TD Cafe #008 - Martin Anderson-Clutz & Jürgen Haas

Juergen: we came up with the idea of modular API to be a generic link between complex configuration in the backend and users in front of a browser, needing great tools to actually manage them and configure them, and only justifying that for ECA. Was not quite enough

Martin: Well, hey, Jürgen, it's always a treat when we get to, to sit down and visit. Uh, definitely enjoyed our conversations that we had at Dev days in Lubin earlier this year.

Uh, lots of great discussions. Uh, learned a lot about some of the workflows that you use as, as sort of a module maintainer, and then you publish those articles afterwards. We can put the links for those probably in the show notes. Um. Yeah. Um, some of the things that I thought maybe we could talk a little, a little bit about today are some of the things you've been writing about recently, particularly modeler, API, and some of the, the changes that that's introduced.

Um, both for ECA, you know, its ecosystem, but, but I think more importantly now, being able to leverage that BPMN interface for ai. So, um. Why don't you talk a little bit about how the model or API is actually different from what folks might be used to using in older versions of ECA?

Juergen: Well, yeah. Hi Martin. Uh, yeah.

I likewise enjoyed Luin. That was really great event and the time we spent together was amazing. Um, I think it was really just the beginning of the development period of, you know, model API and modernizing all that stuff. And, uh, at the same time, certainly all the AI requirements came in as well, which I have to say now with hindsight, it was just.

A fantastic coincidence because, you know, we came up with the idea of modular API to be a generic link between complex configuration in the backend and users in front of a browser, needing great tools to actually manage them and configure them, and only justifying that for ECA. Was not quite enough for me as a maintainer and developer.

Of course it was because it really helped us to decouple things and now I have a even better architecture of all the tools and I'm really looking forward to maintaining them in the future 'cause it, I can be much more focused and all the other people as well, they can join and help us in the area where they are experienced.

But then also having AI with various requirements where they have to configure complex configuration as well, uh, makes it even more obvious that something like modeler API makes sense and. I'm looking forward to hopefully one of the next steps where other teams come up with alternative interfaces to all that.

So everybody who is not deeply involved in the technology looking at ECA, they believe that what they see in the canvas in the UI that's easier. But it's not, you know, this is just one interface on top of ECA. My dream is actually that we can replace that, not because I want to get rid of BPMM, but because I would like to demonstrate that.

It's not tightly integrated, you know, uh, the fact that I love to use BPN doesn't mean that everybody else has to love that as well. So giving people a choice and using a different one, like what's mentioned a lot is N eight N, so N number eight, and N, which is a diagramming tool outside of Drupal. A lot of people get in touch with to automate things, to orchestrate things.

Obviously outside of Drupal, um, or React Flow is another one that has really nice interfaces or lets you build nice interfaces, I should say. Um, and I know there are people working on integrations of those for Drupal, um, and I'm mentoring them a bit. And I said, look, I would really love that your development comes to something that we can start.

Testing using, and if it's really great, we are probably going to retire BPMN. If people don't like it, that developer say, oh no, no, no, no. Please, please don't. Don't get rid of BPMN because I'm using that for all sorts of things. Like, you know, even if it's not a real model that does something in Drupal, I just use it to discuss with my clients how the workflow should be.

So people do really like it, but sometimes they would like to use others and model API really lets us do that hopefully soon.

Martin: So you mentioned N eight N. Um, I do have one quick question, which is it, it seems like it's a Nuer nym, right? So the eight probably is short form for eight letters that should go in the middle, but I haven't actually been able to see anywhere what, what that's a condensed version of.

Do you know what that is?

Juergen: No, I'm just saying to say I have no idea.

Martin: Well, maybe, uh, maybe one of our listeners can, can chime in with that in, uh, the Talking Dral channel. But, um, I think N eight N is interesting specifically because I feel like in that, that much broader conversation around ai, there's a lot of excitement about N eight N is kind of that the same sort of.

Idea of a visual modeling tool. And so I think being able to, to allow people to use the identical tool within Drupal potentially is another way of sort of easing the transition for anyone who is interested in building AI applications, but using a platform that has its own web interface so that you're not having to sort of make that another sort of thing that you have to stitch out to in terms of actually creating the, the interface or the interaction layer.

Um, are there any sort of, I guess, functional gaps with VPMN today that you think using some of these other interface tools might help to address?

Juergen: I don't think so, but before I get into that, uh, one note about NA then, which makes it a little bit problematic as an example tool for exactly that exercise.

One of which is a N wants the world to believe that it's an open source tool, but it's not. And it's a little bit delicate because you know, the fact that they try and don't really provide open source makes it a bit suspicious. And I just heard from an agency here in Europe. Yesterday that they wanted to build an application that incorporated NAN as a standalone platform, but part of the overall solution, and they just realized that they are not even allowed to do that.

So NAN license agreement doesn't allow you to embed NAN in any way into other applications. So that just as a bit of a warning. Um, it's difficult to compare external orchestration tools like NAN or similar ones to what BPMN does inside of Drupal. Which is because inside of Drupal we only have access to Drupal features and we can only orchestrate stuff which is inside of Drupal.

So very specific. All the other tools, they are. Standalone tools that integrate themselves with other platforms. So if you use NA, then you can say, okay, I would like to utilize LLMs, um, other AI tools. And then they integrate with their APIs and they integrate probably with CMSs, with CRMs, database systems and so on.

So. Their purpose is only orchestration and they have a much wider scope where they can incorporate many more tools, which is currently not possible inside of Drupal with ECA and BPMN or other things. However, is that a limitation that. Is implied or is it just one that we haven't resolved yet? And I would want to believe the latter.

So now we have tools in place that allows us to build a platform that could easily compete with N eight and or others, like orchestrating stuff that's outside Drupal. So why not? We just have to build plugins, integrations, and other things that allow us to access them from inside Drupal. But that's not rocket science.

That is more like we just haven't done it yet. The question is, if I say we haven't done it yet, who is we? Uh, I would really like to see that we all come together in the community and start thinking about how can we now leverage what we already have and build applications or platforms that really allow us or the users to actually do exactly that.

Martin: So again, taking what you've just said and, and sort of bringing that back a bit closer to sort of the topic of ai, one of the, the hot topics of late has been this whole idea of MCP. Could, could that idea of MCP as sort of an interoperability standard potentially ease the integration for Drupal? With some of these other systems in terms of providing a more common interchange format?

Juergen: Most certainly. Yes. Uh, I think that's exactly right. Uh, it's technically probably not perfect or even correct, but my personal view is a bit like MCP now kicks off something that we as an IT industry. As a whole could have had a decade ago. You know, there have been ideas that service should have APIs that talk to each other and everybody specializes and then just ping the other service and receives something back, and that never really materialized to an extent where it was widely used.

Suddenly with MCP, that seems to be happening where everybody is happy that their resources get utilized by remote services. Um, and, and I think this is now happening with the pace, which is impressive because you don't have to implement individual APIs anymore. You just need to know how to talk to an MCP server and that makes it much, much easier.

Martin: Yeah, it's a really, really interesting time. And you know, I feel like for me personally, the whole idea of ECA as a very visual way to to customize the functionality of your site has been incredibly powerful. And now potentially. That that's abstracted out, being able to use a similar path, sort of, you know, build AI agents and, and some of those things is super interesting.

In terms of the, the AI initiative, um, are there any discussions going on as part of that initiative that are, that you think. Maybe, maybe not, you know, next week, next month, maybe not even this year, but conversations that are happening there that, that you think are gonna be really impactful for sort of Drupal and its future down the road.

Juergen: I haven't been involved in the initiative on the discussion level when they come up with strategy and, uh, roadmap. Or even weekly planning. Uh, so there is a great team in place, uh, that really directs us as a community where we want to go. And my role was more like in the background, taking on tasks and implementing things, but following the initiative, I am really very excited a, that we have one.

Because you know, it's probably unique. We as a community have a coordinated effort in building something, but it's not just quickly putting things together that show something today, but we are building something which is sustainable and can be maintained hopefully over a long period of time. I'm very hopeful about that.

Um. What I'm really, uh, excited about what the initiative is bringing us is the fact that lots of resources around the world are collaborating in a way that I would like to compare with what we've done 12 months ago with Drupal CMS, where even if the result is not yet a product that goes mainstream.

There are bits and pieces missing, but to me personally, the experience that globally, the whole drupal com or lots of people in the Drupal community collaborated and did something in a very short period of time instead of every agency doing it. Their own thing was the most exciting thing to me and. I hope that we have learned some lessons from that exercise and then repeat that in the AI initiative.

And I see that happening at least to the extent that if you look at all the agencies that are involved, not only the six that are funding it, but also lots of others that provide resources to make things happen. Is to me, outstanding. And I have to say, I'm really proud that the trooper community is able to organize things in that way.

Martin: I a hundred percent agree with everything you just said. The other, I think point that's, or nuance maybe to weave into that is that. Well, Drupal CMS, I feel like was kind of a, you know, from the ground up build, and it was amazing what was achieved in, as you say, a very short period of time. One of the things that I think is really interesting about what's already happened with AI is the fact that it actually started off as a very sort of fractious picture in terms of having lots of different people who are working individually on their own little projects.

Many of which overlap, many of which we're doing similar things, but in different ways. The fact that, that we were able to sort of, um, go beyond those, those individual, uh, projects and, and some of the, you know, individual ideas, but really come together and say, well, we're all gonna put those individual projects aside and really focus our efforts on one common thing.

I think it's is probably something that even in the open source world, not every community would be willing to do. I mean, when you think about. Places where you have maybe paid plugins. The idea of saying, I'm gonna put down something that maybe actually is a source of revenue to focus on something else where maybe I'm only taking a small cut of something that that could be bigger, but at the end of the day, you don't know.

You know, if revenue is, is a part of what's driving your will to contribute, then there may be some hesitation to do that. And I think, you know. Mm-hmm. Yeah. The AI initiative and how it's already come together is also a testament to, I think, the value of sort of true open source and, and really making the focus on collaboration and driving the best solution as opposed to sort of allowing people to, to maintain these kind of individual fiefdoms as well.

Juergen: Absolutely. You're making very good point. I wonder. Is that probably possible in the AI context where the whole technology seems to be like a lot is happening, but business models haven't materialized yet, so nobody really knows if and how we are going to make money out of all those tools Eventually. We seem to be convinced that we need to do something because if we are not part of the right, we are going to lose anyway.

Martin: Yeah, that's a really interesting question. I mean, so I've, I've actually recently been going through a book called Empire of ai, which is sort of this fascinating history of. It's focused mostly on open AI and, and how it sort of came to be, you know, the, the brief involvement with, with Ellen Musk and then now him going off and working on competing technologies.

But, but one of the things that's really interesting in there is the fact that there's, there's so much hardware and resources required to train models that I feel like that's the part that's really sort of driving the need for. These individual companies to, to be going out and creating revenue models and in some cases has, has focused or forced them to shift to being more proprietary than I think they wanted to be at the outset, because at the outset they had this dream, particularly with OpenAI, it doesn't mean it's literally in the name of this being sort of a place to, to avoid.

The possibility of advanced AI being something that was only in the hands of proprietary interest, and yet because there's this need to, to have, you know, and really billions of dollars of resources to train these AI models that's really focused or forced them to to shift. How they interact with bringing some of these things to market and, and trying to do it more through, you know, paid plans and some of these other things.

The other thing that I think is, is interesting is more recently I've heard the argument that because it's becoming so expensive to train new AI models, that actually to some degree gives an advantage to companies that, that don't have to. Fund that directly as sort of their, their primary business model.

So when you think of like a Microsoft or a Google that have a much larger, you know, stream of revenue and source of capital and can invest in these things without having to see an immediate return to sort of keep funding new models, it'll be interesting to see how that, all of that, you know, and the need for capital and the need for, you know, the kind of the arms race of always wanting to train the, the most advanced models.

How does that impact, you know, as you say, the, the sort of what was intended at one point to be a very open source approach to, to creating these and, and sharing them internationally?

Juergen: Yeah, that's right. Whereas I am very optimistic when it comes to the, uh, potential of optimizing on an energy level or generally on a resource level. And it's funny that you were just now focusing so much on the training side of things, which is most likely something that only just the big players can actually do.

But I have been a little bit disappointed really on the execution side of things. If you watch an LLM actually do something, so it's already trained, but you actually utilize it. The level of inefficiency is just amazing. You know, I, I have built a very simple tool with ECA, like just to show that ECA could be a tool that an AI agent can leverage, and I wanted to see that it works and how it works, and if the expected result comes out of it, that there is the AI agents explorer.

You can say, look, take that agent here is my prompt. And then you can watch what it's doing, so what tools it's calling, and what question you ask and what's the input, what's the output, and so on. And it takes a little while, maybe up to a minute, until it responds with the information whether a given user is the author of a given article.

And you think, yeah, okay, a minute with everything that went into it, that's okay. But you would expect, well, it took five or six steps before it came to the solution, and my expectation would've been if I ask a similar question again, it would just have learned something from that exercise and be more efficient the second time.

But it isn't. It just goes through the same path over and over and over again. So having that in mind and now maybe hundreds of millions of users doing that every day, things like that, and more complex stuff. Of course, if we get those LMS to learn from their own experience and shortcut in the future. I think that alone should come with massive savings.

Martin: I totally agree. I think if we can get to the point where the LLMs, as you say, can, can really learn on the go as opposed to having to sort of, um, do a one-time bulk training, I think that that has the potential to significantly reduce the environmental impact of always having to go through this massive exercise.

Of kind of, um, relearning all available knowledge and trying to synthesize that. I'm not sure if, I mean, maybe there are technical limitations in terms of, of how it can understand some of those things, if it's always building on an older way of understanding as opposed to to relearning, you know, that's maybe for smarter people than me to comment on.

But, um, but I agree with you. I think that feels like from, you know, and even from a, as you say, a speed standpoint. If it can sort of, um, once it's learned how to do something, internalize that as sort of a heuristic that it can use to, to more quickly do the same thing in future. That, that definitely seems like it could be huge potential.

So, you've talked a little bit about ECA, uh, one of the things that you mentioned in your articles is that the, the new major version of ECA, that, that was just released. Is really focused mostly on the modeler API, also some sort of deeper level optimizations. Now that you've abstracted out certain parts of it, I wonder if you think a little further out about ECA.

Are there sort of roadmap items in terms of new things that you'd like either ECI directly to be able to do or maybe, um, you know, as part of the ecosystem, new challenges that you want people to be able to, to solve by using ECA?

Juergen: Yeah, that's, that's right. ECA 3.0 is feature wise, almost identical with the latest 2.1 release.

Um, except for internally and we've cleaned up the, uh, code structure. We have implemented object oriented talks. Uh, and lots of other things, uh, using attributes and we thought that, you know, backwards compatibility layers, we don't necessarily have to build in because, you know, the features are the same anyway.

So ECA three is the tool for the future modern code base, and so on and so forth. So, moving on from here. Uh, gets me back to what we discussed before with NAN. Like, I would now like to focus on features that help us to become competitive with those external platforms, um, in, in two ways. Actually, I would like to see from within Drupal that we can hook into external systems like NAN does, but also I would like to make.

ECA tools available to those external orchestration platforms so that they can leverage a group of features more easily. And I've just looked into some of those potential integrations, uh, a couple of days ago, and the difference of a Drupal site. Compared to some other platforms that you would like to integrate into an orchestration platform is the fact that each site is unique.

Now, how can I make tools of your website available to N eight N and without knowing how exactly your dupy is being composed, like which ECA models do you have? Do you have AI enabled? And if so, which agents are available on your Drupal site? So it feels currently a bit like it's not about integrating Drupal into an orchestration platform.

It's about integrating your Drupal site into that. And, and that's not the way to go because we can't do that for every single site. So this is what I'm currently discussing with a few people and on, on how we could. Go about that and we have some ideas and that is, uh, a huge focus on stuff that we are going to do next.

Um, but I would also like to, um, improve the custom event facility that we have in ECA. If we make that a bit more generic, it could be the cornerstone. To be able to build independent tools that we can probably deploy with recipes so that we can say, look, here is a recipe that does certain things, like, for example, the event platform that you are working for, that there is a package that lets you call it from other pieces that you have implemented elsewhere.

And you know exactly what you have to, uh, provide as input and what the output will be. Uh, there is quite a bit what we can improve, uh, on the ECA side for that, uh, and make it really generic. And the nice thing about that, if we build things that way. We can reuse them in lots of places. Like if we have an ECA model that works standalone, it can as well be utilized by AI agents as a tool.

But it can also be accessible through, uh, Drupal's MCP module. So then even external tools can already leverage that. Without having to build extra integrations outside. So that's very exciting to me. And then on top of that, and that is my biggest challenge really. I would like to grow the, uh, the team of developers that are actually actively working for ECA.

Now it's better than ever before because with the new architecture, we have something where even somebody who has no idea about BPMN or modeler API or anything. Now building features in ECA is something that somebody can do who knows how a plugin works, and they don't need to know lots of other things around it.

So I would like to grow the community in in that area. Uh, to have a stronger team actually moving forward.

Martin: Yeah, no, I, I mean, I think the work that you do on ECA both directly on the code is amazing, but also, um, you know, one of the things that, that for years has impressed me about Drupal is like I can go in as a user, as maybe somebody just starting to use ECA, go into the ECA channel of Drupal Slack.

There you are, day to day answering questions, help, helping people with very basic questions, but providing such immense value, as I say, not just in terms of the code that you create, but also you know, the support and other ways that, that you, uh, help people who are getting up to speed in ECA. You see the same thing with search API and lots of other places in the ecosystem.

Um, I will say, because you mentioned event platform, that for me. Actually, one of the things that that has been really, I think, valuable about the, the work that I've been doing on event platform is actually that it's given me a reason to, to really get up to speed on ECA and really understand sort of the inner workings and, and how it can map to, to different use cases.

And all of that has been really valuable. But in the same way that we talk about the AI assistant within Drupal CMS as something that potentially. Will make the complexity of Drupal easier to, to access and build on for people who, who don't already know Drupal, I wonder is there is, are we close in any way to having kind of a similar, um, chat bot agent that can, can create at least the first draft of an ECA model based on a conversation about what you need your site to do.

Juergen: Well, the answer is we are close. But actually I have seen it in action already, uh, which was early this year, maybe six months ago. Uh, Jasper Lemons from Drop Solid. He has developed something, well, an AI agent where you can say, look, I would like you to build me an ECA model that. Starts whenever there is a new blog post and then there is a comment to that and I would like to go through steps and you explain what you need and it comes back with an EC, a config entity, which is just yammel.

That's pretty easy. And I have seen that producing ECA entities that actually work. But the beauty is once you have that starting point, you can then go in and. Fine tune, modify, learn from it, and so on. My initial concern when Jasper came up with that idea, I said, look, I can imagine that an AI agent could build such an entity, but if we want to continue modifying that in the ui, like for example with BPMN.

We need much more than just the ECA entity. We need all the data, which is behind BPMN. And unfortunately, that's XMLA pretty, um, old data format and not very famous either. So how do we, well. The other way around. If you build an E-C-A-E-C-A model manually, you start in the canvas, you draw your diagram, and when you hit save, we just extract the pieces that ECA requires.

Install that separately in an ECA entity. Now, in this case, it's just the other way round. You start with what ECA needs, which is the absolute minimum of information to actually process that. But all the other pieces, like how they are arranged, how they look like, what the labels are, and so on and so forth, that is just not existing at that moment in time.

And I said if, if it were possible to do some sort of upstream conversion. That would be required. But to be honest, I didn't believe that would be reasonably possible because I thought it's so much work. But he didn't give up and he did some research and he found an API, which comes with the BPMN IO library, and he actually managed to do the upstream conversion.

And not only that, he also found, uh, a plugin to that which then allows for auto layout, like automatically arranging the components in the canvas. And from there on, it just was possible to do exactly that. And we now benefit in the new BPMM 3.0 release from those building blocks that he came up with.

Because now we can have an option to say, I don't want to save the heavy XML data in my Cong entity. Just discard that and only store the absolute minimum, which is the ECA younger data, because we can reopen that. And upstream convert that into the BPMN format again, but only when needed. I do remember the days when we leveraged ECA for Drupal CMS.

People were frightened to death when they saw that it's some of the config entities. There was some XML data in there. And I was so happy when I was able to tell them, look, yeah, we can, we can discard that. We can get rid of it. Um, so yeah, it is possible. It already exists so people can try. There is an ECA AI integration submodule, which comes with the main AI module, but Jasper and I, we are currently in the process of extracting that and make it a standalone.

Uh, module so that the AI inter, what's the name of the module? AI Integration eca. This is where that tool lives, and we are going to release that as soon as the, uh, refactoring has been completed. It's impressive, I have to say.

Martin: I think that's, that's really exciting and, uh, definitely something that I'm gonna have to try out.

Yeah, I think that the potential there is super exciting. Um, now we had spoken at the beginning about, uh, Luen and one of the things that was interesting there was, as it turns out, we were staying at the same hotel, uh, which neither of us realized except when you opened up your laptop to connect with the wifi.

You saw my Drupal org username in there. And, um, so one of the, I get the impression that both you and I do quite a bit of traveling. For Drupal specifically, but just in general for, you know, work and so on. Um, I thought maybe we, it'd be fun to, to talk a little bit about travel hacks and what are some of the things when you're on the road that you, you use to sort of make that experience sort of less disruptive, you know, both for work and personal life in terms of, you know, just making it a better experience on the road.

Juergen: It's interesting you say that. My impression is you are traveling 10 times more than I am. Well, I used to travel 50% of my time decades ago. Um, I reduced that a lot and I'm very selective nowadays. So it's more or less like two Drupal cons and deaf days. Um, but if I can afford, that's the most I would like to do.

Uh, but those three I don't want to miss because it's, those weeks are always highlights in the year and when, in the past when I was really on the road almost all the time, it was like I can't get enough into one trip. Whereas today it's exactly the opposite. So when I'm traveling and I'm planning for it, there are actually two things that I try to accomplish.

The first one is travel light. So don't take too much with you, it's just you and yourself on the road. And the second one is, don't get in a hurry. Take your time. With those two, um, site or building blocks kind of thing, I feel much more relaxed while on the road because, you know, there is not much that can go wrong and that helps me to actually focus on the event, on why I'm traveling.

Uh, and that's what I learned from the past and that makes it much easier today. But how about you?

Martin: Well, so the travel router is one that, um, I had read online. Somebody suggested that, and I have to say. That for me has, has actually made it much easier. You know, because I always have at least a laptop and a phone.

Sometimes I'll bring a tablet, uh, on some trips. I'm fortunate that my wife can join me, and so rather than having to get all of those individual devices connected to like a hotel wifi, sometimes they won't even let you. They'll have sort of a cap on how many devices you can connect, and so using a travel router is actually super beneficial for those use cases.

The other thing is. Many of them will have some internal way to connect to A VPN, and so to be able to watch, you know, the YouTube feed it from home or some of these other things can also sort of make it a smoother transition in terms of just feeling like, okay, well, you know. The hotel has cable, but you know, 90% of the content is in a foreign language.

And so, you know, you're sort of hunting around for something that that is, um, you know, entertaining for, to pass the time. Although, I will say when I travel, it seems rare more that I watch TV anyway. But, um, the other thing I, I mean, I've found that, that having, you know, as an example, I have sort of a small toiletries case that is sort of always packed and ready to go.

Um, same thing with like, I've got a, you know, small chargement cable, uh, bag. I completely agree with you about that idea of really trying to pack light. And one of the things that I've, I've tried to develop some discipline around is when I unpack from a trip to try and pay attention for what are the things that I never actually used.

Yeah. So that you can take those outta your bag and as you say, pack later the next time, because the odds that you'll need this, you know. Rare thing. Yeah. You can probably get, get by without or the other, the other thing I've, I've heard people say, for example, if you're traveling to, maybe you live in a warm climate and you're traveling to someplace cold, rather than, than trying to find things locally that are, are gonna be suitable for that climate is maybe leave some extra space in your bag and then just get something while you're there.

Yeah, because they're, they're gonna have, you know, probably better quality, find it a cheaper price, some of those kinds of things as well. So, uh, definitely those. I mean, the other thing that I've been trying to play around with a little bit over the past year is also 3D printing things that can help with travel.

So, as an example, if, if I. Have to fly economy, which, you know, realistically I do more often than I would like because, you know, airline prices being what they are. Oftentimes the um, the distance between my seat and the seat in front of me is, you know, not what I would like it to be. And even putting something like a water bottle.

Yeah, in that, that front pouch will push it out to the point where it's sort of pushing into my knees now. So I've got this like little 3D printed. Uh, I think the model was designed to be like a shower hook. So if you've got like a glass shower enclosure, it'll fit on there, but it actually will fit on that pouch.

And so if you have a water bottle with a hook, you can just hang it in there and it's not pushing the whole pouch into, so there's just like little things like that that I find, you know. Having some of those, you know, small things in a way can make a big difference over the course of a trip as well.

Juergen: That's nice. Yeah. So you have your own 3D printed to do that stuff?

Martin: Yeah, yeah. About, well, I think almost a year ago I had been reading about some of the newer Bamboo Labs printers, and I got one, I think it was the P one s, and I have to say it's really been kind of transformational. I had bought a 3D printer, actually, literally at the start of COVID, unknowingly.

And it turned out to be a good hobby in the sense of, you know, when you're spending lots of time at home, this is a good thing. You just, you know, get a filament dropped off at your door and then you can sort of keep going on that. But the, the one that I had was very technical, required lots of troubleshooting, whereas the.

The Bamboo Labs is, is really cool because it has kind of this ecosystem, it's called, it's almost like an Apple type system where you've got an app on your phone, so you can just sit there on your couch and say, oh, I'd like one of these. Find a model that's already in their store and then have it print one off, and then, you know, whatever.

30 minutes or an hour in some cases, maybe multiple hours later, you can just go see how it's doing. There's also like a built-in camera so you can see exactly how it's progressing. Some of them have built in sort of what they call it, spaghetti detection. So if it goes, uh, completely, you know, um, a stray from from the printing, it'll stop automatically.

But, um, yeah, for me it's, it's been a lot of fun and, um. Have even been able to sort of contribute back a couple models. So it's kind of got its own little open source ecosystem there as well, which has been pretty fun.

Juergen: Nice. Yeah, I fear I would just lose too much time.

Sounds,

Martin: yeah. Yeah, it's definitely been, um. Uh, as I say to me, the, the part that I really like is doing things that are sort of practical so, you know, can make my life better in small ways. I think everybody, when they first get a a 3D printer, does prints out all kinds of fun, decorative things, and then at certain point kind of goes, all right, well, I have all the, you know, Darth Vader mass or whatever that, that I need for a while.

But, um, but to me the, um, yeah, that ongoing practice of figuring out where, where you can sort of, um, do things that, that make life even just a little better. Um, that, that is, is pretty fun. And actually even, uh, modeling my own things using like 3D software and stuff is, is pretty interesting too. So

Juergen: Nice.

But actually, what I also noticed only recently is that you started posting lovely pictures from all over the world. Like it looked like you have been in a garden in Tokyo, and that you have been in San Francisco. But then I realized that cannot be on the same trip that was probably captured in the past.

Is that something you are, you are into?

Martin: Yeah, absolutely. So photography is definitely one of the other big hobbies for me. And of course, you know, doing as much Drupal travel as I do, I, I will often now bring my, I have a, like a Sony uh, micro four thirds. Camera that'll bring with me for, you know, new places.

And, um, so sometimes like on a blue sky, they do like a weekly photo channel challenge. And so I'll go back through like, you know, there were a couple of times where there were shots from Barcelona that were, uh, fun to post and some of those kinds of things. So yeah, you're, you're probably seeing from the backlog of, of different places.

Okay. And not, not necessarily. Recently there was one I, the garden, one that you saw was actually probably from Asheville. From the Sunday hike there.

Juergen: Yeah.

Martin: But, um, but yeah, definitely the photography passion sort of, uh, dovetails nicely with, with Drupal and specifically the travel. So,

Juergen: yeah. Well, if you can make the time to actually look at other things while you are in all nice places where we are traveling to all the time.

I remember in the past, all my friends were al always saying like, why it's so nice? What areas of that world you're going to visit? And I say, yeah, well, other than airports, hotels and offices, there's not very much you can see because there's no time for it. But you seem to be able to just take some time out of that to take something with you that lasts longer than just a memory.

Martin: Yeah, I do feel like I've, I've gotten better about trying to at least book, you know, a day or two before or after to, to actually spend some time exploring, I think. The first Drupal Con Europe that I was able to go to was in Prague, and I think I had basically one day, and so often at night after the conference I'd be, you know, out exploring.

And I think the last night I got like two hours sleep before my flight because I was like, I don't know if I'll ever be in Prague again. So it's crazy to like, I can sleep anytime. I could sleep on the plane, although not usually, but you know, like to have the opportunities to, to go around such an amazing city like.

That has to be the, the priority there. But, um, but yeah, if you, if you have the ability to book a little time so that you can do it without sacrificing sleep, then, then even better for sure.

Juergen: Nice. Yeah. You seem to be

Martin: doing well. I mean, I have to say Drupal has been very good to me and so I'm always happy when I, I get a, a chance to give back and, and sort of help along, hopefully that next generation of.

Of Drupal who will, you know, infuse it with new ideas and, and help it to achieve new things. So, um, I get the impression you're doing well too. I mean, it was great to see freely give as, um, you know, an organization that's done so much for the Drupal community around AI and, and a few other things. Um, but, uh, you know, great to see them funding you to, to do some of the work for the modeler API and some of those things as well.

Juergen: Yeah. I'm very happy for that. Absolutely. Yeah, it really helps. Focusing on stuff you want to get done. If you know there is somebody who is actually paying for it without you always having to reach out and finding somebody. So yeah, really great of them doing that.

Martin: Well, you're gonna, it's always a treat when we get to catch up. Uh, would love to do this again sometime, but, uh, I'm sure we'll get a chance to sit down in Vienna and, and we can catch up more than too.

Juergen: Oh yeah, certainly.