Transcript: Django Fellow - Jacob Walls

Hello.

And we're very pleased to have Jacob Walls, our newest Django fellow. Welcome, Jacob.

test models from 18 years ago and there was a podcast class and a vodcast class i was like

jingo chat has graduated to vodcast oh i like that wow levels okay all right well let's carlton top

is sponsored by hacksoft your partner be on code more on their services later in the show okay

stickers t-shirts being printed up we've got everything and jacob you're gonna make it i'll

we're going to be power and i said about me sat there doing whatever and so we um i think we've

got yeah in i think 15 maybe you know a few more we'll sneak out the um sneak sneak out the wood

workers the next time but that's a nice number we can split into groups and um there's a good

number of um experienced contributors and there's a few um brave souls who've not you know coming

along and you know they'd be mailed out and they'd be like are you sure you know i'm not you know

absolutely absolutely um very excited um so anyway yeah it's you know for the you have to take a

day we're going to do the roadmap um road mapping workshop that we talked about got a django part

um django birthday party for the for the second day and then on the third day i'd like to do a

kind of um round up you know what have we achieved where we're going what we're taking back to the

community um so everyone's clear and you know hopefully thinking about um pushing django on

that made in the future versions you know paulo and i've done this kind of um by ourselves this

time as a proof of concept looks like the proof of concept is going to work and it's okay how do

sponsored the new just with the newsletter but the you know the mailing list is you know the

important communication channel and then um in collaboration as well with the gentleman

who provided the venue um which i'm very grateful for so um that's just yeah it's super all right

a django con us and talk about this but i mean we could briefly do the full circle right so

sarah boyce is is going to be on maternity leave so then we're going to be down to one

or half a fellow because natalia is three days a week um carlton you're on the fellowship

so so you find out oh we need to carlton get a new fellow how does that like there's a blog post and

then you filter 100 submissions or tell us yeah okay so um you know sarah obviously knew that

advance and then there was a certain amount of well we've got um we've got to put out the

an application a call for um applications and we need to do that within a certain time scale you

Now, you know Django within a certain time scale is not Django's forte, right?

Django is as long as it takes, but we keep going forever.

Except the releases. The releases, though.

No, but the releases have a cadence, right?

And Jake can tell us all about that momentarily.

But to have to do something in quite a short period of time really is difficult.

And then we put out the request, and we had over 100 applications.

And many, many of them were just really, really, really super.

And so, um, that was, that was quite encouraging, you know, when you, when you get that number,

but it was, it was a lot to go through.

Um, I was handling the first contact.

I'll just say one more thing and then I'll let Jacob step in.

But, um, I was handling the first contact and my, my hack was to read the applications

as they came in.

So that, you know, rather than touch it once.

Yeah.

Yeah.

yeah because 100 all at once is a big is a big lift well so i was gonna say so obviously you

what uh what about you know what about jacob stood out to the committee well just to make it awkward

into Django before, um, even applying for the fellow application.

So he's all very, he's already very involved, um, in the core framework and he's, um, on

The review.

Being around and being active.

And, um, he's very bright.

I don't tell him I said so, but he's very bright.

And, um, he, he was one of the, um, one of the, um, bunch that we identified in the first

round and then we narrowed it down and, you know.

yeah and in the end we would he was the candidate we made the offer to it's you know and he accepted

sent out you sent off the application but then there's quite a bit of waiting too on your end

and i acknowledged that like you know as you can imagine there were a lot of applications to get

through and for me it didn't feel particularly long like i didn't have a lot of feedback to give

like in that respect that that needed to be changed next time um maybe that's just because of

uh having been uh having to adjust to the academic world where job searches take a year

so six weeks is like pretty good yeah yeah um but um yeah no i mean it's like it must have been a

real challenge to look through so many high quality applications um and then you know there's

volunteer board and you know a small team people may be in and out on leave and make sure everybody's

comfortable like i'm assuming those are all major factors there yeah and it came over the northern

i think if we um oh we would i think i don't think it's a secret i think the the hope is to be able

to put out and you know call for another you know other um people in this role in the year the next

year or so um to help grow grow it and maybe for short-term contracts rather than permanent

positions or whatever but the hope is to continue resourcing this i think next time we do it we

might need to think about you know making that first review phase having some paid time allocated

to it because um you know it's difficult it's difficult to come to a consensus between four

four or five people when those people are all on volunteer and they've got work you know if it's

only 10 applicants it's a much easier thing but when there's 100 that's all hang on um

quickly i want to go back you just said academia jacob so well i was going to ask him to say

the music composition path music composition and technology and that's kind of how i got the

programming bug like you meet almost any composer of concert music and they're all dabbling in

electronics it's kind of just an expected part of your bag of tricks um not hyphen right at the

beginning but you know um that came later um also between degrees um i moved to madison wisconsin

briefly and i was working in a qa job at a healthcare tech firm so that's actually something

that i think shapes kind of i didn't know i would go into software eventually it was just a throwaway

job for a year but like coming back to software later i kind of realized like oh my first time

getting my hands in the clay really with software being in qa kind of did shape how i think about

um yeah that that much experience with testing really did shape how i think about software

um did a phd in music composition and again was dabbling in technology along the way

um and then had a had a course in and python and programming techniques during that phd

and that set me up to right after my fellowship ended and i needed something lined up um there

was a sheet music retailer um that reached out to me they were looking for somebody who had music

theory experience python experience and or software testing experience actually primarily

and was local to the philadelphia area and i was like hey i'm that unicorn like all i had to do

was put those keywords in linkedin probably and get one result um so i ended up there they were

doing this interesting project in um creating a better sheet music search engine that wouldn't

just search keywords from publisher blurbs because every publisher wants to tell you that their band

medley is the best christmas medley that's ever been written that's not useful information so we

were scraping pdfs trying to extract musical information try to extract queryable ranges of

instrument profiles that band directors could save and performs and then covet hit and that wasn't a

money maker and i got redirected to all kinds of other stuff so anyway i had a really interesting

couple of years as kind of a python generalist slash project manager slash sort of finger in

every pie kind of thing and it was around that time uh where the senior developer i was working

with had left um to go start at meta and i realized like i need to learn more right i'm still pretty

junior i don't have a senior developer looking over my shoulder all the time and that's honestly

when i got involved in django because i i thought you know i want to learn more about the web and

there's this community where you can get involved you can volunteer some time working on pull

requests and tickets and you'll get feedback um and so that was that was that was kind of how i

got involved in django and then it just so happened in my next roles after that all involved django

and so i stayed involved and tried to bring my knowledge of what was going on in django to those

roles and say hey we could install this package or extend what we're doing with the orm look how

one of the best ways to level up one of the best ways to learn yeah that's exactly what happened

to me yeah i mean you get to see quite good code and quite clever people give you feedback right

those was a conference week i was at jingo con us in chicago early september yeah and how are you

settling on your feet i mean it's it's it's been great um i you know i'm in the lucky position to

be able to kind of start in it start in a role where i've been doing many of the tasks before

obviously i mean i hadn't been on security team before i hadn't been rolling releases before but

like I'd seen kind of the guts

of how jingle pull requests and tickets go.

So that part was relatively smooth.

I should also shout out to Sarah and Natalia

and the many people at the DSF

who worked very hard to make sure

that I had a smooth onboarding.

That was great.

But yes, it's a welcoming community

and there's a lot of trust

that's bestowed on you kind of right when you start.

the worst what's the worst thing that you can do you know

anyway will you were going to ask about something

i was just going to ask him to share a little bit about his background

mean so natalia's gonna be the release manager on this so you get to watch one and then i think 6.1

you'll be the release manager for jacob is that right yeah be my turn we just take turns

um so i don't know carlton we so i briefly touched upon it but there's

i don't know what maybe what's interesting to people i mean we have two fellows on here right

like from the outside these releases just sort of appear but it's an immense amount of work to

go from feature freeze to release like what if what do people what what should people not know

or what should people know that they don't about you know what fellows actually do to get over that

you've seen you've seen what's going on all the rest but getting that alpha out the door is not

you prioritize things

correctly so that you have as much time as you need to deal with any last minute surprises and

then inevitably there's a surprise on the last morning where it's like oh you know is this feature

ready to ship do we need to document this one thing or back out this one piece and so yeah

there was just a little bit of last minute hustling to just back out a small part of the feature that

cheering away but it was django tasks was coming in right and yeah tell us what happened there

I mean, it's, I think we, you know, it's a DE, it's a Django enhancement proposal that's been around forever, and it's being implemented in phases.

So Django 6.0 is going to have the basic interface, and it's going to have the backends for development and testing purposes.

So the dummy backend and the immediate backend.

I think the database backend is coming in a future version.

And I think that was a really smart division of effort.

So we just had the plan to merge what I just said, the interface and development testing pieces into 6.0 for the feature freeze.

And I went through multiple rounds of review from lots of people from the community.

Obviously, it had been implemented first as a third-party package, and so it had already matured.

Wagtail had already been using it. Others had been using it.

Sarah gave it a thorough review. She asked Natalia and myself to give it a thorough review.

And Jake Howard led the implementation,

did a fantastic job, super responsive.

And then, yeah, just the night before the feature freeze,

I think it was Simon piped up, said like,

did you remember to check this against multi-database setups?

And it's just like, that's the one thing

nobody had on their checklist.

And then like, you know, it didn't make me think like,

oh man, this is, there needs to just probably be a checklist

for every major feature.

Just has the integration against every major part

of Django been checked.

And it's just, we miss multi-databases.

Yeah.

So like, at least for me, whenever I started at new companies or I had people join my team,

I was always like the first like two weeks, write everything down because you're going

to get tunnel vision really quickly.

But yeah, checklist that's added on by future fellows is a good idea.

then you know i saw that well it's great everybody volunteered to help jake was like do you need help

the pr so she could approve it and merge it it's like but then we had we had backup there because

feature freeze i mean because there's two bits right there's the feature freeze where you say

right no more is going in and then there's the alpha and it's it's kind of up to the release

manager to say that the alpha needs to be a couple of days later it's like no i've got one pr that

i'm going to merge but nothing else that's okay but you know obviously the expectation and the

want is to get it out on that day so there's a sort of pressure to to release it um yeah but i

testing it. Come on, folks, get up there.

What you don't want is all the issues

to come in the day after the final release, right?

Okay.

Oh, okay.

Sorry, there's a little bit of lag. Carlton, you go

and then I'll... Well, I was going to ask,

what are you most excited about in 6.0?

I mean, we talked about the headline features.

Yeah.

if you've covered

the headline features i mean tasks yeah obviously it's um the having having a better development and

testing tools for tasks is i think going to be a huge win for developing uh being able to turn

celery off it's going to be great um yeah and then just in terms of like deep cuts or minor features

um i sort of asked around when i started i think my approach as a fellow when i started was to just

try to close as many open loops as i was already involved in before like taking on new projects

and one of them was a kind of pr that had been sitting in the review queue for for ages and

had been a contributor that was sort of what i think sometimes people call taking the vulture

approach where they take a pr that stalled and then they take up the mantle and try to run to

the finish line and this is somebody that i'd iterated with as a reviewer before and i was

kind of building expertise in an area i wasn't totally familiar with it was in postgres full

text search um and it was um creating an expression called alexine that you can use to sort of safely

wrap user input and compose queries with those pieces of user input and negate them and combine

them without having to escape everything by hand which is you know complicated enough that it might

turn you off from trying to even implement those kinds of features um and so i'm poking around in

a you know contrib.postgres scenario i don't know very well and um it got as far as i could with it

And so I kind of asked Sarah and Natalia, like, hey, you know, is there any bandwidth, you know, just reviewers or contributors been waiting? Can I just try to bring this over the finish line? And they said, yeah, if you have the bandwidth, go for it. And we got it into 6.0 within the system, a few reviewers at the last minute that were able to sort of validate that we weren't far off from the mark. So that was cool to be able to land that kind of out of nowhere right before the feature freeze.

and now that you're in a position to to do some of these things that maybe aren't as publicly

visible but you know that within the people committing code like that's a that's a big

deal it feels good for them feels good for you like you know yeah so i have one more um kind

your thoughts on how we go forward it always strikes me that um every release isn't it knocks

it out the park you know you know go back as far as you want four two great release five great

release five one another great release five two another great release we've got another super

release coming up this time and yet it's all done on volunteer it's all done on volunteer basics

it's just what happens to be ready at the time i wonder how you feel about that over the medium

pretty okay didn't you i might have done yeah you did because you asked sorry because it's behind

the scenes i'm pretty sure you posted something and then you asked me to jump in and you know

boost it or something one of us is losing our mind here sorry go ahead jacob yeah no i mean it's it's

like how can you depend on contributions from volunteers and to be able to ensure that something

is ready enough to merge on the other hand it's like if it ain't broke don't fix it you know

jingo's really lucky that it doesn't have like this you know invisible hand of corporate money

behind it like telling it what to do like that's very good for jingo because i'm not saying you

should change anything at all in that respect um and i think you know if this success is going to

continue if we're going to still having still have great feature release after great feature

release it's going to depend on people who just believe that you know new features uh new and

important features can still land in django like it's not that we're closed for closed for new

features i've seen that on reddit i've seen a post like oh you can contribute to django but

they just do bug fixes like you got to understand you know no new features really and that's ready

It is, but...

The major releases and going, well, look at those

and look at those and look at those

and look at those. That's true.

hear your co-workers talking that way,

I think that's my answer to your question.

It's going to take people continuing to push back

against that idea for

the success to keep happening.

I mean, Reddit,

for work, I'm on Reddit responding

to stuff, but you're only...

usually at a negative extreme if you're doing stuff on reddit it's there's there's some signal

there but it's a lot of noise uh i would say right if you go in the hop in the django forum and

you're still finding people pretty riled up about something that's a good signal because

that's a self-regulating community and that's where more quality happens so the reason i'm

to come out with some sort of key things that you know we can take back to the community and say

these were the ideas that were raised at the at the workshop but then do a teeny bit more and kind

of break them down a little bit so that they're addressable to contributors in you know the example

i keep giving is modernized contrib auth right that's just too big you can't do that but right

you know to break it down and maybe you know do a survey of the available two-factor auth options

in the community so that you know we know we've got a starting point well that's really addressable

to a potential contributor and i think if we can identify those and you know help to identify

addressable things then we can it's not that we haven't got contributors we've got lots of

volunteers and it's guiding them to things that the community wants to see go forward

um and that's kind of independent i think of having a big corporate sponsor who's paying

see those unfold

at the

sprints in

Spain and

then to see

that disseminated

in the

community.

part of it's

been historically,

I think part of

the perception

has been that

we haven't had

a visible

roadmap.

We haven't had

anything where

we're saying,

look, these are

the things we're

working towards.

I don't know.

Anyway, that was

kind of what I

wanted to ask you.

APIs going to

come up?

Because that was

one of the

takeaways from

DjangoCon US.

Nothing more to say on it right now?

Well, I fired off last time.

Jacob, what do you think?

I'm glad that we're not just ignoring it.

I think I'm eager to see how much we can do with less, I guess.

Are there forgotten bits of Django that we can build upon

that leverage parts of, I don't know,

doesn't doesn't require ingesting a whole new apparatus um yeah and honestly speaking from

my own personal experience like i can't answer that question i'd love to sit down and play

um and i don't have time during the week as a fellow to just yeah sit down and play

at least not right now not this week um so i'm excited to see what other people are going to

emma who's on the steering council she put dallas called she put out a um proof of concept around

the blog post that she'd written on oh yeah doing that's great i remember that post yeah

so there's a little package called django rest 2 i think it's called and it's it's tiny it's

lovely it's got like a list view and a crate view and you know it's got all the sort of crud views

and it's got a serialization method but if you go in this doesn't there's almost no code to it

and it's like, oh, that's kind of quite interesting

because it's like, I said to her that I liked it

because it was like FindCMS versus, say, Wagtail or Django CMSs.

You've got these giant CMS packages,

and then you've got something that's much more minimal

that just works with the Django admin.

Well, it's the same here.

It's like this is a REST solution that works with very little resource

beyond what Django already provides

to talk to your point about looking for something more minimal.

So I was excited about that.

So you just said, I'm going to bump in one more question

because you just said, and I really am going to let you,

you just said about not having time to potter on things during the week now.

How are you finding that aspect?

Because the federal role can kind of consume you.

There's so many tickets, so many issues, so many things going on.

arguments i guess you could you could you could suggest that like okay you can't be effective in

your role unless you allow yourself enough time for professional development and have a play with

something you need to have a play with and i would say sure that's the feeling i came out of

jango con us with which is that like oh i need to spend 45 minutes playing with this and 45 minutes

playing with this just to you know keep up with the conversation and so there's a time and a place

for that and i think it's going to come but also it's like i've started in the new role we've got

a we've got a backlog we want to start hacking away at i just kind of feel like that's the first

part of my tenure that's the part i want to focus on and um yeah maybe down the road there'll be

time for yeah trying to well carlton you're talking about guiding contributors if guiding

contributors effectively means having a play or just yeah doing a survey of the tools out there

i mean i don't know maybe conferences are also good places where that happens like speaking

with people at django con us i didn't realize that the just the force or weight of opinions

people have about things like drf like i knew people had opinions i didn't know they had

opinions about it um that was just interesting feedback together you know just uh yeah sheer

sheer weight of those emotions um but yeah yeah i mean there aren't many things that janganaut

because seeing jacob be the new fellow and you know go around the sprints i want to ask you about

that um but also i remember i guess natalia and sarah when they became fellows seeing seeing you

kind of pass off the baton carlton in a sense and then seeing them you know not pass it off but also

try to you know bring jacob in and so i was having definitely like you know it's like two you know

sliding doors or something right like or maybe that's not the wrong analogy but like two versions

of the same thing the circle of life circle of life no but i think it was good i mean i hope it

was good for you jacob right i mean i think that's the thing that you knew but now you really know

is just how much trust and

autonomy is placed in you as a fellow. I mean, I know, Natalia, in particular, coming from a

corporate environment, you know, still is like, well, who's, you know, who's the manager? And

it's like, well, you know, it, you know, so a lot of these questions have to kind of come within,

you know, how do I find time to be playful, right? Like, I mean, I, I struggle with this in my own

job, right? The irony is, like, I'm a developer advocate talking about code all day, but how much

time do i have to actually spend three four hours playing with something you know i'm getting better

at doing that so yeah i was just interesting to see and i i think i think you quickly picked up

on it right because a lot of the questions you had natalia also had and it's kind of like well

you know uh i probably even said like off the cuff to you you know in a way as long as people

online are relatively happy and the fellows committee is kind of happy it's really on you to

define the role how you want like there's tremendous trust unless you screw up

go through that and i i hope that you had a chance to meet as many people as possible because i know

you'd been to jango con europe but um even at a jango con us not everyone knows what how the

work um right even among those 300 people it's still kind of a yeah yeah i didn't come anywhere

people and next time i'll rectify that um but you know i met as many people as i could and

the n plus one right so the so the the n plus one rule is that for each each jango con you've been

to that's end and then each day you have to meet n plus one new people now that's that's okay if

it's your first one and your first one you've got to meet two people a day that's you can do that

But, you know, if you're like Russell Keith McGee and you've been to 20 of the things, it's going to get quite difficult.

HackSoft is your development partner beyond code.

From custom software development to consulting, team augmentation, or opening an office in Bulgaria, they're ready to take your Django project to the next level.

And thank you, HackSoft, for being a sponsor.

Helps us do this.

absolutely switching slightly so jacob what what do you want to ask the community right because i

think there's a lot put upon the fellows but you probably already see are there things people can

do with tickets or just online engagement like how can how can people help you and the fellows

i already hinted at which is it's just to like stay positive and like continue to believe that

like we do discussions will come to a resolution that you know features will land in django right

so not to be discouraged and take your energy elsewhere but to like stay engaged um you know

look at a discussion and try to determine what would move that discussion forward and then just

be that person um instead of waiting for i mean because it's funny you just will you just said

like even the fellows ask questions like who's the manager it's like there isn't one you know

so the same thing goes for contributors like be that be that man be that project manager right

help resolve the blockers don't just don't just wait for somebody else to you know whatever outline

the blueprints like help help make the blueprint with us yeah well there there is it may not feel

member and here we come to the the show uh jeff triplet you know so that was one of the things

people said at django con they're like you know the drinking game i forget who said it was it you

jacob someone said you know we have to i have to mention pie charm and we have to mention jeff

triplet as our django bingo but jeff is the uh i think the official fellow liaison with the board

andrew because that was very this on a regular basis to just hear how things are going and also

to play that role trying to remove blockers if there are blockers that are slowing us down and

so it's great to have that support so yeah i think we well speaking for myself only i feel supported

and i believe the other fellows feel totally supported right now which is great i mean and

a meet a monthly meeting being like uh how's it going any any requests you and marish you know

let us know um you know so there was some there was some but it was it was pretty much that right

the structure that coming in is it's a little bit more um specified it's a little bit more i mean

this is mainly natalia she's been driving this and it's a big credit to her and i've been really

impressed with the way she's seen the lack of our florian um apollona described it as a yolo

approach that we have and he's right it's a kind of and it's just been the way it's always been

we've always just kind of got on and and done it and it's been fine but natalia's come in and she's

been like no that's not fine because that's not um great for onboarding it's not great for

continuity it's not great for you know and she's absolutely right and she's identified issues one

by one and ticked them off as trying to make them better so you know just constantly impressed by

natalia's professionalism um but yeah it used to be you text me and i'll text you and we'll be like

why did we have that back channel to the board well because we did the podcast but it can't

it shouldn't rely right well you know like why only employees own or not employees but only

paid work that's on django is only route of communication to the board goes because we

happen to do the podcast like that's well right because we would have i think people don't if

areas and there's some people who overlap on both but they're meant to handle different things so

part of that started Carlton because we would have multiple times where

yeah no one no one was really in charge of the fellows you know it was like what do the fellows

think I don't know like you know that's kind of why I got in the habit of of asking you um and

there was a steering council it wasn't doing anything and but the entire time there's no

guidance coming through and no communication with the community I think yeah what we're trying to

do on the steering council is kind of get that going and if we can establish that in our tenure

and hand that off to the next steering council as something that exists then there should be much

more structure combined with you know fellows having proper management and proper guidance

and then the steering council that's actually helping to steer the technical direction and

then the board can do its thing without there being these kind of gaps but it was back when

marish and i were doing it was well it was marish and i and the the board organized the board did

the board bits but there was no other structure um yeah i do want to do you want to mention to the um

news newsletter and i've committed to uh you know we talked about this at jango con us actually

right like because the fellows take quite a bit of time to list you know reviewed triaged authored

talked about and some people see those but i think they put more time into them than maybe

people see and so one of our actually dinner conversations i think jacob right with natalia

was you know hey if there's a couple sentences that can just go in the newsletter as sort of

like the the headline and then we'll then i stiffed yes yeah i stiffed you last week i didn't

give you well that's i i saw that that's that's okay i i don't i i went in i went in to look and

And I was like, oh, wait.

No, it was a post.

So it's just hacking away at miscellaneous.

Yeah, yeah, yeah.

I try to.

And Natalia, of course, was like, okay, are you going to do it?

You're going to do it.

And I was like, yes.

If you write a paragraph, I'll put it in there.

So I try to look.

So I would, you know, my day would start

as a fellow crackopendashboard.jangoproject.com

and look at the numbers and like,

oh, there's a number if you take it quick, get on that.

There's a, you know, 32 PRs needing to review, right?

Let's see if we can pull that down.

It's kind of like a slight little bit of gamification.

It just needs a few animations and some, let's go.

Are you volunteering to add some?

Is that what you're suggesting?

to help help stimulate the fellow's enthusiasm for you know their tasks you see little you know

stars when they when they mark a ticket as um you know um ready for commit

i want to just mention one thing so what would you magic wand to change anything about django

right so you can mull on that as i say something else which is i think maybe the thing i'm most

sorry. Is that too much

the repository, Django

the community. All of it.

One of the things

I'm maybe most excited about, or not excited

about, but I'm pleased to see is that in

the Django News newsletter, there's a

updates to Django section, which Sarah Boyce

initially started, and now members

of the Django.not space often come in and

do where they'll say

we had X many

uh pull requests merged in and do a little description and i know that like i don't i

don't live in track like i don't always see like that's my entryway in to see every week oh you

know there was 10 15 whatever number of tickets handled and so i'm personally really i i kind of

feel like i wish we had more visibility to that point of how dynamic django is on the inside again

it's double digit basically every week are merged in um and so maybe i should go on reddit and find

that post and link it link to them but i'm really pleased about that i think we could do more with

it but i think it even within django like there's people you know if you're on the review and triage

committee like you're seeing it but then and i think this was sarah's like sort of insight was

like well people aren't even seeing that level let alone you know bubbling up to like a major

feature in a in a feature release so i'm really happy about that and i'm opened ideas on how to

shine a light i guess more on that around so both you the fellows you all just slamming through so

many tickets every week and then also things actually being merged in so how do we disabuse

people that notion that nothing's changing because you know stuff is changing anyways

magic wand so that that's my magic wand just to shine a spotlight now your turn jacob anything

more of a culture of like i'm running my project suite i'm well i'm i'm writing enough tests to

cover the functionality that's mission critical for me and i'm testing against the alpha and my

dependency isn't ready and i'm letting them know you know and it's the same thing that goes for the

community it's like when you're interacting on the repository you want people to be aware of

who they're blocking and try not to be the blocker and if you're a third-party package

the same thing goes it's like you shouldn't be the blocker that prevents your users from testing

against django 6 beta or whatever and and that maybe somebody needs to let you know that you're

the blocker and that person might be you running your project test suite so just i mean we talked

about this earlier you know carlton's saying it's going to be the day after the release you're going

start getting bugs rolling in about django 6 and it's like yeah that's that's what i would

setting that yeah setting that awareness setting that intention that you know this is just a normal

part of developing a django project to test against the pre-releases when they're ready

reported the bugs during the alphas or during the pre-release and who reported them the week

after the final release right they know oh yeah all that one yeah same again is it okay

isn't and it's it's it's not about reputation it's about um being involved and like if if you're

if you're turning up during the pre-release with the bug report look i've run this on my test suite

i've got this regression that that makes you know and it's you know it gets it it doesn't get you

it gets you into the community right it gets you ah like that it makes you a contributor in a way

that turning up the week afterwards it's not quite the same just isn't also there's a huge

It's like, we'll fix it before the release.

But, yeah.

I'll put a link to it. I've been struck by

there was a recent article on

Epic.ai talking about

these benchmarks for LLMs, and

one is SweeBench verified, which is

largely around Python projects.

But this is one of the big, big, when you see

percentage, you know, Sonnet 4.5

is 90%, and

gpt4 is you know 70 or whatever it's 46.2 percent of of it is django is like lms testing against

django like it's not even there's probably if i was smarter there's a way i could show it on the

screen for video we'll do that in the future but it's you know django is just there and has these

testing suites and other people are testing against it sort of um but it's not like there's

not tests and Django.

Well, I saw Adam Johnson

toot about this.

the LLMs were cheating.

They were using,

so they check out the repo

and then they were using

the Git log history

to find the commits by Marish

to solve them.

yeah, they're just Marish AI.

Yeah, yeah.

No, totally.

Let this be a lesson, right?

Shorter and less useful

commit messages all the time.

I'm debating if I can,

should I try to,

we'll just try it.

Come on, let's see

if we can get this software to work.

Oh, okay. Hold on one second. Let me, let me save images. See, we're just learning. We're doing it

that game. So we'll put a link to it, but this is, so again, this, this, we verified, uh, sweet

bench verified um you know and it's uh so the point of the author i think was like you know

it's pretty low diversity of code bases testing this um but you know i mean and why why is flask

0.2 percent where's fast api you know you know anyways interesting is this one of the reasons

because django is 20 years old because it's so well documented because it's so established it's

just sitting there right so it's as mature as any code base is um so all right look at that

all right oops now we're in the wrong area

the issue that you were talking about i opened a discussion about whether we should change jango's

versioning number um yeah one idea i just put it as like well we could use a cowl verb because

i think the problem people have um and i see it after this the point zero release and the point

you know that's what sort of made me think of it again was that um 6.0 will come out and it's in

it's no different than 5.3 it's not it's not like a major version in the in the semantic version

in sense of it and so we it we go point zero point one point two and that's the lts and that's the

cycle and it it has significance but you have to know quite a lot of django before that becomes

meaningful to you and i know will you talk you've talked about this is why i asked you to comment

is that your books people you know they're 5.2 ready and people are like oh but what about 6.0

it's like the 6.0 is the same as 5.2 in almost all ways yes i mean you're just scratching my

um yeah it's it's it's too much to have to update every eight months um but you actually don't need

effectively nothing learning django book learn django book yeah that's not changed in eight months

tell when someone writes in you know some issue oh this isn't working i'm like oh you're on the

you know yeah yeah i can tell which which version or pirated version they're on but uh

yeah it doesn't change that much it almost never changes there's maybe there's like one small thing

that and i'll usually do a blog i'll do a tutorial on it and if they just search for it they can find

it but yeah it's incredibly stable and um i don't know look look at python right like i'm not saying

python's perfect with like every year is like 314 315 316 but it does present a feeling of

stability even as we know there's some python releases that are a lot bigger than others you

know like 36 with f strings right like they're not all created equal i would say from the outside

maybe maybe hugo feels differently but anyways yeah i'm in favor of it i mean i've sort of given

up on trying to push that

up the hill but there was

the discussion I think probably

what the status quo until someone comes up

with a better idea but

it was just acknowledgement of

Carlton and I were going to do an episode on the Django survey

results which come out soon but one thing

again it's I think it was 4500

people so it's not the whole community but it's quite a lot

of people like 83%

something like that claimed they were on the latest

version of django so not even the lts like the latest latest version that feels high to me but

i think it's great if at least 4500 people are doing that right so maybe you know i don't know

the lts thing is right is great but it's a it's a real you will find jacob it's a real burden for

the fellows to support for three years um if that if you if it's not already apparent yeah well i

reporters to the policy take some of the heat off yeah all right so we're i think we're at the

are trying to start a new thing so i'll go first uh project i want to shout out django watch files

from adam johnson so this just came out recently so run server it auto reloads and it goes through

the whole file system and um watch files is a faster written and rust way to do it and so he

implemented this is a third-party package so basically just refreshes faster i think it's

fair to say, and

works... Yes, Carlton, go ahead.

way less CPU to do so.

Oh, sorry, I forget, yeah.

And so the key point there is if you're developing on battery,

that is going to make a massive

difference to the amount of time you can develop

with versus

the built-in stat

file stat run timer, which you're probably

going to use 10% CPU,

which is just ticking away

from burning your battery all the time, whereas the

watch fast approach is much more

mentioned if it picks up momentum that could just like django template partials maybe make its way

into core um it seems like the perfect kind of thing to just incrementally improve uh django

anyway so i've been using that that's cool carlton do you have a project yeah i do um it's another

we don't want to experiment and call we want to you know test things out so it's um rafi khan

who um is part of google summer code did django admin key shortcuts um so the keyboard shortcuts

for django admin and it's lovely looking package it's small it's it's lovely to play with it we

but we need people to test it and feedback

and get some community engagement with it

to make sure that it's stable enough

and all the rest of it.

And whether the design decisions are right

and all those other questions.

Just start using it, play with it.

Do you like it?

Are there bugs?

Are there issues?

Are the keyboard shortcuts defined correctly?

Do they clash with other keyboard shortcuts

that you've come across?

Are you able to define your own custom ones

in the way you want to?

These kind of...

you know is it right um i think it's would be a lovely package to merge in um for accessibility

and usability around the admin which is probably one of the main reasons why people use django

and so um i really recommend people trying that out we do need the community feedback on that

that we'll link to um september 4th jacob did you have one you don't have to but yeah yeah no i i

to be managed and i we were we started using django pg trigger great but we had other things that

weren't postgres triggers so i was looking around for something that would act like that that would

be able to calculate forward and reverse migrations from a single source of truth

instead of spraying your 150 line forward migration for deploying this you know hefty

object and then your another 150 line string for the reverse migration and now your migration file

is unreadable you have several versions i was like there's got to be something in the community

for this and there was this old solution in the community that's recently been forked to be tested

against recent django and pythons called django migrate sequel i think the fork is django migrate

sequel do um and you just put in your install apps and it just works and now we've got like

real clean git diff solder um yeah on your 159 postgres functions also as well you've actually

real as if you'd actually written it um yeah i mean yeah create or replace and then the same

timestamp folders and i was like no no we're not we got to use django for this and i was just

that out wow carlton book you have a book yeah i do so i'm um i'm gonna recommend software can

you see that oh i have a i have a copy of that because you yeah no software design by example

um by greg wilson it's a fantastic book i gave this a review in my very first um stack report

which is my newsletter that i do monthly the very first issue was a view of this raving about it

because i just got it and i picked it up again recently so a year or so later and i'm just like

amazed by it again there's just so much depth in it and the more time i spend with it the more i

find and i think it's not so much in the book within like i'm finding more in it if that makes

sense i didn't read it it's that my my my take on the exercise is the example to have gone so

um like the big ideas it talks about complexity and how you manage complexity and then the

appropriate extraction level for the complexity of your app that you're at which is just

just really wonderful but then the examples of like how classes work really and um things like

um row row orientated versus column orientated storage which was one of the kind of

really fundamental like how why do i need to do it this way to get when i'm doing this kind of

activity um questions why do i need to lay my data out in this in this format and you build

you know you build your examples just python dicks and small things but

that gives you a massive insight that you you kind of previously wouldn't have and i think it's a

wonderful book and i really do recommend it it's also freely available online if you don't want to

buy a print copy but i'm a print yeah do you have one jacob are you okay yeah i uh i guess to go

approximately 1950 by simone de bevoir about her travels in the united states called america day

by day i bought it as a lark and a used bookstore for a friend who had um uh was it was um it's

funny it's we were visiting this curated bookstore you know where everything is you know just so and

the person running the store had said like no we don't have travel writing and it's like you know

i wasn't asking for the lonely planet's guide to italy like i was asking for like memoir or

something like that i went back and i found this on the shelf and i was like look they do have

travel writing here it is and so i just got it just kind of as a lark and then i popped it open

this summer and i was couldn't put it down um it feels like it could have been written yesterday

um you know america's going through a red scare at the time and now we're going through a blue scare

um comments about youth you know and and what's possible um and so anyway if you like if you like

good pros on on travel and on america recommend okay wow awesome uh mine is a short one

recommending it we got on a whole kick talking about books and japanese literature and marikami

and uh yeah it's mid-century really well written there's something about like really good japanese

novelists i don't know why exactly japanese translates in such an interesting but still

good way to english i'm sure it has to do with the translator but it's really it's very i don't

know very japanese very sort of like almost english in a way carlton like very like things

not said but implied that should be on the bingo card as well so anyways i'm enjoying it uh

it's more about the writing than i think the topic it's about like a love affair and stuff

so yeah i can't i'm unlike carlton i can't always read software books to fall asleep at night

i think we've covered most of our stuff anything jacob anything we didn't ask you or you want to

as a fellow and now even more exciting being on this podcast so thanks for inviting me in

uh develop your django development partner beyond code learn more about the

their services in the description and thank you for sponsoring

and thank you everyone next time we'll be we've got we've got some really good guests coming up

we won't we won't give it away carlton but we've got some really good guests coming up

just jogged my memory i wasn't thinking about it at all so when you said it i was like i don't know

who's coming up but then yes i've never thought that we have got some good ones okay so uh we're

next time. Bye bye. See you next time.