 all right good morning everybody I'd like to say welcome to Drupal con I'm excited to be here how about everybody is everybody excited I mean it's nice to be back in person after a few years and so I'm super excited thanks for joining us today we're going to be sharing about a project that's been over two years in the making and so get ready to walk through some approaches from inception through the launch of the first site using this new platform at Penn State so a little bit on the agenda for today these approaches are all rooted in strategy so we're gonna be talking a lot of strategy today we're gonna start with product strategy and which was kind of that guiding light through the entire project and then we're gonna dig into the technical strategy for the project which will guide us through implementation and then we'll wrap up with some insights which I think is really interesting today because that's gonna help us talk through some of the similar current and future project things that will help us in those upcoming projects that you may have all right we'll jump in a little bit about us and I'll kick off since I'm already talking here my name is Mark Shropshire I go by Shrop I've got 20 plus years of experience at the intersection of leadership and technology some of my past roles have been with universities and a nationally recognized graphic communication company some tech startups prior to coming to media current which is where I'm currently at I love helping individuals reach potential big into leadership and mentorship and and frankly empowering leaders and empowering everyone to become leaders who wants to and so I will turn it over to Jim now hello everybody thanks for coming my name is Jim Norris I'm a developer and a program manager with 20 plus years experience in large decentralized research university such as Penn State but also University of California Harvard Dartmouth Boston University my passion is helping moving big things from A to B sometimes as a developer sometimes as a leader in places where change can often happen slowly and my key role in this project was to I like to think is earned by an internal investment from the C-suite leaders who sponsor and and fund change but supporting nonprofit education research and health care enterprises is a form of public service to me that I am very passionate about so Penn State is big Penn State is really big and not only by accounting the number of students and alumni our research enterprise is enormous on the scale with John Hopkins MIT Michigan if you measure by National Science Foundation funding metrics and the like we're one of only two universities with all four land grants that's land Sun excuse me for grants land Sun space and see and the website we're here to talk about today our news website is a single large website that tells and has told for 20 years research stories that highlight the impact of our science and our research search and the impact it has on the Commonwealth and the people of Pennsylvania on the nation and on the world as long as telling a number of other public relations stories so but my perspective comes from this the central office of strategic communications and marketing at Penn State so we functionally own news storytelling it's our website but we're also stewards of the brand itself so it's notable here that Penn State has over you know best estimate over a thousand sub-domains distinct to date that are bound only by the loosest digital brand standards web strategy stacks and operations and just hold on to that thought so I just want to mention a little bit about media current since that's where I'm at and we were a partner with Penn State and Jim's teams throughout the project media current we bring together talented team members and we like to build those world-class solutions such as this Penn State news site we're really happy and proud to support and focus on open source solutions being an open source product company including but not limited to you know projects that are out there like Drupal like Gatsby react things like that and we're deeply passionate about partnering partnering with clients and their teams okay so this was a complicated website build it was a visible project and while involving a lot of new technologies multiple partners and a lot of moving parts and ultimately it was successful but it was one of the most challenging projects I've ever been a part of and I'll be very honest with you there were a number of times we paused and asked ourselves why are we doing this again why would a central communications office select a decoupled architecture for one of the biggest multi-channel highly integrated on brand websites with hundreds of authors and very low fault tolerance well if you stay with me I think that question can start to answer itself we found that something this big and complex demanded that we break it down to each component segment focus on each thing as a service in isolation just to get our minds wrapped around it and thinking like that unlocked many possibilities for us to scale the work for bigger purposes purposes and services in service of the brand and ultimately the institution but there were a few tough months where I and the team were asked difficult questions from our internal stakeholders which led to some uncomfortable discussions with our partners but we all knew that we had the right strategy right we had asked the right questions we chose the right partners the right services vendors technologies many difficult but necessary conversations with our leadership about the total cost of investment and how to truly calculate long-term savings and measure and measure ultimate return when you know these conversations were successful because we had the right long-term product strategy but first we need to tell you a little bit about the business situation we faced three years ago so our legacy website so it was it is it was a big website lots of users loads of integrations up and downstream it was very custom it was it was hosted on premises it was a Drupal seven monolithic architecture supported in part by vendors and in part by in-house system administration labor with some fairly old-school web ops methods and it was becoming hard to maintain it was d7 was end-of-life or so we thought the communications business situation and is and was like functionally we're a decentralized public relations and communication business we're a large institution we have many campuses we have many departments we have many units and they all have their own distinct website or sites their own workflows sometimes their own goals authors put all authors all most authors around the university do have a connection with the new site they're putting their content in our website and they're been in the habit for a decade of fetching their content back out either by API or RSS or some combination of both and the content that's published in our system is available multiple channels so we publish the stories and our website but it's also anything that gets approved by our editorial workflow is it's queued up for email blast syndication and ours this platform integrates with an old less serve that you know we calculate it's low estimate something like 30 million individual email blasts it's a major part of our marketing strategy so big site 250 authors representing 20 plus campuses locations 50 plus you know major departments units the colleges okay it's big our big idea was three years ago what if we broke up what we already had into atomic segments and then glue it back together and manage it chunk chunk by chunk each as its own thing so convert the legacy website into a platform with reusable segments build on what we already had because what we already had was a content hub we already had a business process where users across the Commonwealth are submitting stories to use on multiple channels and getting their content back out we already had apis we had a design that was reflected our best current thinking on the brand nearly every segment in this one large website could be thought of as a reusable component so we thought so some principles started to get clear start to think beyond websites strive to avoid the constant cycle of rebuilds leverage the brand right from the position we sat where we were we stewards and owners of the brand leverage the brand in our position in the university a stewards of the brands as a call to action so I call this my the fever dream slide because you know and everything's going great this is where we're going right so enhancing and protecting the brand is an objective that helped us think but beyond a single website and towards a set of tools and standards that we could anchor a broad web strategy to so we tried to elevate this work and to have impact on the organization strategic plan and on the bottom line itself so as many of you or all of you know academia and dot edu is the ultimate horizontal business it's the opposite of top down where mandates do not work and if you're serious about change you have to make friends and partnerships and facilitate change with carrots so our creative and our development partners and I'm talking in-house and our vendors the people who execute the brand they need things like digital strategies and standards and api's and content strategy patterns rules style guides they also need access to authorized cloud providers Penn State authorized and modern webop solutions so we're not encouraging mere copying of things we're either we're we're building starter kits and we believe that taking a thoughtful point of view particularly on the design system segment of all this and in all of its related strategies enables a centralized digital center of excellence one that can perform things like digital strategy consulting SEO optimization across the domain web product and project management consultative services trying to catch like all major web fact all major web refactors sort of early in the process so we can have some influence on their outcomes central UX strategies leading to you know digital transformation and finally this is my highest concept thing so agile is more than a framework to us it's a philosophy it's a philosophy about how to brace uncertainty when building products and adopting technologies like antifragile things are things that benefit from a little controlled chaos and I think this is a useful mindset for a technical strategist at least and open source itself it's antifragile in the nature of its construction and the way the community supports it and I'll name check Gatsby here is a sort of partner that has learned from the stressors that we have put on their system and they have hardened their infrastructure as a result not just for us but for the whole community for the whole framework so okay that vision is real nice but we still needed to build this big performance news website we needed to migrate seven gigs of data need to re-slug every story we needed to host in the cloud and automate our DevOps drop we need that now please you know just noting some of the requirements you know custom authoring workflow preview security search both Google and an internal search content API RSS single sign-on data migration slug pattern recasting we need to glue it to a single domain strategy lots of integrations etc etc and plus we need preview to work the way authors expected it we needed to be able to publish our stories within five minute increments excuse me and while you're at it please build us a scalable platform so here workflow segregation means to us and power Penn State to own the segments that are closest to our business so our design system and our front ends these are good examples of that cost I choose to think this is in turn think about this is more about cost predictability is the primary driver here so accurate cost forecasting is critical right when you're doing big things especially when it changes how your organization operates so there's major changes happening in the IT business here at Penn State concurrent to this effort there's an emphasis on more cloud services and away from IT of treating IT as overhead for example so our previous setup administrative straight of Lee really was one where the cost of PSU dot edu and the new site they were largely absorbed in IT overhead we had no real good way to cost this so our vendor cost hit the chart of accounts but a less system administration security network infrastructure DevOps emergency response hardware config patching upgraded monitoring logging all that was unpriced and unbilled it was just unseen overhead and this led to our stakeholders to in the past to believe that once a website is done we can stop paying for it right well no being able to price each segment and forecast and plan its maintenance allowed us to change how leaders think about the total cost of investment in these technologies and our websites that are critical to us this is my last slide so when we decoupled everything we landed in a place where we can now be selective about how to best maintain each segment of the platform as a commodity with its own life cycle its own support cadences we can make choices is one segment can it be cloud source can it be outsource can it be in source can we do team augmentation manage each segment opportunity opportunistically as the need rises we're thinking that how we optimize and manage in scale of the design system segment which is the next frontier of this platform the design system in its ancillary code and instructions are how most of our adopters are going to experience the segments of the platform that are relevant to them but design systems you can hear a lot about design systems I feel this week are not enough to accomplish brand unity and great web strategy so we think of it as a library of plays and a playbook strategies that come with codes and guides and other things so one example of what I'm talking about like a single component organism that blends design tokens grid patterns multiple apis all working together is it simple for a major university federated navigation brand bar seems simple right what if you wanted to reuse that across an organization there's various methods so ours contains an alert a be a API from Firebase via Drupal it has Nive items from S3 it's via right now it's via expression and soon to be contentful maybe we have Google search apis patchy solar it's all wrapped together in a single component that can function more or less out of the box on our adopters websites so I've said enough I'm going to let shop tell you how we actually achieved some of these things well first of all Jim thanks for being such a great partner throughout the project putting in all the hard work and of course the teams we're speaking for teams a lot of teams involve lots of people involved making this happen across you know Gatsby Aquia Penn State media current and so but let's shift into technical strategy so I'm promise for everybody in here that's that may not be super technical we're not going to get really in the weeds we're trying to keep this at a higher level but I will still hopefully satisfy some of those that are technical with some technical talk and involved and we're happy to talk about this after the talk this week Drupal com but we'll kind of shift into things and and really more about the decisions of technology why we made some of these decisions for implementation which I think are foundational in this platform so we are talking about the Penn State news site but we're also talking about building platforms and so so speaking of building foundations you know one important message today if you took away anything besides everything Jim said because that's all important but but I would say stop building those websites and start building platforms and this is the this is the thing Jim you talked about over and over I mean we you know he cast the vision for us and so all the teams involved kept seeing it this way that we're building kind of a future something that could be used over and over at Penn State this this Drupal and Gatsby and Firebase and all this all these technical things we're talking about like can be packaged and reused as a platform for the future so so when you take on a task building a platform you still have to think about every decision along the way which is it's pretty overwhelming there's a lot there but you've got to step out of thinking of singularity of purpose you got to start thinking about the larger vision and provide flexibility and maintainability for everyone in every piece along the way in the platform so let's dig kind of a little bit deeper into some of this so I'm going to start here with design systems you know Jim mentioned it already but I feel like this is the this is the important piece to start with you got to start with a great design system it's an initial step it's setting the vision for the platform so we needed some then it's highly flexible component-based I'm gonna say some other buzzwords here about a customizable options and yeah we've all heard these things right but ultimately what am I really saying I'm saying that design systems are ultimately critical and then once you understand that and you start building towards that you can then figure out the technology that you need to implement and so we decided to leverage Storybook which is a React open-source product and so it's it's a well-established component-based system and that allowed us to create a what we call a living style guide but ultimately what that means to me and the team and everybody involved is that there is a there is that that playbook and you mentioned playbook there's this playbook that you can refer to and say this is how things are supposed to look but not just how they look this is how they're supposed to work and that that's critical so it's not just static so you know by building these the components that are flexible editors can arrange the components in different orders and so if you're thinking in a Drupal context you may already fast-forward mentally to where I'm headed with that but to be able to change the order of how things look and feel on a page or in a news story and yes that fast-forward if you thought about it you're right in this situation we've used Drupal paragraphs to make that happen but we've also implemented a great deal of authoring experience work along the way and I'll talk more about that but there's just a ton of work on authoring experience because if you can't satisfy you know the 250 content editors that are going to be using a platform how are you going to satisfy all of the millions of people who are going to consume the content so design systems tell that story they set the stage and everything else is in the platform from there so this is a fun slide be involved in Drupal for so many years this is always a fun slide for me because I've seen Drupal change over the years I've seen it grow and kind of change with the times as the web matures new things happen Drupal has kept up in many ways and adopted new technologies but Penn State has a history of open source I think it's great that made us a great partner we're an open source product company at media current but but we wanted to build on what Penn State already had and that was something that was set up from the very beginning of the project Penn State's got a long history with open source Drupal many other things involved but but here's some of the benefits how we see Drupal fitting this project and and really a lot of other projects but particularly this project so Drupal being a what I like to say and others have said a content management framework and what I mean by that it's not just a system it's a framework you it's it doesn't have to be opinionated you it's very flexible you can build a lot of things from it and from the ground up of course it does have some opinions but you know every system does to some extent but we see it as a framework it allowed us to build and customize that authoring experience and this this happened through fields and taxonomy content moderation user management and more and by the way is a is a big Drupal fan I know others in the room are as well you're at Drupal console there should be I have to say I love the fact that so many features are just in core and I can flip a switch turn it on at least have a baseline to start from right so but Drupal also empowers content editors to control the layout of content so this is true for traditional Drupal sites but it's also true for decoupled sites that we're talking about here so editors can trust what they edit and then preview is what's going to go live into publishing and this sort of fast forwards a little bit more into our implementation with with Gatsby making that happen because we needed Gatsby to actually do some of that but I don't want to get too far ahead it's exciting just hang with me Drupal also has an API first architecture this has been talked about for a long time Dries has gone over this and many keynotes but lots of people have worked on this that I'm highly grateful to in the community but this allowed us to make this Drupal site be the single source of truth and you know Drupal's power is in its content being a content hub so it'll not just have successful governance you know policies related to that along the way and then many consumers you know these are applications or you know it could be mobile apps could be websites could be RSS feeds whatever but you know there's always these points of consumption from the Drupal site but we're excited to be here at DrupalCon talking about innovation building new and modern web on top of Drupal again I'm just always excited to see Drupal move forward and and it's the it's providing the back end for a strong decoupled system so we see the platform here as a potential model for other EDUs as well that want to build on top of this so if anyone's in the room has been involved in higher ed works at higher ed has worked in higher ed I have in the past and I can say that you know one thing I really like about higher ed is that open source feel and vibe is like alive in the academic world so and that's where a lot of that came out of so it's really fun to be involved in a project like this and help out help out other universities so kind of speaking of scale and all the things that had to be done on this project we did spend you know that time on the content authoring experience I'll get a little more detailed here so we've got 250 content editors Drupal's providing a very important service and it needs to be easy to use for those folks so Drupal's ability to you know create those custom authoring experiences and necessary workflows has been highly utilized here and here's some examples so some of those baked-in features users roles permissions content moderation was key taxonomy and I feel like for Drupal 4.5 when I got involved with Drupal which is dating me maybe but like taxonomy has just always been there and it's something everybody talks about and I remember the first time someone said taxonomy I'm like that's what I get it from a library standpoint but what is that really going to do for me but over the years I've seen taxonomy you drive so many things it's it's pretty pretty mind boggling but but beyond those built-in core features we leveraged contributing custom code to as you as you do to cover additional things like schedule publishing we limited this sounds so simple we limited field character limits and we had a UI where there's a module to actually provide a count of those you know how many letters you've typed in the field and I don't want to belittle something it seems so simple but it's so important to a content editor when they view they can see that there's a limit and they see the count that's just helpful it saves time we were able to spend time focusing on clear field labels and descriptions that is not the most exciting thing for a developer to work on I'll just say it but I will say though that like we have strategy partners media current folks like that they're willing to jump in and help with that and it is it is important I mean it has to be clear what does that description field say what how does the user even know the content editor what to put in a field so really really important so so Penn State's partner with aquia to also ensure that Drupal runs well for all the editors and the API in points so there's a lot happening on this Drupal site it's pretty big beefy infrastructure going on with aquia that's providing lots of API in points and at the same time all of those content editors editing and it yes I can talk more later on this week but there is a lot going on to keep everything coming along well with that system and we'll talk more about the API capabilities here in just a moment so so here a little bit little bit more on Drupal's the permissions roles content moderation making it easier really just the key benefit here and it's really critical for the news site specifically it might not be as important for some of the future sites on the platform it might be it just depends but on a on a think about a large-scale newsroom I mean you saw the stats I mean this is turning out lots of stories and images and videos all the time every day so so there's a lot of nuances in the mix here and you would be right content editors for instance can draft and this is a real simplified kind of a description but it makes sense and it probably sounds familiar to some of you and your projects but content editors can draft and work on content but they can't publish but beyond that like that's a real simplified version like a small group of people can actually publish the content but one thing that we spent time doing was actually hiding fields and from the the node edit view making sure that certain fields just aren't visible like so it's not enough the person doesn't have you know content editor doesn't have access to actually work on a field how about just don't even make it show up how about making the interface simpler for the content editor those are things that we worked on okay so I'm gonna jump a little bit to something a little different here and that's separation of concerns and this is this is something that plays off of what Jim mentioned earlier about segments of the system overall and don't worry we're about to get into some interesting Gatsby parts of the project too because that's the other half of what's going on here so we've talked about Drupal and all the features and you're wondering probably like what about Gatsby well let's talk about it so having Drupal and Gatsby operating independently allows content editors to work on their content while publishing only when ready that sounds simple and you might go well that sounds expected and I'm like well you can do a lot of things to kind of make that not happen in a system so on a technical side we benefit also from separating our development workflows so not only do we have that like high level content editor and publishing separation but even our developers can operate independently working in Gatsby React code base and separately in the Drupal code base and you have your teams that are best at each of those things working where they need to simultaneous and they can come together when they need to but a lot of times we find that operating in the Gatsby segment is all it's needed for a certain feature so overall the platform benefits from allowing teams to optimize improve work in any segments and not having these monolithic entities that they're having to work with so we're talking about this earlier Jim but it's like it's like we're kind of talking microservices but not like we're firing up a bunch of AWS Lambda instances not those kind of microservices but it's like breaking up a system into smaller components kind of microservices okay that's a lot of stuff I'm gonna take a breather I'm just kidding let's jump right in speaking of Gatsby stuff so this is important like and you know maybe you haven't used Gatsby before maybe you haven't used maybe you've heard all the talks about decoupled and you haven't used decoupled but but I'm gonna give a little bit of background here that I think is really important about why Gatsby is important to the project and how it works from from my standpoint and so so there are three components here that the way I see it that provide development publishing workflows that that we need for the Penn State platform there's the Gatsby framework and that's the react-based framework that allows you to build those fast secure stable accessible sites in Gatsby which is you know totally makes sense like you got to have the developer tools you got to have the things that spin up and work for developers and that allow them to to build the sites and build the Gatsby components in and then a really critical part Gatsby cloud so this helped us deliver that unified platform of building preview and deploying Gatsby and this last part is so critical it became critical I don't even think we realize how critical was but these are not just buzzwords being able to publish those to a global edge network this is not just words I'm telling you it is critical we tried deploying and when you performance test against not an edge network and on a scale like this the site if you've ever seen it the sites will not hold up so so really important and and we got to kind of be a part of a lot of a lot of the the growth of Gatsby cloud and the release of the Gatsby hosting product so there's also the Gatsby content mesh I don't want to skip over that that is really the decouple data layer that that it's it's GraphQL from a technical standpoint provides the tools for the sourcing normalization updating of content and this is you can think of this is kind of what sits in between Drupal and Gatsby it also sits in between Gatsby and any other source you're pulling from you can pull from markdown files you could pull from some other API you can do all kinds of things with Gatsby but having a common GraphQL source to actually you know query from is is really a benefit there because you're not having to like switch technologies and go oh how do I query that technology now like how do I you know how do I hit this API kind of don't have to worry about that in the same way so these are just a few of the benefits while the integration worked out really well for Penn State and for the project but there's one more which steps us back in a Drupal content editing experience a little bit so this is a little combo of little bit combo of Drupal and Gatsby which I think is great so so a big benefit of Gatsby for the content editors involved here is the Gatsby preview and that lives inside of Drupal and what's really cool about this is I think it's it's a pixel perfect preview and I'm gonna say that again it's a pixel perfect preview inside Drupal of the content exactly the way it's gonna be published and look when it's published on the final site that almost doesn't sound like it makes sense as much until you really think about what it means to a content editor to be able to really see what it's gonna look like immediately is a huge thing it's a big deal there's no loss of fidelity at all editors can see what they're gonna go live with and also keep in mind some of these editors aren't the ones hitting the go live button right they want to be able to hand it off to somebody that can hit the go live button and make sure it's the right thing from their standpoint so so that same design system that we've been talking about today earlier on remember that first thing that's critical starting this a big platform and making all these decisions that that same design system it powers the Gatsby front end it also powers the preview system within Drupal so we're using everywhere and and Jim was telling me some stuff earlier today like all kinds of plans he's thinking about in the future but like it's it's it's making me get goosebumps going oh my gosh it's crazy I can be transparent with everybody right so it's exciting all right so let's let's talk a little bit here about the content API content hub concept with Drupal so I had had the words here epiphany I think that was Jim's words it was like an epiphany the team had Jim had but you know he mentioned as earlier Penn State was already operating as a content hub but there really wasn't a formal strategy for it right like you have Drupal and it's got a lot of API's connected to it what does that really get you well it's great it's functionally working but if there's really not a strategy you don't even know how to maybe contact and work with consumers of your API's and things like that so so really this project helped realize Penn State to have a reusable content hub as a service so thinking about the everything as a service so it's almost like within Penn State what is being done within Penn State and isn't being outsourced to like cloud services like Aqua and Gatsby and companies like that it's almost like operating another service within within the university which is huge and Jim can speak more to that himself because that's what that's what he oversees but but a universal API standard is a foundation to the platform going forward is a big key here and so JSON API is mostly what we use for for this but we're not relying on just the single core Drupal JSON API point we're using it elsewhere and we're also delivering you know from a content content agnostic scenario here to the platform like we don't really we really don't want to be opinionated about what the you know what the consumers who the consumers are what they are we just want to be able to provide a service that we hope about anybody can attach to their services to their sites their apps and all those things and it'll work so Drupal Drupal API first architecture allows this platform to provide content through a variety of methods I mentioned core JSON API that's driving Gatsby publish and preview builds we're driving campus alerts which is is really something close to my heart because that is so critical on a campus and if you've been involved in higher ed you realize you know things can happen you need to get notices out to people quickly and there's a lot of time spent at universities building alert systems and this is the start of something big for Penn State we believe so we're pushing that from Drupal so a Drupal only authorized content editor can publish an alert and that's really restricted and that Drupal pushes that to Firebase and then Gatsby picks that up from Firebase after the page loads so it's not involved in the in the Gatsby build at all which is a huge benefit and then in campus sync which is a J another JSON API endpoint which allows you know all these other partner sites all these other Penn State campuses to actually pull data and and pull that content and use it as they need on their sites across the system so it's that reuse of data so if that sounds interesting think about it we're reusing content we're reusing content we're reusing content over and over again and we're not retyping content we're not here here's a copy of this go put here's a word file go put it on your site it's being reused it's huge so so I think this project to me was and I tried to look it up I don't remember which Drupal con and it blurs in my mind but Dries at a keynote talked about the vision of the API first infrastructure you know Drupal being API first and being a content hub he spoke about that a number of years ago and that stuck with me and I think this to me was like the dream realized a bit so anyway that's that's content hub and I'm gonna jump into a little bit here about how we maintain the platform like maintaining this platform is like never done right that's true Jim mentioned that for any of our sites that we work on in projects but you know we may see a cycle of you know go through a cycle of seeking out like those are the gaps in our code and defects and what opportunities can we improve on a platform so the key here is really about taking those segments that we've talked about throughout this talk and then say you know what we can maintain and optimize and work on each segment independently we don't have to break away and and say you know hey we've got to take down the entire site to do a release no we don't we can release each segment independently this is really exciting recently so we had some search improvements to make and we're using we're using aqueous search which is you know it's aqueous version of solar search they've got some other stuff wrapped up with it and and then we have an API from Drupal and then Gatsby when you and I'm in Gatsby front end you do a search it actually talks back to that Drupal API and and and returns results but that's the overview of the system but we had to really focus on each of these separately so it was really it was really neat because all of the changes that we needed to make were in Drupal like we were able to focus on fixing some Drupal API pieces and and it took care of all the things and we did a Drupal deploy just you know to put those out there when we were ready and it didn't it you know Gatsby said they're operating has no idea anything's going on as far as that deploy it just continues to operate as normal so to an end user nothing had to be taken down all systems were still running this is important for this platform and but to help visualize a little bit of this this is this is the slide so I'm gonna pull it up this is a little this is a little Friday right because it's a lot going on I know there's a lot going on and and feel free to take pictures and all that stuff but I will post the slides afterwards this is the technical overview it's not as crazy as it seems once you get to know the platform but I did want to hit on a couple of quick things and that is that that notice notice some things we've talked about already right Drupal and Gatsby have their own get repos which help those separation of concerns and not always on optimization us to be able to manage those the flows those development workflows separately you know we've got Acquia providing hosting for Drupal Gatsby's providing Gatsby cloud hosting for the Gatsby front end but we've also got some Amazon S3 involved which is providing our image optimization and file services and there's some other pieces that Amazon's doing as well but but I want I did feel like it was important and Jim and I talked about this to show show this and and to you know be transparent about we're talking about these segments what do they look like that's gonna be the next question how do they work so we're gonna jump in now to a little bit about insights thanks Rob I'll blow through this somewhat quickly so one key insight from me is I think everyone in this room right has done are you you will do one really really hard thing in your career when we've all worked on a project where there was you know one key person who had to make or break an effort like a real MVP this project had about 10 of those Shrop being one of them and then they came from all three you know segments of the partnership so I was really really proud to be involved with this sort of matrix of people and so a key insight for me is like I said at the beginning know your business right and have a strategy and believe in your strategy maybe point to a product owner if the project's big enough right really good RFPs and try to choose some partners who have seen some stuff as Shelby foot said it had four o'clock in the morning courage if that means anything to you all right so one of the things as far as challenges that they hit on here is new tech always pushes the limits this is something you know Jim you talked about a lot and you know being on the bleeding edge definitely means that a lot of times there's there's a few peers to talk about the challenges with you know there there's times where like Jim and I were on a zoom going we don't know what we do and you know we reach out to Kyle and Kyle in Gatsby helps us out with something and we're like but like that like it's a small circle of people that have worked on something this large and it's exciting at the same time it's very scary I mean I'll just be honest with you but um but so you know one of the pieces I think that helped though along the way was it led to us working through proof of concepts I think proof of concepts is important in our work and like these large projects we have to have ways to prove that an idea might work and limit that so you know we had some examples like we did proof of concepts along the way to solve the front-end hosting of Gatsby and we went through different scenarios of using AWS and Netlify and and we landed on Gatsby called Hust hosting and and I have to say in it it's not just because it's newer but I really for I believe that that Gatsby Cloud hosting is working so well because it's it's geared for the platform it's it's engineered for platforms optimized for a platform so so that was important though to go through those pieces and you know we're always pushing on these systems and so so it's anyway it's it's really exciting to work on that and the in the proof of concepts or something to always be thinking about would be kind of my advice on that we did have some also challenges with long build times as you might imagine 65,000 plus news stories over 111,000 images I think when we launched after we migrated from seven to nine and this meant Gatsby Cloud like if you just kind of threw all that at any type of system that is in a build architecture like Gatsby as there's other systems similar you know that do that but like you just throw that out there without thinking about it like it's it's a lot like it's just you can have I'll just shorten it you can have some challenges right so so it took that partnership that we've been talking about and and I know you're like he said this a bunch Jim said it a bunch about partnership but I'm telling you partnership between Penn State Gatsby Aqua Media Current I mean it takes all that to solve these big challenging problems and it's a totally real thing so I'm gonna let Jim jump in on it so one one challenge maybe this is more of an insight I would share and I'm speaking mostly I think right now to sort of marketing managers and other people that sort is we're reminded here that great technology doesn't implement itself right that there's lots of new bleeding-edge tech in here but it took a lot of time and thought to sort of build it and wrap it around a context that made sense to pay off it's not doesn't always come off the shelf another challenge was getting our major stakeholders and again these are people functional owners editors people that control funds flow and release funds for projects getting them to adjust to a new mental model of web took some doing and you know just one tiny example is the sort of value trades we're always making but one is trading TTLs for resiliency you know sometimes authors simply despite all of the last 40 minutes they want that story to be published immediately well it's not immediate it's very it's dang fast but it's resilient and it's safe and it'll hold up if it's ever under amazing stress so don't skimp on the change management set the right expectations I'm still talking you got it I'm still talking good so successes the author experiences still Drupal right so while a lot of stuff changed our authors didn't really feel it it's a little leaner and so our unit users adapted to paragraph styles without much difficulty how we're thinking about federated components and the design system has some really exciting applications for analytics now we can now build in analytic tokens to track and filter sort of by component right and so per adopter we can track and see how these various components are doing and optimize accordingly if we talked a lot about campus alerts we are kind of we are we are very excited about it's a good example but it's it's really just a Drupal content type it's one that hits Firebase it becomes an API data that's immediately available across domains and we can push out alerts potentially potentially down the road to you know hundreds of domains a key point is the legacy website was delivered on its own sub-domain news dot PSU dot EU and we migrated now to live on PSU dot edu slash news so PSU dot edu is not Drupal and it is hosted in S3 so we've glued this whole thing together to PSU dot edu via reverse proxy and that's part of a long-term strategy to move away from sub-domains you see a lot of in academia towards amp away from it some debates towards amplifying the value of PSU dot edu domain itself which will ultimately we think raise the entire university's SEO value and if there's time to talk about our open source contributions yeah I'll I'll kind of briefly mention this but there's a lot here and again all of this stuff I know it may open up lots of questions come see us this week but to talk about it happy to do that but I just have to say like this is this is open source working I mean it doesn't mean there wasn't challenges and struggles and like maybe frustrations and all kinds of things in the middle but but we we were able across you know Gatsby and and media current to partner together like there were a lot of contributions that went back to community that's built made Gatsby a better platform for many people that things that maybe they only could see problems with on this platform but now like helps every site from the smallest Gatsby site to the largest Gatsby site there's like Drupal contributions not just on the Gatsby integration that went on but throughout Drupal as you do on these large scale Drupal projects like there's things you contribute back to and I just encourage everybody to always be thinking about how can you contribute how can you give and it's not it's not always code I just have to say that and so with that I'm gonna say this is this is the slide this is this is a domain whatever you say on that this is exciting part here this is the this is the product go check it out yeah it it works it's real and so we I know we're kind of short on QA time again like there's this there's so much we could talk for a lot longer as you could tell and so please ask us questions shoot us emails talk to us at the media current booth you know Jim and I both will make ourselves available to chat happy to do that and and I have to mention I do have to mention Drupal Camp Asheville this is a really awesome Drupal Camp encourage you to come check it out they're accepting sessions now and this is one it's worth flying international for I'm telling you it's that good but anyway thanks thank you so much everybody