Transcript: Katie McLaughlin

This week we're joined by Katie McLaughlin, who wears many hats in the Python Django world, director of Python Software Foundation, also director of Django Software Foundation, conference director for PyCon Australia, and currently at Google.

Katie, welcome to the show.

Hi.

So there's so much we want to talk about.

Maybe we'll just start with the classic how did you get into programming?

One of the first things that really got me interested with computers and what they can do in general was I had a Windows 3.1 machine in my garage that my dad had for reasons, and I worked out that you could change the colors of the theming to be glarish colors, which is now known as hot dog theme.

But from there, I um I didn't really get started into any programming stuff until really high school, where one of the end-of-year classes they offered was an intro to HTML programming in Microsoft Word.

But next year I picked up a book on HTML learnt tables and from there just started getting involved in like working out how JavaScript works, taking the ICT class in final years of high school, then finding a way to be able to do a university course as part of high school, and then getting a bachelor of information technology, and then getting a job, and then jumping to so many different languages, and now I am apparently one of

the directors of a foundation or two foundations to do with technology, and okay, this is fine.

I mean, again, because you you wear so many hats in the community.

How how did that come about?

I mean, how does you know, among the other people who are directors and conference organizers, is there a similar um path of how they get there, or is it really vary for everyone?

And from there going, huh, maybe I could speak at one of these things and then, huh, maybe this is interesting, maybe I should give a talk about this, and then oh, I got accepted for a talk in New Zealand.

Okay, this is fine.

Um, and then just being around and volunteering to help, something as simple as um making sure that a speaker is introduced and that I make sure that people ask questions in the form of a question, and then realizing that there's a whole lot of opportunities around helping on boards and councils and making myself somewhat popular because popularity v

oting is absolutely a thing.

Because I mean, that's something I I hope we can talk about um today is just how much work and how many volunteers are involved in any conference, because I I like I was just at PyCon for the first time and that's overwhelming.

But there's an overwhelming amount of yeah, work behind the scenes to make it so smooth that you, you know, more than anyone since of all the hats you wear have have a sense.

So how does like how does that break down as for example as a conference organizer?

What's what's the sort of ratio of um you know attendees to volunteers to like like how do you keep track of it all in your head, right?

Because there's so much going on at any one conference.

Python world and they are funded by the PSF, and the PSF has actual staff members because one of the things with volunteering is that there are some tasks that you really cannot get a volunteer to do, and financial accounting for a non-profit with how many millions going in and out is not something you dump on a volunteer.

Yeah, paying people for that like professional work is absolutely a thing, but it's also a case of many hands make light work, and volunteering for a three-hour session sharing thing and a couple of dozen people doing that just makes the entire thing run so much smoother.

Um, because it would be great if we could give everyone compensation for their time, but we can also give a hearty handshake and a strong felt thank you, which is for some people more than enough.

But right.

Well, and it's and um I mean for me at PyCon, so I I volunteered to uh do conference check-in, and that's also just a great way to meet people for for anyone if it's your first conference or you're new to a conference.

Um you know, you're all it's there to it's there to meet people, so stuffing bags, you know, just there's always gonna be a list of volunteering opportunities, which is as good a chance as anything to uh to meet a lot of people.

So encourage people to do that.

Um well before we get more into conferences, uh you mentioned you've done a lot of different programming languages.

Can you talk about the shift to Python and Django, which is where um you I think you still are, or talk about, you know, in a professional sense, how how are these things used, right?

Like what's the what's the vibe?

Because everyone listening knows Python and Django, but they probably also use different programming languages.

Oh wow.

Um But due to a series of events, I have been um employed as a software developer in a not a junior, not a senior, but as a oh, I can eventually learn the language and then get stuff done.

Um so I was doing PowerShell for a while, and then I was doing Ruby for a while.

Um, but the first conference talk I gave internationally was at Linux Conference Australia.

And Linux Conference Australia is part of Linux Australia, and Linux Australia runs other events, including WordPress um WordCamps, including DrupalCon, and including PyCon Australia, which it was strongly suggested was a very good event that I should attend.

Not having a lick of Python in me, I submitted a talk and it got accepted.

Yeah.

I did submit and publish a package in Python before my first PyCon by reading the docs and working it out, and then some very marvelous people made sure that my code wasn't terrible and ported it to Python 3 for me.

And it I did not know a lot back then, but I still don't know a lot, but I know a little bit more.

And so when when um professionally did you were you able to work with let's say let's say Django, right?

Was there um have you been able to do that for a while?

Or that was I think only at the m at Divio most recently?

Um at Divio I was a site reliability engineer, so I was paid to help keep the Django's running, but I have not been paid to professionally develop in Django, and yet I am a now two year serving board member of the Django Software Foundation.

That's amazing.

Like you don't have to be, well, you don't have to be Carlton to be able to serve on these things to be able to help the community.

Um I mean, every community needs a Carlton or twelve, but not everyone has to be a Carlton.

Do you do you remember um, you know, so learning Django?

This is a question I like to ask people because you know, most people, whether people have come from a previous web development background or not, tends to have a big impact on how they view Django.

I'm curious for you.

I I assume you'd you'd done some other web stuff before you touched Django for the first time.

I have been paid to do that.

Um I've also done the odd Sinatra app or Flask app here and there.

But most of the way that I learned how to use Django, I mean, I still don't know a lot about it, but the bits that I had to learn was the stuff that helped me in a system administration sense.

So literally, there is a Django app that's playing up.

How do I go in and work out what's wrong?

And oh, there's a Django shell.

What is that?

And how do I manipulate this?

And I gave a talk at DjangoCon US 2018 where I described literally going from my Oracle roots and my SQL knowledge into how that maps to the ORM, which is a learning process that is absolutely valid, being able to use generic skills like SQL and being able to apply that to domain-specific information like the Django ORM and knowing that there are

similarities and how that maps to each other as a diving board into the great unknown.

Yeah, we'll link to that that talk because that I I would say that yeah, the ORM, the more time I spent with Django, that's the part that is the most interesting but still a little mysterious.

Um I feel like that in some ways maybe has the greatest depth.

Um I mean even uh Tim Graham, who we had on um as a guest, said that was the one area that he felt he didn't have mastery, though, you know, take that with a grain of salt when Tim says that.

So uh another thing you've um talked about is um deployments, which this is something Carlton and I often talk about.

Um so I'm curious, you know, you've done some talks.

What are your what are your thoughts on Django deployments before I give mine?

I now work at Google where there are many different ways that you can host your Django.

Um I was accepted, but sadly could not attend DjangoCon Europe, where I was going to give a talk entitled, What is Deployment Anyway?

So I haven't actually written that talk yet.

Um but there is absolutely a stepping stone from I've completed the Django Girls tutorial to so where do I host this now?

And going from there, what is the best way to host production level ready and scalable Django?

And what is 12 factor and what is serverless and why do I care?

And because I have a very strong operations background, not just a development background, I am able to talk both languages and be able to help convey the TLDR from the operations world for developers.

Because hopefully, in most situations, it's not just the developer having to do all this stuff.

But what it is, you don't need to know the entire spectrum, you just need to know enough to get your job done.

And hopefully, if if that talk gets accepted anywhere, we'll be able to update these show notes to a link to it once I give it.

So they have to handle all the deployments.

You know, if it's full stack Django, they have to learn JavaScript and all these things.

And they're you know, the reality is most Django engineers are focused on Django, focused on back-end APIs, and have an awareness or can dive in as needed, but do not have all of this in their head at all the levels required.

So it's

You do not have to know everything.

I hope that anyone listening to this doesn't have to be in that position where they are the one-stop shop, that they have co-workers and colleagues who have the specializations to help them with that.

However, it is also very useful to know what those folks are doing.

And so this is DevOps.

That's why I always tell people to like that.

Sometimes I'll I'll be out of my depth on some production thing, and it's like, well, you know, DevOps is a career for a reason.

Like Yeah.

I am but a single person, so obviously I'm a DevOps.

Oh, that's true.

Yes.

Um one thing that I thought it might be interesting to talk about is um what you see as the similarities and differences in these various communities that you're so involved with.

So PyCon Australia, and then maybe even the DSF versus the PSF.

Um, because uh, you know, even for me, I only have a vague sense of what these organizations are and do.

Um but you you know you have that experience.

I'm curious what you know pro uh, not pros and cons, but just what are the sort of similarities and differences um that you've noticed?

And for very successful conferences, it turns into a space that turns into an annual pilgrimage for the community to join together and fly halfway across the world to listen to people to share ideas to learn, and it is absolutely something that a bunch of the um languages that I've used don't really have

Like, I'm not aware of a lot of um such events like PyCons for the Java world.

I'm sure they exist, I'm sure that there are many people who attend like um Oscons or All Things Open, where it's not language specific, but just the sheer scale of how many Pycons there are around the world means that there is nearly always one close to you, which is so important, especially for those who can't travel.

So like there's Rails Conf and stuff, but maybe there's a RubyConf.

Um Python.

There are RubyConfs around the place, but there are there's the pycon.org is a newly relaunched and revamped website from the PSF that lists all the Pycons, including a calendar of all the PyCons.

And just in June alone, there is one, two, three, four, five, six events and at least one Django Girls event in Israel, in London, in the Czech Republic, in the US, in Switzerland, in Colombia, just in June.

Yeah.

One thing I'm still again trying to understand is where does where does the Python world see Django fitting in?

Because um my sense is that Django used to be more of a first first pillar of Python, and now it's more of a utility, which is a good thing.

I mean, certainly the excitement at PyCon was not around Django, it was around data science and async.

I'm curious what your take is on that.

What is you know, because the Python world is so much broader than Django itself.

Django has proven itself to be a very stable, boring platform on which to work on, which is extremely good because having uh putting all your eggs in one basket on a beta product and then having that disappear is not good business practice.

No, it is not.

I I spent a year and a half working with MeteorJS, which um was fantastic and I think is still an active framework, but was changing um more rapidly than our business business needs would have liked at the time.

And Django itself is so mature, and the DSF is in a position where it can fund a person to help be a full-time or part-time or Carlton will correct me.

Yeah, two people.

The DSF has money to be able to put towards helping conferences happen, towards ensuring that Django continues to exist.

And the PSF does that at a larger scale, which is absolutely incredible, but there is still so much more that goes underfunded, unfunded volunteer effort.

The PSF has a grants program, the DSF has a grants program, but those aren't at the scale that would provide a salary or 12 for professional software engineers to be able to fix all the problems in just one little bit of C Python, for example.

Um that's an entire other conversation, and I highly suggest that you get Dr.

Russell Keith McGee on board because he has a lot to say about funding.

And if you can't, or if you have only 45 minutes, I strongly suggest you watch his keynote from PyCon 2019 in Cleveland.

It is a very good talk.

I was fortunate enough to be yeah, there in the room for it.

I was I was fan-girling the entire time.

He's uh sometimes it's hard to make a technical talk interesting, but he never has that problem.

He um we'll link to that one, but he's given a lot of talks, you should check them all out.

Um so actually, and speaking speaking of Russ, um you've also done a number of open source stuff, including Beware, um, which um maybe you could speak a bit about that project and and your involvement.

I happen to be at these development sprints and ended up sitting next to Russell and he was mucking around with this thing, and I'm asking, well, what was that?

Oh, I'm trying to get Python to run in the browser.

I'm like, huh, can I try that?

Well, no, I haven't got it up on GitHub yet.

I'm like, well, could you do that?

And then I can try it.

And turns out his code wasn't exactly installable at that point, um, but I was able to help debug that because anyone knows it's like it works on my machine, is not a good uh unit test.

It is a um one and a half inch, very hefty coin made of nickel, um, which has the BWAR logo on one side and a yak on the other.

And this has been an amazing factor in getting people interested in our project because people like shiny and people will contribute to your project for shiny.

Um, but we don't just have one coin, we have a second one.

If you help two or more people earn their yak shaver coin, you get a yak herder coin because we also like our puns.

And it's the same form factor of coin, it's just that the yak on the other side is on a wonderful green grass on a beautiful clear blue sky day.

Um, I'm sure that we will also link pictures to these particular beauties in the show notes.

Um, but that's part of Russell's attempt to get engagement and also a use case of some of our um uh donations and funding support.

So a bunch of those pressings were done directly by folks like Travis CI and Max CDN and GitHub giving us money to make coins.

But a pressing of coins is not a developer's salary, sadly.

But no, it's it is a way to, yeah, it is something.

I mean that would help, but then of course let's pick you know some benefactor, they would inevitably want it to be custom to them, right?

That seems to be the sort of tension whenever you have um successful companies using open source software is they that they want to use it, but they inevitably want something that may not align with what the broader community directly wants.

And to take it away from BWA for a moment and go to say another large bit of software that is production ready and stable that has helped many businesses make many hundreds of millions of dollars.

Django, for example.

There is many, many hundreds of millions of dollars are being made off the backs of volunteer sweat, and there is nearly not enough, not nearly enough um paying it back to that.

And I totally get why it's it's it's like this is open source software, it's got to be free, right?

But there is a point where you need to start thinking about the sustainability of what you are using to literally make your money.

And if you don't secure what you're using, you are being negligent.

You are not ensuring that the product that you use that you depend your entire platform on continues to run.

If you do not ensure that that continues to exist, then you have a major liability.

So it's not just a case of kicking back a few hundred bucks to a project, it's a case of being an actual sponsor about giving back in a financial significant way that doesn't just serve your interests but the interests of the project in general.

And I understand this is really not the best thing to be saying when I work for a company that is as big as Google, but it's one of those social things where free and open source software is so new and we're still working out what it all means, and we are we are seeing now the retirement of some of our f founding fathers in this space, like with Guido stepping down, and it it it's

incredible and amazing and fantastic, but also brutal at the same time, and it's yeah, we live in interesting times.

And um I mean one thing that I've seen it's interesting the different funding approaches out there.

I mean, for example, at PyCon, um, there's a company, Tide Lift, actually based in Boston where I am, that is trying to do like a subscription for open source stuff.

Um there's even one of my one of my friends, he works in the PHP world, and Laravel's interesting because that's a web framework that is for profit, largely run by one person, and that person, for example, sells a $99, I believe, SaaS off-the-shelf, you know, like third-party app.

And that alone contributes greatly to the community.

So it's I hope that it's a business problem, not not like a cultural or social one, to find ways to make open source projects like Django sustainable.

Um, but it is interesting seeing what people other people out there are doing.

I mean, for example, Django, I think last fall for the first time had a um sponsorship with um JetBrains where for a month there was some donations.

Um yeah, something needs to be done, but it's it's tricky to balance, as you said, the nonprofit and community aspect with finding a way to get the funding that the project needs to uh to go on.

I'd I've just listed a whole bunch of things.

I wonder if any of those those that I mentioned uh to you seem interesting or if you've heard of other approaches um that you know make it maybe a combination of easier and also like, oh, we have to do that um for um the companies that are using and relying on this software, specifically Django.

There are multiple people who can survive as an income off WordPress plugins, paid WordPress plugins.

Umatic is an entire company based off this, WordPress have their own hosting models.

There is a lot of money going in and out of WordPress.

And even though it might seem as someone who has always been able to just download packages for free, maybe if you have to have a subscription which makes sure that your particular package gets patched, maybe that's not a bad thing.

There is so many people trying to spin up these sort of things, but it's always a case of we're software developers, we're not economists.

There is

Yeah.

And I mean for me, because I I came from the business world before I got into programming, so the idea of Stack Overflow, that certainly does not exist in the business world.

I mean, there's just you there's no way that someone would freely help you with something.

So that's one of the things I I love about engineering, and yeah, you don't want that to go away.

I mean, with Python 2 now being deprecated in 2020, I believe it's Red Hat is gonna offer some sort of support for it, but you have to pay for it.

The idea being that you're a large company.

Do you know more specifics on that?

What can be recorded?

The reality is that enterprises cannot meet that deadline.

So there are a number of providers who, for a fee, will continue to make sure that Python 2 at least remains with security patches.

However, that is a big risk because when the entire community is no longer looking at Python 2, no one's going to be trying to find security vulnerabilities anymore.

No one is going to be pushing it as hard as they once did.

So we're not going to know if things are still broken.

It would be a very, very good idea that if you have any Python 2 code running, that you do migrate it to Python 3, preferably before January 2020.

Yep.

Python 3 is marvelous.

Python 3 supports emoji.

Please migrate to Python 3.

And you know, something about Django traditionally is it seems it's it wasn't used as much for agencies um in the way that certainly WordPress is.

But with Wagtail, that's I think Jang uh Wagtail is the second most popular um third-party project now.

Um that really is becoming a use case, and so that might be a way that it's easier for Django to look at, I guess, some of these WordPress um monetization techniques, though as we as you've said, um there is a danger there and losing the magic of open source.

You can spin it up yourself if you have the resources, but if you don't, here, pay us and we'll do it for you.

I mean, an example beyond WordPress that Django could look to for that.

Because I think with Python, there's starting to be some of this.

There's uh I mean Python Anywhere has this, um, Replit has this.

Are there other good examples out there of in the Python world of people doing hosted options?

Divio.

I know.

I I don't have personal experience with Divio, which is why it doesn't come top of mind to me.

But yeah, um DiviO.

You can also use Django CMS and host it all yourself, or you could pay them and they will host it for you, and then you don't have to worry about having to do all that.

Yeah.

How do how does that work?

Aldrin.

Um sorry.

Aldrin Django is their code name for their opinionated Django, which works in their setup, but they made Django CMS and they will host Django CMS for you with a uh you can go on to a URL that I can't remember right now, and it will spin up a copy of Django CMS for you, give you 15 minutes on it, and if you like what you see, you can click continue and pay for that particular instance to

continue for you.

Oh, interesting.

Yeah.

And and so that seems very much geared towards agencies, but is that incorrect?

They're also hard um, or maybe startups as well, larger companies, because that's the at least my perception is at a certain point people need to go custom in the past.

But um certainly, well, you're you're at Google now.

I mean Google that's not the case if you use Google.

It also scales to enterprise because Divio have expert DevOps people behind the scenes who can make sure that your stuff will scale.

Um, and it can also just be for the developer because it comes pre-configured with Django CMS or Wagtail or Oscar um out of the box.

And so if you don't want to have to worry about the hosting, you can pay someone to host it for you.

However, not everyone has that as a valid business model.

Well, personally speaking, I don't like doing DevOps, so I'm happy to pay a bit of money and not have to deal with all this myself.

Um, but it is uh not everyone feels that way.

Are there uh things we haven't talked about that you want to cover?

Uh tickets will be on sale at the beginning of June.

And if you happen to like us the thought of coming to Australia, perhaps convince your boss that you should learn more about Django and Python by going to a conference in Australia.

The weather's nice and there are no drop bears in the CBD.

Is it is the is that correct?

Your ticket will cover access to both events and also any other talk at the conference.

Awesome.

Well, we will link to that, of course.

Um thank you so much for coming on the show and sharing your experiences in the Python and Django worlds.

Thank you for having me.

Bye.