navarre.aiClarisTalk AI · the FileMaker AI podcast
EngageU Live: Matt talks with Joris Aarts and John Renfrew
Episode 21 Joris Aarts and Johan Hedman are the two brilliant minds behind the EngageU conference in Malmö Sweden. This year they had over 200 attendees, and reached many other milestones. Matt sat down with Joris and John Renfrew to talk about their sessions and the conference. Joris did a session on the exciting things coming soon in Claris Studio, and John presented on OData and why it's a grea
Apple PodcastsSpotifyYouTubeAll episodes
Transcript
Transcribed by machine (whisper, large model), 2,357 words. Names and product terms are sometimes misheard; the audio is the record.
Claris Talk AI Claris Talk AI and the ETS features. You're showing some ETS features. I'm going to summarize a little bit because I was really excited. So there's that you're showing some ETS features. There's features of studio that are the things I've wanted since the beginning. Yes. Yes. Custom forms that you can have different segments that you can configure any way you want. I use Wix for my website design. It's very much like that. Wix kind of feels like FileMaker in a lot of ways where you can just grab elements out, do it. It automatically resizes. You can pick themes and colors. And then the really cool thing that you were showing is how you can use calculations to get data from different places and do things that much more quickly than you could do with FileMaker.
Yes, it is. And so this is exactly the excitement I wanted to share. I'm involved with the kitchen since early, early this year. And so I shared his first time in Berlin permission from Yes, because then was still still early. But I saw the excitement there. And that's what I also was like, let me let me do this demo again. And just watch how people react. This is what finally people are looking for. Yeah, it's going to be in production in studio version is so limited that people feel like math is not for me. Yeah, and we can totally I think it's going to be a very interesting and soon useful extension of file. Yeah, yeah, I think what you showed is pretty, pretty strong. Yeah. But I also think you showed a lot of areas where it kind of needs work, like a little calculation thing and kind of the small windows. It kind of reminded me a little bit of like FileMaker earlier days. Yes. When you get a little teeny window to select your stuff and go, Oh, yeah, actually, now there needs to be a window for full screen. So you can select from the thousands of, you know, this kind of things from the kitchen, my experience is telling to us, like two weeks later, you see already things improving there. It can do things pretty fast. They can move pretty fast. It's to have like the kind of liberty there that you don't feel in FileMaker decorations. That's absolutely true. And because they can push it up there and the updates just come out with really good release notes that are actually also really fun to read. Yeah, yeah, yeah. These release notes are really, yeah.
Yeah, it has a sense of humor. Like that. Yeah. Yeah. And then John, you've been preaching about OData, which you've been talking about for years, apparently. Three or four years. It's my favorite. If I was on Mastermind, it would be my specialist subject or so. So, and I don't know a ton about it. I haven't really had a chance to use it yet, but it's been really the buzz in the last couple of conferences about its ease of integration and its really, really fast speed and utility for a lot of other things. Give me your little spiel. So what we did here was obviously in the brand new release of FileMaker Server, which has just come out. Literally a couple of weeks ago. Yeah, there are 21.1, 21.1.
So this was the first opportunity for people to be at a conference where I could say, here's all the new stuff we can just demo that it works now, go away and play with it. So it's like a fast track to, oh, here's what I need to look at. So it wasn't a how to use OData or anything like that, but the big driver now is having announced that there is a future for OData in the platform and a change in how you can query the basic server. Also, internally, some stuff to do with Studio is that there's going to be some degree of connection between FileMaker and your studio using OData. So because Claris had worked out there's a use for it, then it will get lots of love and attention and other things.
I think there's something sort of in common with this, right? These are two areas that are getting love from FileMaker. This and the amazing chart feature that's built in that has had some, oh wait, that hasn't had any. Oh, okay. So the point about OData is it, so the first big point is actually not what we can do in FileMaker. Anybody who knows OData can connect to our FileMaker server. So if you have a specialist in your business in Power BI that might have been connecting previously to SQL, they could just connect to your FileMaker data. I'm coming for the job, yeah. We're recording video, so don't let them from me. So, you know, so that's like, it's not just about what we can do.
All of our heads are cut off. Damn. I'm not sure when that started. That's better. Yeah. So you query the data through an API call, effectively. And there are some common elements between ODBC and SQL in the way that you write the queries. But what you get back is some JSON. And the advantage is you can query any file on any server as long as you have the right username, password, and permissions. Love that. So there are really interesting use cases for both reading and writing with it. What are the limitations in the payload for the JSON? So there is a limit of 10,000 records is the maximum number of records you can query in one go. You mean get the results of, you mean?
Yeah. So you can query millions, I'm presuming. You can query millions of records. You only get back 10,000 at a time. But like any other system that's paged, you get a link to here's how to get the next set of records. That also is more than you'd want for most anything, anyway. So in the main, the systems I write, I'm working with one or two records max at a time. There are other people who are doing different use cases for it. So querying data, what about like querying a script? You can run a script. The difference is that when you're querying the data, you're just talking directly to the table of currencies and saying, can I have this and this? To run a script, it's more like a piece of session.
Because a script has to have a context. It's a layout, so you're into what data API will do when it comes to script and PSOS. So there are some performance downsides to doing that, but there are some advantages to things you can do by separating your systems. So I don't have to have my system full of external data links just so I can do a thing in another file. I can trigger a script in a third-party file just from OData. So as I understand it, correct me, but authentication is also one of the things that's simpler. Yes. And you send authentication credentials, and then from then on, you don't have to resend it? No, you send the credentials every time, but the credentials are basic.
So it's the base64 of the username and the password string. So it looks like a token. You're not actually sending the real username and password. You're sending the base64 string. But you just include that as a header in every call. So unlike the data API, where you have to get a token, restore the token, reuse the token. And here, just every single call. Symbol of credentials. Symbol of credentials. Like an API call. It is exactly an API call to your file maker server. What are the AI implications? I'm not ignoring you. No, no, no. The AI implications, because I'm actually dying to know about that for Studio as well. So would you think that you would query the data and then run AI on the stuff that you get?
Or would you call something and have the AI stuff running on the server and get the results of it? So the issue here is it's a tool. So the problem is... You have a new tool. Every problem looks... You know, if you only have a hammer, every problem looks like a nail. None of us experience that in the file maker world, though. No, never. But obviously the point there is... Well, the answer to that is... Well, both of those would be interesting use cases. The question is, what are you trying to do? Where are you getting the data from? What do you want to do with the data that you've brought back? So maybe you're running a script to run an AI query before you get the answer back.
Which obviously has a time delay. Or bring data back, run a query on it. Do something else with it. So the issue is, here's a tool which enables you to do things. The question now is, oh, so what are the use cases that you could think of for working with this as a transport mechanism? Not purely as the thing. Because you always... Whenever you get any answer back from a web service, you have some JSON. You have to decide, what do I do with the answer I've been given? Where do I put the data? What do I do with it next? So the answer is, yeah. What do I do with it above? Yeah, okay. So it's not like an obvious best practice. It's like... Not yet. Whatever way. Yeah, okay.
Because it's all new and people are just beginning to get on it. I speak to conferences and 60% of people have never even touched data. A couple of sessions I've seen about it, or demos I've seen, didn't make it look like the hello world was at all easy. Would you say that there's like a sample file that can make it ridiculously easy? It's like, oh, call these three script steps and poof, you've got data to start? Yes, but it's not documented well anywhere. Oh, okay. There aren't lots of blog posts. Maybe that's time for me to start writing some of those. I would say yes. Would you say yes? Well, you know, I think like demo files and videos on YouTube. Yeah.
Because that makes it, for me, that makes it really, really easy. And then also, of course, I use ChashiPT a lot. Just like, how do I get this? Because it's obviously a lot of text and query. And then to figure out what's going on with it and correct it and make it look good. So without wishing to just be a self-publicist, I have just done five sessions on Richard Carlton's training. And I've done a lot of training on his training TV about OData as a proper, for people who've never touched it before, up to here's some things you can do. Okay. Obviously, clearly, five hours is a lot of material. Yeah. But it is in depth, I hope, to help people on the journey.
So there are things in there that are proper examples you could replicate. Yeah, but you could get started in a half an hour. Yeah, absolutely. Okay. Let's talk about a third studio. Yes. Because it hasn't really come up. No, but there is the possibility. There is the promise of, we already can see that in the Clarice Studio Connector, we will have in the future, the following iteration, we will have the ability to create hubs and create views in the Clarice Connector. Oh, okay. And so that's something new for the Clarice Community, being able to generate schema and generate layouts and UI on the fly. Okay. And that's what I feel the urge today. Because I think the point in Clarice Connector is...
Almost over. Yeah. We'll go to a new era where we're just talking about the build with this, and then we start pointing to that. And that's what I see. And in the studio, I totally believe that this will happen. Yeah. With FileMaker, not so much. I don't know. I think actually there's a lot of things with FileMaker that's coming, too. Maybe not so much in, like, coding and FileMaker. There are AI features in FileMaker. Yeah. Building a FileMaker solution using FileMaker, I don't believe it's possible because FileMaker is... You know what I mean? My knowledge... You know what I mean? You know what I mean? FileMaker still has no way of generating UI elements dynamically.
Yeah. You have this XML, and you have this XML patches. Yeah. But my impression is that it's still a struggle, that it's still not able to reach any corner of the game where you can go. I guess my brain goes to, like, what can you do right now within it? And what calls can you make that actually do things that are really useful and cool? I saw Todd Geyer doing a demo a couple of weeks ago where he built a script. I asked you, Chad CPT. And so he teached Chad CPT how XML works. And then Chad CPT can more or less generate many things that XML, and then copy-paste that into something. Which is, yeah, that's not a theme line. Right. But then there's some other tools that I think are kind of...
We'll do another... We'll have another discussion about that. Actually, I think doing screen sharing is a really, really cool thing that I've seen really cool. Simtheory.af. That might bridge the gap for a lot of this stuff. It's in super early alpha right now. But I think it's actually going to have a lot of users. So cool. Gentlemen, thank you so much for your time. It was great talking to you.