of DjangoCon, someone asked a question about well, what is an individual member and where's it about?

Because if you look on the website, it doesn't really say, and then you know, um Tim Schilling who maintains um Django debug toolbar and has done for years he's like well I I've only just applied to be a Django a member of the DSF clearly he would be you know a candidate but he he doesn't know what it's involved.

Yeah and you realize someone like this who's a pillar of the community has doesn't really know and a lot of people wrote in saying I I yeah I you know either I don't feel like I'm qualified um which is you know it can be anyone or I don't know what it is involved and therefore I'm hesitant to put forward because what am I stepping up in for you know yeah yeah even even people like that.

you first.

Like we could just like Python has one like I just sent it around to some people we can just put it on the sidebar um but here's the thing it doesn't it can't be a board person like so like I'm happy to do to be the working group of the newsletter but uh you know I've there's other things to be done but just like put it up say someone all you do is once a month send it out repeat what we already have we can put down the line sponsors on there like boom like I we have to have a newsletter like that's that's kind of what everyone has like it's it's 20 minutes

But anyways yeah hot take um so I got to sit next to David Lord at the Sprints who maintains Flask he lives in San Diego and um knows Andrew Godwin and some of the other people um and that was interesting.

Uh so he you know he he he has a full-time job and then he single handedly runs Flask and I can't remember all a whole bunch of other things in the ecosystem and he he made the point which I hadn't really realized that it's that Flask is downloaded 10 times as much as Django.

So you know okay we sit here and say well you know yeah Flask is for kitties and add-ons and you know it's not for you know you do you you know snarky things like you just rebuilding Django but like it has unbelievable usage and he was saying a lot of times in science and other fields where people don't want to be a web developer they you know but but it's a professional setting they're using Flask.

But anyways he's managing all of that and he was asking about our you know our board right because he was like wow seems like Django's figured this out right like there's a whole community it's not like literally one person who took it over from Armin.

And you know some of the things we're talking about he's like I he's like you know so he's being a fellow and a core contributor and everything doing and a board topper and all the rest and not his and not his not his day job.

And so he's um was asking about you know yeah how do we structure things is it possible to have kind of like a fellow like role but then there's this huge trust component there's no clear person to him in the community who could take that on even if he had the money and everything else to do it.

One thing he's done is he's recently gone under the branch of the PSF.

So they if you contribute to uh Pallettes which is the Flask organization it routes you to the Python Software Foundation and in return for this is public um I think 10% they handle all the administrative they handle the processing fees you know they they let him which which is great like because otherwise he can't you know so so PSF is expanding its umbrella to these areas because most of these things aren't Django they don't have conferences they don't have this wide community even if they

have insane usage and so they can provide legal accounting um and so that's something that we Django can also do Jeff with Jeff Triple has been making this point you know can we have a like a more formal liaison with with the PSF and actually tomorrow uh Friday um Jeff has set up a call between the PSF executive director Deb Wilson uh Deb Wilson the new one um me and Anna are president so we're gonna chat it out and that

's something that you know if not a board member we should have some direct line of communication beyond like I have one with the accountants for conferences but like anyways long rambling thing that's my hot take is that wow flask wow David Lord and like we in the Python community can and should try to pool our resources and share tips on how to make this all sustainable.

do that.

ess them for it.

So what's your hot take?

es it blows up.

And there's in the existing docs there's this uh timeout you can add to you can add a number of seconds that it will sort of re-retrive for before it's uh raises the same error.

And then there's this other thing that the WAL or the writer head log where uh you can write much faster uh to SQLite.

And um I was like but we we really have no idea how fast this goes.

So Simon in his brilliance just whips up a benchmark puts it all together and he's like yep if you uh with the default journal mode and you know no no timeout settings and all the rest it it you you hit this error quite quickly which is you know the historical advice but if you set the timeout to about 20 seconds it kind of disappears for some you know you scale up quite a lot before you hit that error again but he was like but then if I enable the the well mode the writer headlock um it just kind of disappears and you know 4,000 requests

a second sort of thing and multiple gunnercom workers and you know all the things that you think no surely this is going to to um blow up the the the locked error just kind of went away um and uh that benchmark needs a bit of work if they're moreistic right more realistic or different write patterns which will create which will trigger again trigger the error because uh you can you know the write ahead log isn't flawless if there's a there's a these checkpoints I I forget the details it's all in on the SQLite uh

docs but you can hit errors but why aren't we hitting them in in Simon's benchmark you I mean you want you want the benchmark to hit the expected error at some point right um but the figures he was getting out of his um example were more than enough for you know almost every website you've ever worked on

what we say because for a lot of websites it might be that that's enough and then it takes you know if you if you're running a small site and you're worried about cost it takes yet another cost center out because to spin up a a an RDS or another managed database instance it's not free at all.

Whereas if you've already got a server and you can just put a SQLite folder file on it you know it essentially

So you can have persistent storage instead of an ephemeral thing where it gets wiped that's also seems like um yeah but no that was I mean

And I was speaking to Simon Chalette on the uh way to the airport about it and he was like well yay but but you know if you need access from multiple hosts and you need to or need to migrate to a different one it still may not you still may be better off um starting with C um with Postgres anyway so that you know it's not a totally one-sided story that you should all just use SQLite but for you know where in that measured case you you you know SQLite fits your use case why

what like for these for these toy sites and for in most cases like yeah you can do Postgres I'll show you that but you can also just use SQLite.

That would be that would lower the barrier.

Well I as as you're talking one thing I maybe maybe the thing I I feel like my biggest contribution is at these sprints I was sitting with uh so uh so David Lord came over and then he was talking something about like automating deployments and asking questions and um someone was like well Simon's or Willison is over there he's like you know he's got like hundreds of projects and you know should talk to him we kept talking and you know I'm for a developer I'm outgoing so I went over and said hey Simon like could you come over here you know I was

like are you in the middle of something he's like well yes I was like can I bring you over here he's like okay so bless him he trusted me and Simon came over and and he and David um had never met like I just sort of assumed that these original these you know these these gods of web development all are are chummy and some of them are but it was like amazing to see it was just like oh like hey I'm Simon like I'm David like obviously they knew who he were but they'd never met and then you know they and then they were off to the races for you know I I I had to leave for the airport but like

an hour in they were showing like all these little things and hacks and so it's a beautiful thing to see but it's also a reminder like yep it's not like everyone knows everyone right and especially at these conferences the bigger point I wanted to make is I think Tim Schilling was tweeting about this but you know most developers are on a little more on the introverted scale so like it's just natural in conferences not every interaction is going to be smooth but everyone is there and trying and and wants to engage.

If they don't they can go to their room you know so um taking that extra step to you know if you know that two people might work well with one another don't know each other is always worth it.

Yeah no exactly and like even Simon you know superstars like I say Simon like even you know even superstars that that you think oh well they must be really that they can be shy people you know too and they're they're feeling it just as much as you are when you're a bit nervous about talking to people like it's it's not it's not you it's everybody.