What's Jammin'? Jamstack FUD
This is a joy to witness (◍•ᴗ•◍)♡ ✧*.
Welcome to Remotely Interesting brought to you by Netlify.
People who were remotely interesting:JAMSTACK FUD (Fear, Uncertainty, & Doubt)
Recently WordPress's Matt Mullenweg's came out as not a fan of the Jamstack and we've heard lots of feedback from the community about things people may not understand about the Jamstack. It got us to thinking, "What can we help clarify about the Jamstack to help everyone make informed decisions about whether the Jamstack architecture is right for their teams?" To show you just how dedicated we are to getting you the answers we have a very special guest on today, Matt Biilmann, who had a huge part in creating the term and the concept. Let's dig in!
why are we excited about the Jamstack?
As always, we hope you find it remotely interesting.
TRANSCRIPT:
Phil Hawksworth: [00:00:00] Previously on Remotely interesting.
Tara Z. Manicsic: [00:00:04] When don't you just make a Tika Tok?
Cassidy Williams: [00:00:07] Hello, and welcome to Remotely interesting.
Phil Hawksworth: [00:00:10] This is remotely interesting.
Divya Tagtachian: [00:00:11] Well, that seems a little
presumptuous.
Sarah Drasner: [00:00:14] No, thats the name of the show.
Tara Z. Manicsic: [00:00:23] Hi, everyone. Thanks for joining us today on an episode of remotely interesting. We're calling this session fondly, what's jamming, and we're going to focus on jam stack FID, which is fear, uncertainty, and doubt. this is in regards to, there has been a little news lately out in the. JAMstack community about, the presses, Matt mullenweg's came out as, maybe not so much of a fan of the JAMstack.
And we've heard lots of feedback from the community about things people may not understand about the jam stack and it got us to thinking. What can we help clarify about the jam sec to help everyone make informed decisions about whether the jam architecture is right for their teams and to show you just how dedicated we are to getting the answers.
We have a very special guest today. Matt Billman. Who had a huge part in creating the term and concept of the JAMstack will answer all of our questions no matter what,
Sarah Drasner: [00:01:25] no pressure
Phil Hawksworth: [00:01:28] you regretting this already matched after hearing, hearing what you're in for that we could ask you.
Matt Biilmann: [00:01:32] Absolutely. It's still assume call, right? I can certainly suffer unexplained connection error.
Phil Hawksworth: [00:01:42] It happens.
Tara Z. Manicsic: [00:01:43] Yeah, it's really early in the morning for you too.
Matt Biilmann: [00:01:45] I
mean, it's not that bad. 10:00 AM is like I'm on my fourth cup of coffee or something by now. So it should be okay.
Cassidy Williams: [00:01:55] Oh, he's ready to go.
Phil Hawksworth: [00:01:56] Oh, hummingbird after four coffees,
Tara Z. Manicsic: [00:01:59] the rest of the podcast is just Matt talking nonstop,
Sarah Drasner: [00:02:06] just throwing out ideas.
Tara Z. Manicsic: [00:02:09] I'm going
Cassidy Williams: [00:02:09] to pivot the company by the end of this.
Tara Z. Manicsic: [00:02:13] There'll be a whole new architecture by the end of this. One of the first topics that I thought would be great for us to talk about, is, I think all of us on this call are very excited about, the way that we can build the web with the jam stack.
And, we have all built the web in many different ways. And not just the JAMstack. And so I'm curious with all of you, how you see this new presence of this architectural approach of the jam stack and like what's changed and why you may be excited about it. Doo doo doo doo doo doo doo,
Phil Hawksworth: [00:02:53] I'll go first.
I'll go first because I was interested. Tara that you said this, this new approach, and it's definitely feels like it's a, like a new term and we've got all of these new tools and toys and services that are there, but it feels like to me that this is built on top of an age, old approach as well. I mean, it kind of hearkens back to the fact that, you know, we can create things that are.
Ready to serve, you know, they're, pre-generated, they're on a CDN. They're ready to go. and, and it feels like that's, that's the way maybe not the CDN layer, but, but yes, it's been established for a long time that we could build things out as like pre-generated assets that we could serve. Very simply, you know, later CDNs came along, but ultimately we had these.
Assets that we can get out into the world. So at that part, it feels like it's been around for a really long time. And it's, to me, it's the new stuff. And the stuff that excites me is the, the services, the glue layer of which notify as one. And the services that, that, you know, we can hand off the more complicated tasks to other things and use them through API.
So I don't know if I, I feel like this is a whole new thing. It's more like. Oh, really nice maturing and evolution of the things I enjoyed from the beginning. That's, that's always how I kind of think about it, even though, you know, it definitely feels like it's a, a new term and a new, any new kind of era, I suppose, for the web.
Cassidy Williams: [00:04:16] It's kind of how I feel about it too. Because when I first was hearing the term, I was just like, Oh, this sounds like this whole new thing I have to learn. Great. But then once I started looking into it, I was like, Wait, this is kind of everything I already know. I don't have to learn anything, but there's some nice bonuses that come with it.
And suddenly there's just things that I don't have to do that make my life easier as a developer, while still letting me build off of the knowledge I already had just of web development in general. And that's, that's probably the most exciting thing for me. Like. It's kind of feels like the you're teaching an old dog, new tricks kind of thing, but you're, you're able to do a lot of.
Classic web development practices. I don't want to say old, but, but ones that have been established for a long time,
Tara Z. Manicsic: [00:05:01] mature
Cassidy Williams: [00:05:02] look. Sure. Yes, but enhanced with new technologies and, and abilities to do all sorts of things that you can just couldn't do before.
Tara Z. Manicsic: [00:05:11] It's funny that you say it that way. Cause that's exactly.
I just, I I'm doing a talk next week. Titled building a better web with the technology you already know. And I think that's like a really great highlight of it is like you're not learning all new
Matt Biilmann: [00:05:26] things. It's also very much been, been the journey for me. Right? Like it's, it was fun to see Matt Mullenweg bring up movable type.
Right. Because like way back then there was sort of like there their competition between WordPress and movable type where movable type had a mode of essentially working as a, as a static site generate and just outputting the generated site. Right. But back then, There was a lot of reasons that that approach just didn't work out that well.
Right. Like, and then, and, and I was kind of familiar with them because I, I, I worked at C Joe for this company where we were building websites that are very light scale, just for small to medium businesses. Right. So a hundred websites every week. So when we had different diversions of the internal platform that we built to deliver all of that with, and the really first version was actually also in the same way where every site would get.
Sort of, it was not completed stag, but it would like output a prebuilt sort of PHP site that would get FTP to a server and then be ready. Right. And any change the client made would regenerate and build it. Right. And we then moved away from that approach and built like a pure content API, but with dynamic front end, that, that, that would pull everything.
And I didn't build like cloud hosted version as its own startup multitenant, all dynamic and so on. Right. So I sort of saw all of those steps, but also saw when I was running these more complex dynamic systems. Right? Like there was still something really beautiful about like the operational simplicity of like once you've rebuilt the site and it's out there, right?
Like you, you, you really did it a different piece of mind than any kind of like really complex dynamic Ellie operating system with a lot of moving parts. And so Andre. So I was spending a lot of time of thinking, like, why was it we moved away from that, right? Like, why did move up with type field? Why did like our early model not really feel like it was working and so on.
Right. And. And there was all these pieces that were sort of taking a logical and platform oriented around it. Right. Like if DP as a deployment protocol, for example, it's really horrible, right? Like there's no concept of atomicity of like one operation. Right. And you you're basically just missing. With a program with, with files in a folder that, that people are also reading and so on.
Right? Like it's really not a nice way to manage like a deployment destination, especially if you start having thousands of them. Right. So, so that was one thing that, that, that really had to be fixed. And then of course, As time has gone on, right? Like back with movable type build times would very quickly get very long for like any reasonable sized site.
Right? Like, so that's just a pod where sort of evolved evolution of, of computation and disk speeds and memory speeds and everything it's meant that. That now we can expect like much shorter turnaround on, on a built thing. Then we could like, obviously 10, 15 years ago. Right. And then the other part was just that, like, I can back in the times of movable types of these technologies, like having a CDN, that was really not something normal.
Right? Like there was something like maybe CNN had a CDN or something like that. Right. But you, as a normal person would not have a CDN. Right. Like, so. Seeing that, that by now was something that could be democratized and just be available to pretty much anybody also really shifts like the actual, like performance results you get from pre-building like the site tremendously, right?
Because suddenly you get this global distribution and you, and people will load this side out of it, out of a destination. Right? Like, and then the other part I've really lived through was that. Originally, like when, when we build all of these sites, any site that involves checkout and payment, right. We had to really do this whole, the Riverside integration with a payment provider and like check out flow.
And the only way to do it was really draft. Like I didn't have dynamic website. Right. So it was sort of forced to do that. Right. And the same when for so many other things like comments or anything that needed to be interactive, right. And then over time, I just started seeing more and more shifting away from, from our own backend, even when we were still running our own back end.
Right. Like where quickly we would start, like when Stripe came along, right. You would start feeling really silly if you were still doing your whole payment flow on your server with some like old fashioned bank system or something. Right. Like. So even if you still ran those servers, you would say that pulling out like those kinds of things and just do them Glenside side and talk to strive in the browser.
Right. And the same happened for, for a whole lot of different things. Like feedback is so form submission, pressing all those stuff. Right. So suddenly I sort of also saw that now we were running all this really complex dynamic infrastru. Yeah. But so much of the actual interactions work completely independent from all of that.
Right. And now a lot of this infrastructure is essentially just generating. HTML on the fly and, and being bedded. Right. So, so I think just over time, it's seeing like the technology around how we build stuff, it's evolved to a point where suddenly it said, Hey, now, now if we went back and had more of this approach of thing, how much can we actually pre-build out in front a.
Let's let's do that. Let's distribute it right directly on a CDN that that's fine deployment methods that are way more efficient than, than, than. And if TPP is that can really guarantee that each deploy we make is like a snapshot of, of, of state and atomic and instant and so on. Then we really go from, from this being something.
That, that feels old fashioned to being something that, that feels much nicer to actually work with as a developer and much easier to reason about, and that like just operationally, it's a completely different dimension from running these really big monolithic dynamic systems.
Sarah Drasner: [00:11:23] Well, I just wanted to clarify too, for listeners who are, are kind of listening in, and aren't sure about some things and kind of breaking apart some of the FID that a WordPress versus JAMstack does not have to be right.
You can absolutely use WordPress in a headless manner. I have a couple of examples online of like a next site that uses WordPress in a headless manner in the writeup on smashing about how we did a similar type of thing. of, integration and migration for smashing magazine and the O'Reilly book that, Tara has made an audio book of and Matt and Phil, wrote, also has a lot of example, technical examples of that as well.
One thing that I was curious about that Matt Mullenweg brought up that I'd love to get your points of view on as well was. So I used to work in agency as a WordPress dev back in the day, and he did bring up this. point of view that using microsystem microservices is brittle because you can, you have these different pieces of the system and what if something falls apart and things like that.
And I actually have found the opposite to be true. And what was interesting to me about that? Having been a WordPress dev was that I found that often about the plugin ecosystem in WordPress. And I'm not, I'm not, You know, saying that it's bad because no technology is fully bad or good. Right. There's trade offs for each one.
But I did find that I had to do a lot of research for the plugins that we implemented, because it did start to get into this area where something would. Drop off something would work the way that I thought it was something wouldn't work the way that I thought it was. So I did a lot of investigation for each one of those plugins.
And there's some symmetry here with the microservices and API commerce, right? Like you, when you are working with microservices, you could. Build microservices in a poor way, but you could also build my microservices in a way where you personally have a hand in it where you're doing good research, where you're making sure that what you're investing in is something that everybody in the community is using.
Stripe is a great example of a service that I really trust and depend on. Right? Like I. Stripe is not an, you know, an API where I'm mashing it up with other API. So I think like that the concept was a little bit flawed. Yeah. And that I wouldn't expect that to be a brittle piece of that experience. I've never found it to be.
so I wonder if that is also a shift. Yeah. The ecosystem that maybe has not been tried yet yet, or some piece of it was tried that wasn't. Ideal, but it doesn't express everything. I was curious to hear your thoughts.
Matt Biilmann: [00:14:00] I think that's a piece of it that that's still a little bit ahead of us for the gems tech, right?
Like, and that's the part where when you have like a very opinionated monolithic system like WordPress, right? Like you can, as a developer, go in and say, just enable this plugin. And now, and now you're done with the initial work, right? Like now you have some sort of new functionality. And since the whole system is very monolithic and very opinionated, right?
Like it gets edited at the right place that it does what it's supposed to do. And you're sort of like, you, you get like a very quick, first win of like, I just click this button and now I have a new functionality for my users. It's in the admin panel, it's in the front end website, everything is tied together.
Right? Like that sort of. The inherent strength of the very opinion in monolithic approach. Right? Like, but I think to your point around the brittleness, I think that actually we've also seen come with a lot of brittleness, right. Because it does mean in that ecosystem that every plugin you install have like full access to through your main database, with everything from users and passwords to him.
Two to content to what goes on the website, two styles, two snippets that gets in the injected, right? Andy, and you get this world where every time there's an update to WordPress, you want to apply that update because it could be an important security update. And if you get too far behind, you can never apply the security updates.
But now that update might. Change a little bit about what WordPress does in general, and that might break one of your plugins. So now we have to figure out, does that plugin also have an update made available and does that update maybe require some manual intervention to actually apply in Strawn? Right.
And suddenly people with a lot of plugins and have word prison still gets, gets get stuck in a, in a world where things can actually feel very brittle, right? Like, and where the very tight coupling feels a lot worse where. In this sort of microservice way, if an microservice breaks or something, it's typically like a very isolated piece of your site.
It's not something that will let a hacker take over your full database or even get X's through your server infrastructure, stuff like that. Right. It'll I mean that, there's like a localized component. That's not doing what you want. Right. So I think in terms of brittleness, I don't really. I don't really see the big advantage from, from there, from the plugin ecosystem.
Right. I think it's more about that initial, like opinionated set up piece. Right. And I think of course. A full sort of very monolithic system will always have like an advantage there. Right. But I do think that we're seeing a lot of, a lot of progress in the gem stack ecosystem and thinking through like, how do different services connect to each other and work together, right?
Like we're seeing on the framework side of things, more thoughts around like, how do we get data from, from, from these providers and make it easily accessible. We are seeing on. Tools like Tina's CMS or stack bit or Prismic slide machine that, that, that shows like how can you even have templates and themes and then have like normal content editors just plug in new components, drag things around and have things work in, of course we've done a lot of work with.
Building things like our identity service, where you can quickly add like an login widget and, and manage that. And there's tools like zero and so on for those kinds of functionalities. And I think, yes, the whole ecosystem progresses. We will see it. We'll see more and more tooling that makes it really easy to also get that initial sort of upfront experience very quickly putting together pieces.
And then I think. Overtime. Like if you add diligent, like, like Sarah talked about and look at like, what services are we using? Like evaluate the different providers and so on. You can actually get to a place where. Where picking the right ones gives you partners that are really taking care. Yeah. Of all the upgrades, all the maintenance, everything for their part of your stack.
And, and you can actually get to a world where things fields a lot less readily there. And then they can actually often feel like in, in, in this world of big monolithic systems where everything constantly, they need updating that, that works across like horizontally across all of your plugins and your main system.
Tara Z. Manicsic: [00:18:33] I think that was a really big thing for me about learning with the jam soccer texture, because you are as a team, right? Like either, if you're doing like a microservice architecture or doing more, more monolithic, you are always having those conversations anyway about what technology you're using for what, but being able to delegate those really hard things.
And it's the huge stark difference for me with security, because I've made a lot of. WordPress sites where we had to be up, you know, make sure we were updated on those plugins to make sure we were like countering and being either reactive or proactive about different security vulnerabilities. Whereas with the jam stack, it was just always amazing to me that I could pass off security in general, like MetLife identity and zero, because security is.
So hard because people, they, people have their job is to hack into your site. Right. And so they're constantly like doing their research and probably there's probably like a, how to hack someone's site weekly newsletter that they're all following. So like, to be able to keep up with that in itself was, you know, especially if, you know, you have limited developer resources, just being able to have a team who said.
Specialty is like that third party service of keeping your website safe. Just that kind of delegating to me is so huge in the JAMstack.
Sarah Drasner: [00:19:57] Yeah. I mean, even if you enable an analytics on an LFI site or, or, you know, something like that, what you might see is that there's a lot of people pinging that WP login or WP admin file.
And I think that that's true probably of every site on the planet, right? Like, everybody's just like, Oh, is there any. Security vectors. I can exploit right here. And that was also, you know, a big reason why we saw a shift in the landscape was thinking through how to be more secure all over the place and all around the world.
Phil Hawksworth: [00:20:28] And, and I liked the fact that, with this model of decoupling these various services and consuming them from APIs rather than collecting, you know, a group of plugins that you've. You know, you've, you've found through, I mean, they're very discoverable these plugins in the WordPress ecosystem and that's, that's, that's a huge advantage.
However, they're all living in the same pot. And so just because you're using just, you might not be using something to do identity, for instance, however, you still might be introducing. Compromise is that could, could touch that part of your infrastructure because they are all living in that, in that runtime.
The thing that I really like about this, this, this, the jumps like approach is that you're not putting all of your eggs into this one basket together. You can sit in the same way that. Tara was saying that you can inherit this expertise. You also put distance there as well. I really like the fact that you bring that capability into your team, so you don't have to become compliant for all, for all of the different things to do with security, but also you put it, put it.
Away from where your site is. So there's a great distance and there's a great decoupling there. So you don't have to take on that burden, you know, as you slowly gather and collect these, these various APIs and services, each one has its own areas of concern. And each one is brings with it, its team of experts and that's their core business.
That's the thing that I love, you know, I. I don't want to have to roll my own. One of these when there's someone whose entire business depends on them doing that properly. And they have to have all the compliance and regulation, all of those things that I've never, I never get to want to get into. So very
Tara Z. Manicsic: [00:22:01] undeveloped to say, you don't want to make your own
Phil Hawksworth: [00:22:05] you're right.
I mean, strictly speaking from
Tara Z. Manicsic: [00:22:08] the ground up every time.
Phil Hawksworth: [00:22:09] Yes. If I can't write a blog post blog post that says do X, Y, and Z from scratch and do that for every service. then I'm failing. I think I could get all of the hits that I'm missing.
Sarah Drasner: [00:22:21] Phil not invented here. Hawksworth
Matt Biilmann: [00:22:26] another area where that distance you, you mentioned becomes really interesting is this distance between the built and, and the, and the runtime, right.
Where especially again, when, when we talk plugins and so on, right? Like whenever you go to a WordPress site, right? Like, WordPress has to run. Like it will have like this, the loop, right. That sort of loops over every piece of content and so on and is from the database and so on. Right. And obviously like plug into you add or remove can have a lot of impact in what actually takes place in, in, in that loop and in that run and so on.
Right. Like, so, so you also easily with any kind of monolithic system like that, get these effects where a seemingly like simple thing you add is. Without your knowing, it might be exponentially increasing the amount of database queries you do for every single. Load off of your site. Right. And you might not notice it initially because it might just mean that under normal traffic, you get a little bit of flow down.
Right. But then you get a peak of traffic and suddenly there's like an exponential amount of increase in requests. And each request has an exponential amount of increase in the database queries that does, and suddenly your whole system just blows up. Right. And I've finding like multitenant dynamic systems.
So I've, I've really seen that somewhere, right? Like where. Someday. Like when, when I ran my, my, my web pup, like my CMS product way before Natalie Feinstone right? Like we had an at massive like increase of database queries. They were slowing the whole system to a crawl. And it was just like this one developer that had built some extension for like, Persisting like analytics data for, for users that came in that turned out to like in a worst case scenario, do 63,000 database queries per requests.
And suddenly the face that extension, it was on God, like an advertising campaign from United healthcare or something like that put, edit. Right? Like, and, and every single page load was trying to do a 63,000 database requests and everything just died. Right. But. But those are the kinds of things where if you add some extension like that to interact with data, and that happens during build time.
Sure your build we'll, we'll get a little slower, right. And over time that might be annoying and you'll want to go back in and understand it and try to make it the build faster again. Right. But there's a really big difference from your built getting two seconds slower to every single page load, getting two seconds.
Right. And to the compounding effect that this has when you suddenly get a spike of requests, right? Like, so that's another sort of security or our operations related vector. We're keeping this barrier between builds and query time during build and runtime and query time during runtime, it can be really helpful.
Sarah Drasner: [00:25:19] Yeah. And the ability to roll back. I mean, I can't tell you when I was in agency, how many times? I wish I had a feature like that when we, I mean, we used to have these clients who were really, really, you know, get huge amounts of traffic and we would, of course, during those spikes, that's when you least want things to go down.
Right. And that, that would. It would be a gamble sometimes to be honest, with some of the, with some of the infrastructure that we were using with some of the plugins that were using that, whether or not things would actually stay up with millions of people, hammering it. And you know, at times when you're hot patching something or something like that, I wish that we had the ability to just roll back to another state of things.
For sure. I probably shouldn't go into any more detail than that, but yeah.
Tara Z. Manicsic: [00:26:06] Spill the spill, the tea, everything. We've
Sarah Drasner: [00:26:09] all been okay.
Phil Hawksworth: [00:26:11] That's cool. Let's he? Is that,
Tara Z. Manicsic: [00:26:13] is that how we say that? But the kids are saying or do we drink it?
Phil Hawksworth: [00:26:21] The beans,
Cassidy Williams: [00:26:22] but that's, that's when you're like gossiping. So it would be accurate. You were right, Tara. I'm still here.
Tara Z. Manicsic: [00:26:26] Yeah? Oh, good. Okay.
Sarah Drasner: [00:26:33] We actually share something. That's like not the full story of what the client would like. So I worked for this agency that had some clients who were romance novelist, and I didn't know anything about romance novels before I worked. For that, I guess I have like books on my shelf.
People are like, what is this for a while? Like, it's, there's a. There they are the number one most, you know, like highest profiting industry on for, in terms of books. Like they're, they're a billion dollar industry that people don't really know about that much. And they're very, very popular. I just had no idea.
And so one of my clients put out a contest once and we got. 30 million for the company. Doesn't it involve this custom post type and format. Yeah. Calendar thing that I made and a few other people made, and we were just like, we didn't, we actually, we knew that there was going to be traffic. We didn't know it was gonna be that much trap.
but this is an author who. Has books and every Walgreens and every single super market. And so was quite popular. And wow. That was a learning day.
Tara Z. Manicsic: [00:27:47] It was my,
Sarah Drasner: [00:27:48] it was definitely a day where we were like sitting by the monitoring as much as you could in those days. I mean, I'm a neck beard, so it was a while ago.
we didn't have great observability then either. Right. I think something else that's improved quite a bit. in times of JAMstack is the ability to see what's working. What's not working what's up, what's down. And how much were, how fast that latency is around the world as well.
Tara Z. Manicsic: [00:28:13] That actually brings me to, kind of the final topic that I would like us.
To chat about a few of us have actually talked about, sorry, things, like, like spilled times how we saw that there is, there's a place where bill times can get slowed down. But now as we evolve with the JAMstack architecture, we have. Things like finally of build plugins, where I know that Phil and Jason have worked on things to help you cash things.
So your build time gets quicker. and there are more people doing more things to help, really speed up your build process time, because we've seen that that can be a sticking point. and because the stamp stack architecture is, you know, a new evolution of how we build websites. We're seeing this ecosystem grow, to solve these.
Issues. And so I like to call, I wanted to call this section the rise of Skywalker. but I mean, this rise of JAMstack, but, okay. Here's the thing though. We do a lot of things in the cloud, right then technically clouds are in the sky, so we're kind of walking the sky. So I'm still, right.
Phil Hawksworth: [00:29:20] This is a joy, this is a joy to witness.
You're such a craftsman. It's amazing to see you work your magic.
Tara Z. Manicsic: [00:29:28] I got a call five. You so, so one of the things, we also talked about where it was, doing things like using the CDN, which Matt said earlier, it's like big companies, like very appropriately titled CNN would have a CDN. but now we're seeing more and more that. Not only are, is it more available to us so that we get that like distributed globally and the redundancy, if, if anything goes down, you can be on multi-cloud providers.
but now the evolution of CDNs also gives us things like, like we have edge hands, which will give you just yet another way to kind of render things on your site, and even speed up how you talk to, different service providers. but. With all of you, what are things that you think about when you think about the rise of the JAMstack and what the ecosystem is doing to help?
basically give us more dynamic and more reliable and faster web answer me now.
Matt Biilmann: [00:30:30] And the beauty is that the ecosystem is executing, executing across all these different layers, the stack, right? Like there's obviously like provide us like, like us and others in the space that are looking at the whole sort of workflow and infrastructure is based.
Right. And how we can, how it can continues to democratize things that used to be really hard to do. Right. Like again, like if you go back 10 years ago, like, Looking for teams that had like a CIC D. Pipeline that could do like individual builds for pull request and branches. And that would like roll every build out for like an atomic deploy on, on a production, like on a replica and the production infrastructure, it with CDN based hosting and instant cash and validation and so on.
Right? Like do it have to look really long for teams that have built it. And you would know that that would be like, Really big teams with very specialized, like DevOps groups that had really invested a lot of work into this pipeline. Right. And now any web developer can essentially get that in a few seconds, sun on nearly fell right in.
And we keep looking to that kind of like. For example, you mentioned itch handlers, right? Like I've, I've talked to really big teams that have again have like big DevOps groups and like network teams. And so on that build really like complex things at the edge layers through tools like or Fastly or sometimes CloudFlare workers and so on.
Right. And. And we've been looking a lot of like, what, what are they doing and what are the kinds of problems they're solving and so on. Right. But it's still like, typically, like the real world use cases are limited to two pretty advanced organizations with like dedicated teams just to build that layer.
Right. So that's one thing I'm really excited about from, from our, the point of view of taking, taking this and then. Seeing, how can we make this, something that, that any normal like web developer can, can use. And that's like approachable and testable and it fits into the whole workflow and where you can write, like custom-made code that you can run with your friend and locally that you can open a pull request with and you get like the whole production version of it running.
You can test it recent about it and so on. Right? Like, So that's one example of like this infrastructure layer and how like, in terms of workflows infrastructure. So all of that providers like us can, can contribute. Yeah. Then there's the whole layer I mentioned earlier, right? Where we have players like stack bid and Tina CMS, and it prints make slices and many others that are really thinking through okay.
For the, for the people that are moving the content editing side and do. And, and the like marketing side space, like how, how can we, how can we build tooling that gives them the same sense of like live previews and, and, and, and direct interactions with the site that, that either want to list like WordPress or even site builders, like Squarespace, like can, can give people, right?
Like, how can we take like an architecture that's still nicely decoupled and still great to work with for developers and bring some of. Some of that to, to, to marketeers and induces and contented to strike, then there's a whole. A whole layer of, of developed men and research, I think around like four developers as we start building with all these separate data sources, right?
Like that's another sort of change from like the world where we had like my server and my database. Right. And there was sort of the two things we worked with. Right. And now. We've gone serverless, right? Like we, we no longer really work with that abstraction of like my server, but we also seeing that shift away from like my database to just all these different data sources where you might have your users in one place and your orders in another place and your content in a third piece and some custom data in some place.
Right. So now we also also seeing like a whole. Whole world of, of companies in the ecosystem tackling, how do we actually work with the developer across all these different data sources? Right, right. Apollo is, is doing really interesting things around Federation. It has Torah in launched there, the idea of Federation, right, where you sort of both have your connection to the database, but now you can also connect to, to other things like Contentful in store, and then just query it through one graph.
QL API, one graph is like, have also built like there. One graphic for us, all these different iPads and services where you get like one way to query it in companies like Christmas. Yeah. Thinking about like, how, how do we, how do we build like, sort of, if I'm more universal query language for, for different data type basis and so on.
Right. So I just see that as sort of a whole area. Where, where the ecosystem is making improvements and targeting, like how do we make it easier for developers to essentially like glue things together? And I think that that's the fascinating thing to see now that there's just like this explosion in, in the ecosystem targeting all these different areas of the stack and really seeing like, how can we mature it?
How can we make life easier? Both for developers, for content editors. Yes. If our operations people and them.
Phil Hawksworth: [00:35:52] It's funny. Cause there's, there's so many exciting things within that and you could harvest any one of those. And that would be like a really valid answer to that question, but I'm also excited by something which is really mundane by comparison.
And that isn't, I mean, or. Of course, I'm excited by the things that are new and arriving, but I'm also excited about the things that have been here for awhile and independent developers and maybe small agencies have been discovering by embracing these tools. You know, they've now got access to, you know, Instant cash, invalidation and CDN management without ever having to touch it.
The kind of stuff we talked about earlier on that you'd have to get an account management account manager and a project and try and figure out how to do all of these things. They've got access to these things. Now, as a result of this kind of explosion of the ecosystem, I'm excited about much larger enterprises.
Discovering these in larger numbers, because certainly enterprises are embracing this now, but I think there are so many more places which still don't have access to these things because are, they've invested a lot. In doing this themselves and in a way are kind of behind, you know, they, they have yet to discover how to use these things and, and benefit from it.
So I'm excited each time I see another large enterprise, embrace a lot of these things and get access to the instant rollbacks that Sarah was talking about and not having to worry about. Their success being leading to their downfall when they have, you know, their black Friday sales suddenly becomes a real issue for them, is probably a bit of my agency history kind of showing through that.
That was the thing that gave me the sweaty palms, but I'm excited for more people to discover the stuff that a lot of smaller agencies and independent developers have been getting excited about and benefiting from for a while. So. That's my 2 cents.
Sarah Drasner: [00:37:41] Yeah. I'm really passionate about developer experience as well.
And I think that this is something that wasn't even a con concept back when, right? Like the idea that someone cares about how a developer is feeling about anything in general, but nowadays thinking about how everything is matured and everything's at your fingertips in terms of choices. I mean, we even mentioned that you don't have to move away from WordPress in order to use.
JAMstack premises and make things secure and get rollbacks and things like that. that this piece, I mean, what I love about Nella Phi is this ability to just connect a, get hub repo, and immediately something is all over the world. And I think that that was the satisfaction that we got from pH or from FTP back in the day.
So it was kind of funny to explain where I worked to my parents. Because I think they were used to FDP a site back in the day. And so, yeah, I kind of had to go through the whole
Tara Z. Manicsic: [00:38:36] history of development,
Sarah Drasner: [00:38:38] explain, you know, what we do and, and how we do it. But I, I love the premise that all of these different pieces, the API economy, what Nettleford does, what all of the services that Matt was just talking about.
It, there's a focus. On making sure that all of these pieces of connective tissue work really well for the developer, because that developer activity, isn't just a nice to have. It is literally the thing that makes us not have to worry about the things that we don't need to be focusing on, on so that we can build the experiences that we want to be built and focusing on the things that we do.
Tara Z. Manicsic: [00:39:15] I like imagining that you're at the dinner table and your mom's like, so what do you do, Sarah? And you pull out a presentation and your dad's like kicking your mom.
She does this every
Sarah Drasner: [00:39:28] time too far from the truth. They asked me to stop speaking in acronyms, like CSS, but I was like, my brother's a product product manager at Google. And he was like, no, no, no. We use those terms. Like, like I'm like, okay, I can say things. Without the acronyms, but I think you're not going to, it's not going to provide more clarity.
Phil Hawksworth: [00:39:51] It's no acronym day. No, no. One's will have to use any acronyms at all for that, but you would be
Tara Z. Manicsic: [00:39:59] that sounds exhausting.
Cassidy Williams: [00:40:02] I ended up branching off. What all of you have said. The, the thing that I am very passionate about is making things accessible for new developers and people who are learning and stuff.
And I love how accessible everything is to newbie developers on the jam. Second in this ecosystem, like the fact that. Someone who might've just started learning JavaScript or reactor or any other framework or something can build a fully featured web application with authentication and security and, and images and video and address all, all kinds of different things, from the ground up.
Is so cool to me because that's something that I wish I could have had access to when I was first learning things. And so the fact that that's, that's not even just like, Oh, it'd be a nice to have to be able to do this. It's is here and anybody can build, build all that stuff is that's the coolest thing to me.
And then I love being able to talk to people who aren't sure where to get started. And you can just say, Oh, well learn this, this and this. And then. Go and, and they there's just so many resources available and so many services available that way. Make it easy
Tara Z. Manicsic: [00:41:07] for them, for people who aren't familiar with the term headless.
It's not as morbid as it sounds. this is basically saying that you can have content coming in. And the, it, you just throw that content in and it's something like maybe you, why that content contributors can be familiar with, but it doesn't care how you consume it. So you're getting the data in there and we use, for a nonprofit site that I do for, for, Asian Natty.
we have a lot of people who, contribute the blog posts. And so they just see this user interface where they put the title and they put the content and they upload pictures and they use that in the strappy UI. And then we consume that, and get it all through on the back end. Right. And style it all out with our, With our static site generator.
And, it's just, we never have to explain anything to them. This is a form that they're so used to seeing on every website. So that's the kind of a very, we'll probably have a whole episode on headless, technology because maybe a Halloween episode on headless technology but we are reaching time. So, if anybody wants to one last comment, otherwise, we'll head out for today.
I
Sarah Drasner: [00:42:16] just want to thank Matt for joining us. I mean, it's so cool to work with the founder of jams tech. So it's such a pleasure to hear your thoughts on it, especially because they're, they're so mature, right? Like thinking through what was in the past, but also what's in the future and I really appreciated you.
I think we all really appreciate you
Phil Hawksworth: [00:42:34] joining. Yeah. There's so much stuff we didn't get to cover, so we'll have to. Get you boozy and talk
Sarah Drasner: [00:42:42] after hours,
Tara Z. Manicsic: [00:42:45] we'll have to pause you from advancing the way we make the web in order to talk to us on our podcast.
Matt Biilmann: [00:42:52] Thanks so much for the invite. Well, that's our show.
Cassidy Williams: [00:42:54] Next time we're going to be talking about virtual tech in quarantine conferences. What works or not for both attendees and organizers, it's going to be a blast.
I've been Cassidy. Strawberry
Phil Hawksworth: [00:43:04] jam Williams. I've been filled peach jam Hawksworth
Matt Biilmann: [00:43:08] I've been mad. Rhubarb jam feel men
Tara Z. Manicsic: [00:43:11] I've been Tara big
Sarah Drasner: [00:43:14] and I've been Sarah, Lala Berry jam. Drasner
Tara Z. Manicsic: [00:43:17] that's perfect. Thanks everyone. Hi everyone. Thank you very much for joining us.
Matt Biilmann: [00:43:24] You.
Sarah Drasner: [00:43:30] Everyone was
Phil Hawksworth: [00:43:30] smiling at me, horrifically what's happening.
Something happen.
Tara Z. Manicsic: [00:43:37] Lovely. So it's fine.
People who were remotely interesting:JAMSTACK FUD (Fear, Uncertainty, & Doubt)
Recently WordPress's Matt Mullenweg's came out as not a fan of the Jamstack and we've heard lots of feedback from the community about things people may not understand about the Jamstack. It got us to thinking, "What can we help clarify about the Jamstack to help everyone make informed decisions about whether the Jamstack architecture is right for their teams?" To show you just how dedicated we are to getting you the answers we have a very special guest on today, Matt Biilmann, who had a huge part in creating the term and the concept. Let's dig in!
why are we excited about the Jamstack?
- the age old approach
- building off the technology we already know
- beauty in the operational simplicity
- technology around how we build sites evolved around services and pre-rendering
- from old-fashioned monolithic systems to modern architecture
- Sarah finds the opposite to be true
- Any architecture can be built poorly
- benefits to opinionated platforms
- tight-coupling of a plugin system can also feel brittle
- delegate complicated, dev-heavy tasks like security
- services help you decouple your site to their area of concerns even in your company
- build time vs run time
- having the ability to rollback
- Cassidy explains what's cool to us
- Sarah spills the tea/dirt/beans
- Build plugins to help build times
- the ecosystem across the different layers of the stack
- making advancements that more technical teams are making more accessible to more teams
- layers for different contributors on a team like content, marketing, etc.
- companies focusing on federation and other ways to enhance DX across multiple sources
- enterprises finally have access to these services instead of creating it themselves
- finally people are caring about dx! so devs can focus on building
- the Jamstack & its ecosystem is more accessible to new users
- fyi: what headless is (a quick high-level explanation)
- Matt Mullenweg interview with The Newstack
- Matt Mullenweg interview with The Newstack
- Ohad's post
- Matt Biilmann's post
- How To Create A Headless WordPress Site On The Jamstack
- Modern Web Development on the Jamstack Book
- Audiobook (to be announced soon!)
- Tina CMS
- Stackbit
- Prismic Slices
- Netlify Identity
- Auth0
- Build Plugins
- CDNs
- Netlify Edge Handlers
- Federation:
As always, we hope you find it remotely interesting.
TRANSCRIPT:
Phil Hawksworth: [00:00:00] Previously on Remotely interesting.
Tara Z. Manicsic: [00:00:04] When don't you just make a Tika Tok?
Cassidy Williams: [00:00:07] Hello, and welcome to Remotely interesting.
Phil Hawksworth: [00:00:10] This is remotely interesting.
Divya Tagtachian: [00:00:11] Well, that seems a little
presumptuous.
Sarah Drasner: [00:00:14] No, thats the name of the show.
Tara Z. Manicsic: [00:00:23] Hi, everyone. Thanks for joining us today on an episode of remotely interesting. We're calling this session fondly, what's jamming, and we're going to focus on jam stack FID, which is fear, uncertainty, and doubt. this is in regards to, there has been a little news lately out in the. JAMstack community about, the presses, Matt mullenweg's came out as, maybe not so much of a fan of the JAMstack.
And we've heard lots of feedback from the community about things people may not understand about the jam stack and it got us to thinking. What can we help clarify about the jam sec to help everyone make informed decisions about whether the jam architecture is right for their teams and to show you just how dedicated we are to getting the answers.
We have a very special guest today. Matt Billman. Who had a huge part in creating the term and concept of the JAMstack will answer all of our questions no matter what,
Sarah Drasner: [00:01:25] no pressure
Phil Hawksworth: [00:01:28] you regretting this already matched after hearing, hearing what you're in for that we could ask you.
Matt Biilmann: [00:01:32] Absolutely. It's still assume call, right? I can certainly suffer unexplained connection error.
Phil Hawksworth: [00:01:42] It happens.
Tara Z. Manicsic: [00:01:43] Yeah, it's really early in the morning for you too.
Matt Biilmann: [00:01:45] I
mean, it's not that bad. 10:00 AM is like I'm on my fourth cup of coffee or something by now. So it should be okay.
Cassidy Williams: [00:01:55] Oh, he's ready to go.
Phil Hawksworth: [00:01:56] Oh, hummingbird after four coffees,
Tara Z. Manicsic: [00:01:59] the rest of the podcast is just Matt talking nonstop,
Sarah Drasner: [00:02:06] just throwing out ideas.
Tara Z. Manicsic: [00:02:09] I'm going
Cassidy Williams: [00:02:09] to pivot the company by the end of this.
Tara Z. Manicsic: [00:02:13] There'll be a whole new architecture by the end of this. One of the first topics that I thought would be great for us to talk about, is, I think all of us on this call are very excited about, the way that we can build the web with the jam stack.
And, we have all built the web in many different ways. And not just the JAMstack. And so I'm curious with all of you, how you see this new presence of this architectural approach of the jam stack and like what's changed and why you may be excited about it. Doo doo doo doo doo doo doo,
Phil Hawksworth: [00:02:53] I'll go first.
I'll go first because I was interested. Tara that you said this, this new approach, and it's definitely feels like it's a, like a new term and we've got all of these new tools and toys and services that are there, but it feels like to me that this is built on top of an age, old approach as well. I mean, it kind of hearkens back to the fact that, you know, we can create things that are.
Ready to serve, you know, they're, pre-generated, they're on a CDN. They're ready to go. and, and it feels like that's, that's the way maybe not the CDN layer, but, but yes, it's been established for a long time that we could build things out as like pre-generated assets that we could serve. Very simply, you know, later CDNs came along, but ultimately we had these.
Assets that we can get out into the world. So at that part, it feels like it's been around for a really long time. And it's, to me, it's the new stuff. And the stuff that excites me is the, the services, the glue layer of which notify as one. And the services that, that, you know, we can hand off the more complicated tasks to other things and use them through API.
So I don't know if I, I feel like this is a whole new thing. It's more like. Oh, really nice maturing and evolution of the things I enjoyed from the beginning. That's, that's always how I kind of think about it, even though, you know, it definitely feels like it's a, a new term and a new, any new kind of era, I suppose, for the web.
Cassidy Williams: [00:04:16] It's kind of how I feel about it too. Because when I first was hearing the term, I was just like, Oh, this sounds like this whole new thing I have to learn. Great. But then once I started looking into it, I was like, Wait, this is kind of everything I already know. I don't have to learn anything, but there's some nice bonuses that come with it.
And suddenly there's just things that I don't have to do that make my life easier as a developer, while still letting me build off of the knowledge I already had just of web development in general. And that's, that's probably the most exciting thing for me. Like. It's kind of feels like the you're teaching an old dog, new tricks kind of thing, but you're, you're able to do a lot of.
Classic web development practices. I don't want to say old, but, but ones that have been established for a long time,
Tara Z. Manicsic: [00:05:01] mature
Cassidy Williams: [00:05:02] look. Sure. Yes, but enhanced with new technologies and, and abilities to do all sorts of things that you can just couldn't do before.
Tara Z. Manicsic: [00:05:11] It's funny that you say it that way. Cause that's exactly.
I just, I I'm doing a talk next week. Titled building a better web with the technology you already know. And I think that's like a really great highlight of it is like you're not learning all new
Matt Biilmann: [00:05:26] things. It's also very much been, been the journey for me. Right? Like it's, it was fun to see Matt Mullenweg bring up movable type.
Right. Because like way back then there was sort of like there their competition between WordPress and movable type where movable type had a mode of essentially working as a, as a static site generate and just outputting the generated site. Right. But back then, There was a lot of reasons that that approach just didn't work out that well.
Right. Like, and then, and, and I was kind of familiar with them because I, I, I worked at C Joe for this company where we were building websites that are very light scale, just for small to medium businesses. Right. So a hundred websites every week. So when we had different diversions of the internal platform that we built to deliver all of that with, and the really first version was actually also in the same way where every site would get.
Sort of, it was not completed stag, but it would like output a prebuilt sort of PHP site that would get FTP to a server and then be ready. Right. And any change the client made would regenerate and build it. Right. And we then moved away from that approach and built like a pure content API, but with dynamic front end, that, that, that would pull everything.
And I didn't build like cloud hosted version as its own startup multitenant, all dynamic and so on. Right. So I sort of saw all of those steps, but also saw when I was running these more complex dynamic systems. Right? Like there was still something really beautiful about like the operational simplicity of like once you've rebuilt the site and it's out there, right?
Like you, you, you really did it a different piece of mind than any kind of like really complex dynamic Ellie operating system with a lot of moving parts. And so Andre. So I was spending a lot of time of thinking, like, why was it we moved away from that, right? Like, why did move up with type field? Why did like our early model not really feel like it was working and so on.
Right. And. And there was all these pieces that were sort of taking a logical and platform oriented around it. Right. Like if DP as a deployment protocol, for example, it's really horrible, right? Like there's no concept of atomicity of like one operation. Right. And you you're basically just missing. With a program with, with files in a folder that, that people are also reading and so on.
Right? Like it's really not a nice way to manage like a deployment destination, especially if you start having thousands of them. Right. So, so that was one thing that, that, that really had to be fixed. And then of course, As time has gone on, right? Like back with movable type build times would very quickly get very long for like any reasonable sized site.
Right? Like, so that's just a pod where sort of evolved evolution of, of computation and disk speeds and memory speeds and everything it's meant that. That now we can expect like much shorter turnaround on, on a built thing. Then we could like, obviously 10, 15 years ago. Right. And then the other part was just that, like, I can back in the times of movable types of these technologies, like having a CDN, that was really not something normal.
Right? Like there was something like maybe CNN had a CDN or something like that. Right. But you, as a normal person would not have a CDN. Right. Like, so. Seeing that, that by now was something that could be democratized and just be available to pretty much anybody also really shifts like the actual, like performance results you get from pre-building like the site tremendously, right?
Because suddenly you get this global distribution and you, and people will load this side out of it, out of a destination. Right? Like, and then the other part I've really lived through was that. Originally, like when, when we build all of these sites, any site that involves checkout and payment, right. We had to really do this whole, the Riverside integration with a payment provider and like check out flow.
And the only way to do it was really draft. Like I didn't have dynamic website. Right. So it was sort of forced to do that. Right. And the same when for so many other things like comments or anything that needed to be interactive, right. And then over time, I just started seeing more and more shifting away from, from our own backend, even when we were still running our own back end.
Right. Like where quickly we would start, like when Stripe came along, right. You would start feeling really silly if you were still doing your whole payment flow on your server with some like old fashioned bank system or something. Right. Like. So even if you still ran those servers, you would say that pulling out like those kinds of things and just do them Glenside side and talk to strive in the browser.
Right. And the same happened for, for a whole lot of different things. Like feedback is so form submission, pressing all those stuff. Right. So suddenly I sort of also saw that now we were running all this really complex dynamic infrastru. Yeah. But so much of the actual interactions work completely independent from all of that.
Right. And now a lot of this infrastructure is essentially just generating. HTML on the fly and, and being bedded. Right. So, so I think just over time, it's seeing like the technology around how we build stuff, it's evolved to a point where suddenly it said, Hey, now, now if we went back and had more of this approach of thing, how much can we actually pre-build out in front a.
Let's let's do that. Let's distribute it right directly on a CDN that that's fine deployment methods that are way more efficient than, than, than. And if TPP is that can really guarantee that each deploy we make is like a snapshot of, of, of state and atomic and instant and so on. Then we really go from, from this being something.
That, that feels old fashioned to being something that, that feels much nicer to actually work with as a developer and much easier to reason about, and that like just operationally, it's a completely different dimension from running these really big monolithic dynamic systems.
Sarah Drasner: [00:11:23] Well, I just wanted to clarify too, for listeners who are, are kind of listening in, and aren't sure about some things and kind of breaking apart some of the FID that a WordPress versus JAMstack does not have to be right.
You can absolutely use WordPress in a headless manner. I have a couple of examples online of like a next site that uses WordPress in a headless manner in the writeup on smashing about how we did a similar type of thing. of, integration and migration for smashing magazine and the O'Reilly book that, Tara has made an audio book of and Matt and Phil, wrote, also has a lot of example, technical examples of that as well.
One thing that I was curious about that Matt Mullenweg brought up that I'd love to get your points of view on as well was. So I used to work in agency as a WordPress dev back in the day, and he did bring up this. point of view that using microsystem microservices is brittle because you can, you have these different pieces of the system and what if something falls apart and things like that.
And I actually have found the opposite to be true. And what was interesting to me about that? Having been a WordPress dev was that I found that often about the plugin ecosystem in WordPress. And I'm not, I'm not, You know, saying that it's bad because no technology is fully bad or good. Right. There's trade offs for each one.
But I did find that I had to do a lot of research for the plugins that we implemented, because it did start to get into this area where something would. Drop off something would work the way that I thought it was something wouldn't work the way that I thought it was. So I did a lot of investigation for each one of those plugins.
And there's some symmetry here with the microservices and API commerce, right? Like you, when you are working with microservices, you could. Build microservices in a poor way, but you could also build my microservices in a way where you personally have a hand in it where you're doing good research, where you're making sure that what you're investing in is something that everybody in the community is using.
Stripe is a great example of a service that I really trust and depend on. Right? Like I. Stripe is not an, you know, an API where I'm mashing it up with other API. So I think like that the concept was a little bit flawed. Yeah. And that I wouldn't expect that to be a brittle piece of that experience. I've never found it to be.
so I wonder if that is also a shift. Yeah. The ecosystem that maybe has not been tried yet yet, or some piece of it was tried that wasn't. Ideal, but it doesn't express everything. I was curious to hear your thoughts.
Matt Biilmann: [00:14:00] I think that's a piece of it that that's still a little bit ahead of us for the gems tech, right?
Like, and that's the part where when you have like a very opinionated monolithic system like WordPress, right? Like you can, as a developer, go in and say, just enable this plugin. And now, and now you're done with the initial work, right? Like now you have some sort of new functionality. And since the whole system is very monolithic and very opinionated, right?
Like it gets edited at the right place that it does what it's supposed to do. And you're sort of like, you, you get like a very quick, first win of like, I just click this button and now I have a new functionality for my users. It's in the admin panel, it's in the front end website, everything is tied together.
Right? Like that sort of. The inherent strength of the very opinion in monolithic approach. Right? Like, but I think to your point around the brittleness, I think that actually we've also seen come with a lot of brittleness, right. Because it does mean in that ecosystem that every plugin you install have like full access to through your main database, with everything from users and passwords to him.
Two to content to what goes on the website, two styles, two snippets that gets in the injected, right? Andy, and you get this world where every time there's an update to WordPress, you want to apply that update because it could be an important security update. And if you get too far behind, you can never apply the security updates.
But now that update might. Change a little bit about what WordPress does in general, and that might break one of your plugins. So now we have to figure out, does that plugin also have an update made available and does that update maybe require some manual intervention to actually apply in Strawn? Right.
And suddenly people with a lot of plugins and have word prison still gets, gets get stuck in a, in a world where things can actually feel very brittle, right? Like, and where the very tight coupling feels a lot worse where. In this sort of microservice way, if an microservice breaks or something, it's typically like a very isolated piece of your site.
It's not something that will let a hacker take over your full database or even get X's through your server infrastructure, stuff like that. Right. It'll I mean that, there's like a localized component. That's not doing what you want. Right. So I think in terms of brittleness, I don't really. I don't really see the big advantage from, from there, from the plugin ecosystem.
Right. I think it's more about that initial, like opinionated set up piece. Right. And I think of course. A full sort of very monolithic system will always have like an advantage there. Right. But I do think that we're seeing a lot of, a lot of progress in the gem stack ecosystem and thinking through like, how do different services connect to each other and work together, right?
Like we're seeing on the framework side of things, more thoughts around like, how do we get data from, from, from these providers and make it easily accessible. We are seeing on. Tools like Tina's CMS or stack bit or Prismic slide machine that, that, that shows like how can you even have templates and themes and then have like normal content editors just plug in new components, drag things around and have things work in, of course we've done a lot of work with.
Building things like our identity service, where you can quickly add like an login widget and, and manage that. And there's tools like zero and so on for those kinds of functionalities. And I think, yes, the whole ecosystem progresses. We will see it. We'll see more and more tooling that makes it really easy to also get that initial sort of upfront experience very quickly putting together pieces.
And then I think. Overtime. Like if you add diligent, like, like Sarah talked about and look at like, what services are we using? Like evaluate the different providers and so on. You can actually get to a place where. Where picking the right ones gives you partners that are really taking care. Yeah. Of all the upgrades, all the maintenance, everything for their part of your stack.
And, and you can actually get to a world where things fields a lot less readily there. And then they can actually often feel like in, in, in this world of big monolithic systems where everything constantly, they need updating that, that works across like horizontally across all of your plugins and your main system.
Tara Z. Manicsic: [00:18:33] I think that was a really big thing for me about learning with the jam soccer texture, because you are as a team, right? Like either, if you're doing like a microservice architecture or doing more, more monolithic, you are always having those conversations anyway about what technology you're using for what, but being able to delegate those really hard things.
And it's the huge stark difference for me with security, because I've made a lot of. WordPress sites where we had to be up, you know, make sure we were updated on those plugins to make sure we were like countering and being either reactive or proactive about different security vulnerabilities. Whereas with the jam stack, it was just always amazing to me that I could pass off security in general, like MetLife identity and zero, because security is.
So hard because people, they, people have their job is to hack into your site. Right. And so they're constantly like doing their research and probably there's probably like a, how to hack someone's site weekly newsletter that they're all following. So like, to be able to keep up with that in itself was, you know, especially if, you know, you have limited developer resources, just being able to have a team who said.
Specialty is like that third party service of keeping your website safe. Just that kind of delegating to me is so huge in the JAMstack.
Sarah Drasner: [00:19:57] Yeah. I mean, even if you enable an analytics on an LFI site or, or, you know, something like that, what you might see is that there's a lot of people pinging that WP login or WP admin file.
And I think that that's true probably of every site on the planet, right? Like, everybody's just like, Oh, is there any. Security vectors. I can exploit right here. And that was also, you know, a big reason why we saw a shift in the landscape was thinking through how to be more secure all over the place and all around the world.
Phil Hawksworth: [00:20:28] And, and I liked the fact that, with this model of decoupling these various services and consuming them from APIs rather than collecting, you know, a group of plugins that you've. You know, you've, you've found through, I mean, they're very discoverable these plugins in the WordPress ecosystem and that's, that's, that's a huge advantage.
However, they're all living in the same pot. And so just because you're using just, you might not be using something to do identity, for instance, however, you still might be introducing. Compromise is that could, could touch that part of your infrastructure because they are all living in that, in that runtime.
The thing that I really like about this, this, this, the jumps like approach is that you're not putting all of your eggs into this one basket together. You can sit in the same way that. Tara was saying that you can inherit this expertise. You also put distance there as well. I really like the fact that you bring that capability into your team, so you don't have to become compliant for all, for all of the different things to do with security, but also you put it, put it.
Away from where your site is. So there's a great distance and there's a great decoupling there. So you don't have to take on that burden, you know, as you slowly gather and collect these, these various APIs and services, each one has its own areas of concern. And each one is brings with it, its team of experts and that's their core business.
That's the thing that I love, you know, I. I don't want to have to roll my own. One of these when there's someone whose entire business depends on them doing that properly. And they have to have all the compliance and regulation, all of those things that I've never, I never get to want to get into. So very
Tara Z. Manicsic: [00:22:01] undeveloped to say, you don't want to make your own
Phil Hawksworth: [00:22:05] you're right.
I mean, strictly speaking from
Tara Z. Manicsic: [00:22:08] the ground up every time.
Phil Hawksworth: [00:22:09] Yes. If I can't write a blog post blog post that says do X, Y, and Z from scratch and do that for every service. then I'm failing. I think I could get all of the hits that I'm missing.
Sarah Drasner: [00:22:21] Phil not invented here. Hawksworth
Matt Biilmann: [00:22:26] another area where that distance you, you mentioned becomes really interesting is this distance between the built and, and the, and the runtime, right.
Where especially again, when, when we talk plugins and so on, right? Like whenever you go to a WordPress site, right? Like, WordPress has to run. Like it will have like this, the loop, right. That sort of loops over every piece of content and so on and is from the database and so on. Right. And obviously like plug into you add or remove can have a lot of impact in what actually takes place in, in, in that loop and in that run and so on.
Right. Like, so, so you also easily with any kind of monolithic system like that, get these effects where a seemingly like simple thing you add is. Without your knowing, it might be exponentially increasing the amount of database queries you do for every single. Load off of your site. Right. And you might not notice it initially because it might just mean that under normal traffic, you get a little bit of flow down.
Right. But then you get a peak of traffic and suddenly there's like an exponential amount of increase in requests. And each request has an exponential amount of increase in the database queries that does, and suddenly your whole system just blows up. Right. And I've finding like multitenant dynamic systems.
So I've, I've really seen that somewhere, right? Like where. Someday. Like when, when I ran my, my, my web pup, like my CMS product way before Natalie Feinstone right? Like we had an at massive like increase of database queries. They were slowing the whole system to a crawl. And it was just like this one developer that had built some extension for like, Persisting like analytics data for, for users that came in that turned out to like in a worst case scenario, do 63,000 database queries per requests.
And suddenly the face that extension, it was on God, like an advertising campaign from United healthcare or something like that put, edit. Right? Like, and, and every single page load was trying to do a 63,000 database requests and everything just died. Right. But. But those are the kinds of things where if you add some extension like that to interact with data, and that happens during build time.
Sure your build we'll, we'll get a little slower, right. And over time that might be annoying and you'll want to go back in and understand it and try to make it the build faster again. Right. But there's a really big difference from your built getting two seconds slower to every single page load, getting two seconds.
Right. And to the compounding effect that this has when you suddenly get a spike of requests, right? Like, so that's another sort of security or our operations related vector. We're keeping this barrier between builds and query time during build and runtime and query time during runtime, it can be really helpful.
Sarah Drasner: [00:25:19] Yeah. And the ability to roll back. I mean, I can't tell you when I was in agency, how many times? I wish I had a feature like that when we, I mean, we used to have these clients who were really, really, you know, get huge amounts of traffic and we would, of course, during those spikes, that's when you least want things to go down.
Right. And that, that would. It would be a gamble sometimes to be honest, with some of the, with some of the infrastructure that we were using with some of the plugins that were using that, whether or not things would actually stay up with millions of people, hammering it. And you know, at times when you're hot patching something or something like that, I wish that we had the ability to just roll back to another state of things.
For sure. I probably shouldn't go into any more detail than that, but yeah.
Tara Z. Manicsic: [00:26:06] Spill the spill, the tea, everything. We've
Sarah Drasner: [00:26:09] all been okay.
Phil Hawksworth: [00:26:11] That's cool. Let's he? Is that,
Tara Z. Manicsic: [00:26:13] is that how we say that? But the kids are saying or do we drink it?
Phil Hawksworth: [00:26:21] The beans,
Cassidy Williams: [00:26:22] but that's, that's when you're like gossiping. So it would be accurate. You were right, Tara. I'm still here.
Tara Z. Manicsic: [00:26:26] Yeah? Oh, good. Okay.
Sarah Drasner: [00:26:33] We actually share something. That's like not the full story of what the client would like. So I worked for this agency that had some clients who were romance novelist, and I didn't know anything about romance novels before I worked. For that, I guess I have like books on my shelf.
People are like, what is this for a while? Like, it's, there's a. There they are the number one most, you know, like highest profiting industry on for, in terms of books. Like they're, they're a billion dollar industry that people don't really know about that much. And they're very, very popular. I just had no idea.
And so one of my clients put out a contest once and we got. 30 million for the company. Doesn't it involve this custom post type and format. Yeah. Calendar thing that I made and a few other people made, and we were just like, we didn't, we actually, we knew that there was going to be traffic. We didn't know it was gonna be that much trap.
but this is an author who. Has books and every Walgreens and every single super market. And so was quite popular. And wow. That was a learning day.
Tara Z. Manicsic: [00:27:47] It was my,
Sarah Drasner: [00:27:48] it was definitely a day where we were like sitting by the monitoring as much as you could in those days. I mean, I'm a neck beard, so it was a while ago.
we didn't have great observability then either. Right. I think something else that's improved quite a bit. in times of JAMstack is the ability to see what's working. What's not working what's up, what's down. And how much were, how fast that latency is around the world as well.
Tara Z. Manicsic: [00:28:13] That actually brings me to, kind of the final topic that I would like us.
To chat about a few of us have actually talked about, sorry, things, like, like spilled times how we saw that there is, there's a place where bill times can get slowed down. But now as we evolve with the JAMstack architecture, we have. Things like finally of build plugins, where I know that Phil and Jason have worked on things to help you cash things.
So your build time gets quicker. and there are more people doing more things to help, really speed up your build process time, because we've seen that that can be a sticking point. and because the stamp stack architecture is, you know, a new evolution of how we build websites. We're seeing this ecosystem grow, to solve these.
Issues. And so I like to call, I wanted to call this section the rise of Skywalker. but I mean, this rise of JAMstack, but, okay. Here's the thing though. We do a lot of things in the cloud, right then technically clouds are in the sky, so we're kind of walking the sky. So I'm still, right.
Phil Hawksworth: [00:29:20] This is a joy, this is a joy to witness.
You're such a craftsman. It's amazing to see you work your magic.
Tara Z. Manicsic: [00:29:28] I got a call five. You so, so one of the things, we also talked about where it was, doing things like using the CDN, which Matt said earlier, it's like big companies, like very appropriately titled CNN would have a CDN. but now we're seeing more and more that. Not only are, is it more available to us so that we get that like distributed globally and the redundancy, if, if anything goes down, you can be on multi-cloud providers.
but now the evolution of CDNs also gives us things like, like we have edge hands, which will give you just yet another way to kind of render things on your site, and even speed up how you talk to, different service providers. but. With all of you, what are things that you think about when you think about the rise of the JAMstack and what the ecosystem is doing to help?
basically give us more dynamic and more reliable and faster web answer me now.
Matt Biilmann: [00:30:30] And the beauty is that the ecosystem is executing, executing across all these different layers, the stack, right? Like there's obviously like provide us like, like us and others in the space that are looking at the whole sort of workflow and infrastructure is based.
Right. And how we can, how it can continues to democratize things that used to be really hard to do. Right. Like again, like if you go back 10 years ago, like, Looking for teams that had like a CIC D. Pipeline that could do like individual builds for pull request and branches. And that would like roll every build out for like an atomic deploy on, on a production, like on a replica and the production infrastructure, it with CDN based hosting and instant cash and validation and so on.
Right? Like do it have to look really long for teams that have built it. And you would know that that would be like, Really big teams with very specialized, like DevOps groups that had really invested a lot of work into this pipeline. Right. And now any web developer can essentially get that in a few seconds, sun on nearly fell right in.
And we keep looking to that kind of like. For example, you mentioned itch handlers, right? Like I've, I've talked to really big teams that have again have like big DevOps groups and like network teams. And so on that build really like complex things at the edge layers through tools like or Fastly or sometimes CloudFlare workers and so on.
Right. And. And we've been looking a lot of like, what, what are they doing and what are the kinds of problems they're solving and so on. Right. But it's still like, typically, like the real world use cases are limited to two pretty advanced organizations with like dedicated teams just to build that layer.
Right. So that's one thing I'm really excited about from, from our, the point of view of taking, taking this and then. Seeing, how can we make this, something that, that any normal like web developer can, can use. And that's like approachable and testable and it fits into the whole workflow and where you can write, like custom-made code that you can run with your friend and locally that you can open a pull request with and you get like the whole production version of it running.
You can test it recent about it and so on. Right? Like, So that's one example of like this infrastructure layer and how like, in terms of workflows infrastructure. So all of that providers like us can, can contribute. Yeah. Then there's the whole layer I mentioned earlier, right? Where we have players like stack bid and Tina CMS, and it prints make slices and many others that are really thinking through okay.
For the, for the people that are moving the content editing side and do. And, and the like marketing side space, like how, how can we, how can we build tooling that gives them the same sense of like live previews and, and, and, and direct interactions with the site that, that either want to list like WordPress or even site builders, like Squarespace, like can, can give people, right?
Like, how can we take like an architecture that's still nicely decoupled and still great to work with for developers and bring some of. Some of that to, to, to marketeers and induces and contented to strike, then there's a whole. A whole layer of, of developed men and research, I think around like four developers as we start building with all these separate data sources, right?
Like that's another sort of change from like the world where we had like my server and my database. Right. And there was sort of the two things we worked with. Right. And now. We've gone serverless, right? Like we, we no longer really work with that abstraction of like my server, but we also seeing that shift away from like my database to just all these different data sources where you might have your users in one place and your orders in another place and your content in a third piece and some custom data in some place.
Right. So now we also also seeing like a whole. Whole world of, of companies in the ecosystem tackling, how do we actually work with the developer across all these different data sources? Right, right. Apollo is, is doing really interesting things around Federation. It has Torah in launched there, the idea of Federation, right, where you sort of both have your connection to the database, but now you can also connect to, to other things like Contentful in store, and then just query it through one graph.
QL API, one graph is like, have also built like there. One graphic for us, all these different iPads and services where you get like one way to query it in companies like Christmas. Yeah. Thinking about like, how, how do we, how do we build like, sort of, if I'm more universal query language for, for different data type basis and so on.
Right. So I just see that as sort of a whole area. Where, where the ecosystem is making improvements and targeting, like how do we make it easier for developers to essentially like glue things together? And I think that that's the fascinating thing to see now that there's just like this explosion in, in the ecosystem targeting all these different areas of the stack and really seeing like, how can we mature it?
How can we make life easier? Both for developers, for content editors. Yes. If our operations people and them.
Phil Hawksworth: [00:35:52] It's funny. Cause there's, there's so many exciting things within that and you could harvest any one of those. And that would be like a really valid answer to that question, but I'm also excited by something which is really mundane by comparison.
And that isn't, I mean, or. Of course, I'm excited by the things that are new and arriving, but I'm also excited about the things that have been here for awhile and independent developers and maybe small agencies have been discovering by embracing these tools. You know, they've now got access to, you know, Instant cash, invalidation and CDN management without ever having to touch it.
The kind of stuff we talked about earlier on that you'd have to get an account management account manager and a project and try and figure out how to do all of these things. They've got access to these things. Now, as a result of this kind of explosion of the ecosystem, I'm excited about much larger enterprises.
Discovering these in larger numbers, because certainly enterprises are embracing this now, but I think there are so many more places which still don't have access to these things because are, they've invested a lot. In doing this themselves and in a way are kind of behind, you know, they, they have yet to discover how to use these things and, and benefit from it.
So I'm excited each time I see another large enterprise, embrace a lot of these things and get access to the instant rollbacks that Sarah was talking about and not having to worry about. Their success being leading to their downfall when they have, you know, their black Friday sales suddenly becomes a real issue for them, is probably a bit of my agency history kind of showing through that.
That was the thing that gave me the sweaty palms, but I'm excited for more people to discover the stuff that a lot of smaller agencies and independent developers have been getting excited about and benefiting from for a while. So. That's my 2 cents.
Sarah Drasner: [00:37:41] Yeah. I'm really passionate about developer experience as well.
And I think that this is something that wasn't even a con concept back when, right? Like the idea that someone cares about how a developer is feeling about anything in general, but nowadays thinking about how everything is matured and everything's at your fingertips in terms of choices. I mean, we even mentioned that you don't have to move away from WordPress in order to use.
JAMstack premises and make things secure and get rollbacks and things like that. that this piece, I mean, what I love about Nella Phi is this ability to just connect a, get hub repo, and immediately something is all over the world. And I think that that was the satisfaction that we got from pH or from FTP back in the day.
So it was kind of funny to explain where I worked to my parents. Because I think they were used to FDP a site back in the day. And so, yeah, I kind of had to go through the whole
Tara Z. Manicsic: [00:38:36] history of development,
Sarah Drasner: [00:38:38] explain, you know, what we do and, and how we do it. But I, I love the premise that all of these different pieces, the API economy, what Nettleford does, what all of the services that Matt was just talking about.
It, there's a focus. On making sure that all of these pieces of connective tissue work really well for the developer, because that developer activity, isn't just a nice to have. It is literally the thing that makes us not have to worry about the things that we don't need to be focusing on, on so that we can build the experiences that we want to be built and focusing on the things that we do.
Tara Z. Manicsic: [00:39:15] I like imagining that you're at the dinner table and your mom's like, so what do you do, Sarah? And you pull out a presentation and your dad's like kicking your mom.
She does this every
Sarah Drasner: [00:39:28] time too far from the truth. They asked me to stop speaking in acronyms, like CSS, but I was like, my brother's a product product manager at Google. And he was like, no, no, no. We use those terms. Like, like I'm like, okay, I can say things. Without the acronyms, but I think you're not going to, it's not going to provide more clarity.
Phil Hawksworth: [00:39:51] It's no acronym day. No, no. One's will have to use any acronyms at all for that, but you would be
Tara Z. Manicsic: [00:39:59] that sounds exhausting.
Cassidy Williams: [00:40:02] I ended up branching off. What all of you have said. The, the thing that I am very passionate about is making things accessible for new developers and people who are learning and stuff.
And I love how accessible everything is to newbie developers on the jam. Second in this ecosystem, like the fact that. Someone who might've just started learning JavaScript or reactor or any other framework or something can build a fully featured web application with authentication and security and, and images and video and address all, all kinds of different things, from the ground up.
Is so cool to me because that's something that I wish I could have had access to when I was first learning things. And so the fact that that's, that's not even just like, Oh, it'd be a nice to have to be able to do this. It's is here and anybody can build, build all that stuff is that's the coolest thing to me.
And then I love being able to talk to people who aren't sure where to get started. And you can just say, Oh, well learn this, this and this. And then. Go and, and they there's just so many resources available and so many services available that way. Make it easy
Tara Z. Manicsic: [00:41:07] for them, for people who aren't familiar with the term headless.
It's not as morbid as it sounds. this is basically saying that you can have content coming in. And the, it, you just throw that content in and it's something like maybe you, why that content contributors can be familiar with, but it doesn't care how you consume it. So you're getting the data in there and we use, for a nonprofit site that I do for, for, Asian Natty.
we have a lot of people who, contribute the blog posts. And so they just see this user interface where they put the title and they put the content and they upload pictures and they use that in the strappy UI. And then we consume that, and get it all through on the back end. Right. And style it all out with our, With our static site generator.
And, it's just, we never have to explain anything to them. This is a form that they're so used to seeing on every website. So that's the kind of a very, we'll probably have a whole episode on headless, technology because maybe a Halloween episode on headless technology but we are reaching time. So, if anybody wants to one last comment, otherwise, we'll head out for today.
I
Sarah Drasner: [00:42:16] just want to thank Matt for joining us. I mean, it's so cool to work with the founder of jams tech. So it's such a pleasure to hear your thoughts on it, especially because they're, they're so mature, right? Like thinking through what was in the past, but also what's in the future and I really appreciated you.
I think we all really appreciate you
Phil Hawksworth: [00:42:34] joining. Yeah. There's so much stuff we didn't get to cover, so we'll have to. Get you boozy and talk
Sarah Drasner: [00:42:42] after hours,
Tara Z. Manicsic: [00:42:45] we'll have to pause you from advancing the way we make the web in order to talk to us on our podcast.
Matt Biilmann: [00:42:52] Thanks so much for the invite. Well, that's our show.
Cassidy Williams: [00:42:54] Next time we're going to be talking about virtual tech in quarantine conferences. What works or not for both attendees and organizers, it's going to be a blast.
I've been Cassidy. Strawberry
Phil Hawksworth: [00:43:04] jam Williams. I've been filled peach jam Hawksworth
Matt Biilmann: [00:43:08] I've been mad. Rhubarb jam feel men
Tara Z. Manicsic: [00:43:11] I've been Tara big
Sarah Drasner: [00:43:14] and I've been Sarah, Lala Berry jam. Drasner
Tara Z. Manicsic: [00:43:17] that's perfect. Thanks everyone. Hi everyone. Thank you very much for joining us.
Matt Biilmann: [00:43:24] You.
Sarah Drasner: [00:43:30] Everyone was
Phil Hawksworth: [00:43:30] smiling at me, horrifically what's happening.
Something happen.
Tara Z. Manicsic: [00:43:37] Lovely. So it's fine.