Our guest today is Steven Pemberton, a renowned researcher in the field of computer science and information technology.
Pemberton is best known for his contributions to the development of the World Wide Web, particularly in the areas of markup languages and web standards.
In this podcast episode, Steven and Noz discuss the future of web development and content strategy, bringing to the table their extensive collective experience of 55 years in omnichannel strategies—30 years by Steven and 25 years by Noz. This discussion untangles the complexities of web technologies, emphasising the pivotal role of structured content and the innovative approaches shaping digital experiences.
“My fundamental point of view is that if it works, you should do it… My problems with AI is that firstly, there’s no I in there yet, that it’s machine learning but it’s not artificial intelligence. So that you can’t trust it.” – Steven Pemberton
By the end of the episode, you’ll learn:
- The Importance of Separating Content from Format: Understand why and how content should be distinguished from its presentation in the digital realm.
- Insights into Declarative Programming: Explore the benefits and applications of declarative programming languages, specifically XForms, in creating more adaptable and efficient web technologies.
- The Role of XML and JSON in Modern Web: Learn the evolution of markup languages and their impact on the development and management of web content.
- Challenges and Opportunities with Headless CMS Systems: Learn about the shift towards headless CMS systems, their advantages for content management, and the challenges they pose.
- Future-Proofing Digital Content: Discover strategies for designing content and interfaces that are adaptable across different platforms and technologies.
- AI in Software Development: Gain insights into the potential and limitations of using AI for speeding up software development and the critical perspective on AI’s reliability.
Also listen and download on…
Steven Pemberton is a renowned researcher in the field of computer science and information technology. He is currently affiliated with the Centrum Wiskunde & Informatica (CWI) in Amsterdam, where he has been actively involved in cutting-edge research and development for several years.
Pemberton is best known for his contributions to the development of the World Wide Web, particularly in the areas of markup languages and web standards. He is a co-author of several key web technologies, including HTML, XHTML, and CSS, and has been instrumental in driving the evolution of these standards over the years.
In addition to his work on the web, Pemberton has also made significant contributions to other areas of computer science, such as the design of programming languages, the development of software engineering methodologies, and the study of human-computer interaction.
Pemberton is a highly respected figure in the academic community, and has received numerous awards and accolades for his work.
As a speaker, Pemberton is known for his engaging and informative presentations, which draw on his deep knowledge of computer science and his passion for technology. His talks are always thought-provoking and entertaining, and he has been invited to speak at numerous conferences and events around the world.
Full session transcript
THIS IS AN AUTOMATED TRANSCRIPT
Hello, everybody. So welcome to the episode I’m here with Steven Pemberton. Steven is someone I’m very happy to have on the episode. We have we met at our my first Janus boy conference. In Denmark a little while ago, and I immediately hit it off as old dyed in the wool. XML unstructured content folks. Steven do want to say hi and introduce yourself a little bit.
Hello, yeah, I’m Steven Pemberton. I’m ethnically English but judicially Dutch I live in Amsterdam. The Netherlands. I’m a researcher at the CWA which is the Dutch National Research Institute for mathematics and computer science. When I went to university my tutor in England was was deep Grimsel who built the first transistorized computer. His tutor in itself was Alan Turing, which makes me a grand to Tiv of Alan Turing. And after university a quite Coincidentally, I went and worked in Turing’s old department working on the fifth computer in the line of computers. That Turing himself worked on before he died. He worked on on numbers one, two and three. Then I was a university lecturer for a while and then I came to Amsterdam as a researcher where I co designed the programming language that Python is based on. I was the first user of the open Internet in Europe, when the CWA set up the first internet node open Internet node in Europe in 1988. And I went on to get involved with the web right from the start, which means that I then co designed a lot of web technologies like HTML, CSS, X forms, RDFa and a number of others and I’m still working on a couple of internet based things. And I’m great I’m very happy to be here. Fantastic. Will leave space for an edit there? I just realized my mic is way off to the side here. Which is not where it should be. So no problem, you’re fine. So what we’ll do is we’ll read we’re repatched me and if your levels have changed too much, do I sound very different to you? No, no. Okay, great. So hopefully it’s not too bad. So take a breath and pick it up.
Noz Urbina /UC
Okay, fantastic. Yeah, that it just warms my heart. Hearing your bio. I don’t talk about too much of my, my older background in the beginning of my career on the podcast, but I was born into the world of structured content, XML RDF, straight out of university, so I was just a young and kind of discovered the web through that lens through a, what we now call semantic structured perspective. And I only kind of halfway through my early career did I discover that people were ever doing it another way, much to my shock and dismay and confusion? I found out that people were just throwing their content of these big buckets called Word documents and webpages and and chucking them into even bigger buckets, which they called Content Management Systems. And obviously, we’re suffering many problems, which we continue to suffer today. I think things that we’ve been talking about since the very beginning, is the simple concept and your listeners of the podcast. Many of you will be familiar with it, but I’ve got Stephen peppered in here. So I’m gonna bring it up, which is we’re still talking about the fundamental issue of separation of format, and content away it looks from what it actually is used still running across this and when you’re talking to people steal?
Well, absolutely. I mean, I mean, separation of concerns in general is, is a is is a concept that doesn’t seem to have sunk into a lot of heads. So I mean, not only content and presentation, which of course is absolutely essential, but but also content and data for instance. You can separate those out. So there are a number of things that that are beneficial to separate because then you you have more freedom to deploy over different different sorts of of media. But yeah, there’s still a lot of work to be done.
Noz Urbina /UC
So tell me about that content from data. That’s interesting, because so the data content and if I if I want to get out of I don’t know if I want to open this can of worms when we get into data versus content versus information versus knowledge, like there’s you know, especially in the in the you know, authors circle, we’re we’re, we’re having bitter fights about these terms all the time. So when you say separating content from data, what are you referring to what’s an example of where I would separate?
So So an example of a simple example that I use is, imagine you’re getting a contract from let’s say, a book publisher. Now, there’s, there’s one level of separation content from presentation in the sense that you know, there’s a whole big text and you know, whether that’s in one form, or another, or displayed in color, or whatever, that’s all presentation issue, because what you’re interested in is the actual content of the text. But there’s another way that the publisher for instance, would look at it as old. That’s a standard contract. But there are three pieces of data here that change the person’s name, the royalty percentage, and the title of the book. Yeah, and so for the publisher, the presentation includes all that text that we think of as content because the only content for the publisher are those three bits of data. So there are there are different levels. You can look at it going from the data down to what gets President presented on the screen.
Noz Urbina /UC
Thank you. So that’s a very interesting way of putting it. I’m going to talk I’m going to talk to that through the lens that I’ve been discussing it recently because we’re we’re discussing content models all the time. And what I have found when I’m, we’re looking at a lot of the new CMS systems, the headless CMS systems that are coming onto the market we’re doing content models for the headless CMS is a very popular topic with our listeners and in the world. The new breed of headless CMS, for better or for worse, seems to be finally bringing the idea of separation of content from presentation to the wider world. I’m not gonna say it’s mainstream yet, but it’s, it’s you know, as we’ve been talking about this, since since you were half my age. Since so, we’ve been talking about this for a better part. of half a lifetime. And now we’re starting to get more mainstream, some more mainstream success of just the idea of separating out content from presentation, as you said, tons of work to do, but what you just talked about, is when we talk in content modeling terms, the idea that not all content types are equal. And that there has to be a sort of perspective on the hierarchy, the sort of what I call the the concentric circles, circles, have a content. So you if you have for example, you just gave a perfect example like a contract, where we’ve got certain clauses, which are standard standard clauses, which we’re going to reuse again and again and again, they may have, you know, we call them clauses, we might also have disclaimers, jurisdiction. I don’t know I’m making up the content model for as we speak. In we do a lot of pharma work. So you’re talking about studies, studies have results results have different types, you have safety results and efficacy results. You have plot you have data with endpoints outcomes of the study that were planned when you went into the study and then other things you happen to discover during the study. So unplanned treat, different companies have slightly different words for these things, but you have these recurring structures within those then you may have smaller data points or smaller variable content. If you will, that changes.
So a product a product name or a particular molecule, or as you said, the author name on on a on a contract for a book. Yeah, these are tiny little things, which deserve to be and do need to be managed. But they are not well handled in some of these new systems that we’re seeing. It’s it becomes very difficult to properly say, Okay, well, this, this is our broad structure of what a contract looks like. That That doesn’t change very often. Then we have within the clauses we might see we might you know, have a bit more play there. Some clauses come in and in and out, in and out. And then we have these tiny little pieces. At the you’ve given the term data level, which which will change change all the time, and maybe every time Yep. And this this thinking of the hierarchy of not only am I structuring content, but I have to have levels to separate out frequently changing frequently personalized, personalized by or changed or adapted by region or content or the winner are your levels. Where am I levels for? General Ledger? Theory? You know? Yes. out where are ones levels? Yeah. Yeah. That’s, that’s fantastic. You brought this up? Do you have any ways that you kind of help people wrap their head around that?
Well, I mean, I mean, one of the technologies that I’m currently co designing is x forms, which is based on a lot of abstractions. So the idea of abstraction is that that you try and get to the essence of the thing. So for instance, although the than the name suggests it’s about forms. It’s not it’s a programming language. But what we do is, we have controls, which are which represent what they do, but not what how they look. So for instance, if you think of a drop down menu, where you select one thing or, or a pop up or radio buttons, they all do the same thing, namely, select one from a list of things, and you present them in different ways. And so we just have a control that says this is a thing that selects one thing from a number of things and has a number of possible presentations. But you represent it in the code in one in just one way. It’s there’s no visible in the code. There’s no visible difference between the different representations. But then where it gets that data from is completely separated out. So like a style sheet we also have, as it were, data sheets, which we call instances which contain the bits of information. So so that you have a complete separation of the data and the presentation and the control, which which in fact has turned out is a is very it makes it much easier to program in the first place, but also to be able to move through different different media’s different presentation forms. Yeah. So I think that some people can think about that kind of thing where if you’re defining User Options, I like that drop down. Example. So yeah, that dropped down, kind of radio boxes, tick boxes. There’s lots of ways you can represent here’s some options how to choose and we’re moving to a world where we want to define these in truly channel agnostic ways. As we’re seeing voice to text interfaces, you know, using language models to be the interface where we speak to the content, and users will be presented options, this ability to have this logic defined in a way where you understand you may show this to people in radically different ways, radically different ways. Different what we call different modalities.
Noz Urbina /UC
This is yeah, no Absolutely. No, no, no you know but this is a this is a term that I dug it out because we published our book 11 years ago, and now it’s coming into the into the into the dialogue again, because of Google. Google is releasing their their models and going Rs Rs is multimodal, whereas there’s just text. Well, in fact, one of the implementation named bridges of actions get demonstrated this by showing the same for the same program running on a screen with with typed input. And then on a voice browser, same same program but but in two different modalities one where you’re interacting using your voice and the other way you’re interacting using a keyboard. And we really have to head towards that sort of specification where you don’t specify once for one sort of modality and another way for another modality, because the abstractions are exactly the same. You want to choose one from a list. And it’s the implementation that should then okay, decide how it’s going to present that to the user and not that you have to specify that differently for every different modality. Fantastic. See, so we’ve been we’ve been saying this message for the content for years. Yeah. And those who took the message saying, okay, you know, if you if you properly structure your content, and separate it from the fact that it’s web format, or app format, or print format, then you will be so much more future proofed against maybe you’re not using those channels today. But when the market or management shows up and says, We want an app, or we want a chatbot or we want to who’s him a flipper, whatever, whatever is about to come out. They don’t want to wait three years for you to rework all of your content.
Oh, sure. Absolutely. So what you’re saying is the same thing now for the functionality, the interaction design? Absolutely, absolutely. But I you know, I have bad news for you in a way because you and I think you and I are thinking sort of about the future and and and we have a vision about how it could be but but in fact, we’re dealing with a generation that has grown up with editing documents in Word where everything is presentation oriented, and that a lot of people haven’t got into the idea of saying, Oh, this thing is a heading, but sort of select it, make a bigger font, turn it into bold, and that’s the heading and they’re thinking the whole time in presentation concepts and not in the underlying abstraction. Now the bad news is that similar things happened with the introduction of the book, and I like to compare the introduction of the web when entrusted a book because many similar things happened. That with the introduction of the book, books in though in the early fifth, the first 50 years of the book, looked exactly like manuscripts, they used handwriting fonts. They didn’t have page numbers, they didn’t have tables of indexes, indexes or contents. It was exactly like a manuscript and it took 50 years before people started having the bright idea of, of page numbers and readable fonts and all that sort of stuff. Now why was that? And then page numbers, can you have such a thing as an index? Exactly. What Why Why did it take 50 years? And the answer is because they were in a in a in an environment where people were used to manuscripts, and so they were providing for them. And so it had to they had to wait for that first generation to die out before they could start producing books for the new generation who didn’t have the expectation of manuscripts. And unfortunately, we are still living in an environment that still thinks in terms of bits of paper. I mean, we even say web pages, as an example. And they’re not pages at all, and that in order to achieve the things that we think should be achieved, we’re going to have to wait I mean, it’s hard to educate the whole of society in one big go. So I fear that that the first generation has got to die out before we can really achieve the things that we’re talking about, but that there are such huge advantages to do the things we’re talking about that they The sooner the better, because it reduces costs. a hundredfold. I mean, I have examples of people producing systems with X forms that are in 100 of the cost of the equivalent system using using old technologies. What Yeah, if you think about it, so when we talked about just just structuring your content properly so that it’s can be rendered out can be delivered out in three different formats. And the little variables like we were talking about earlier, so this is this is the British version, you know, British language and British metric units and etc, the data values, and also in other all the other English countries, and then we’re going to translate that, but we’re all going to share one source. And then we’ve got all these different products which are 90%, the same except for these 10, little 10% little differences. And you can configure those products depending on your industry and application. So we’ve been you’ve got from one set of source files, you’ve got hundreds, hundreds of actual variants of the possible output, right, just the content. And so now you’re adding the layer of the interaction and the functionality of writing the interfaces so that when when the boss says I want an I want that on an app or I want that invoice interface, you don’t have to go hire a team to recode the whole thing. So it’s interesting what you say there because you’re talking hundreds of different variants. But in fact, the number of variables that create those hundreds of variants are actually rather small. There are only three or four things that you mentioned. And, and so purely coincidentally, I’m at the moment designing actually only for myself, but and a recipe app that that reflects all these things like names of ingredients, but but metric or imperial imperial units and so on but also how many people you want to cook for. And so you know, you can change those things and you get a version of the recipe that matches those but the number of variables are very small so that you only have to change two or three things then you get one of the hundreds of options out that’s that’s that’s the joy of such a thing that you don’t have to translate your your content into 100 different versions. You just have one or maybe two versions, and you just leave the holes for calculations for what the reader is going to expect. at that spot. I actually had an old slide which I need to pull out more often, which was called multiplicity and it had exactly what you’re saying. So we had kind of like layers so it said like five five deliverable document five deliverables documents for better words. So collections of information, so about five sources, times, five regions times, five industries times, five for et cetera, et cetera, et cetera. And I think I went like four or five layers that wasn’t always five. So sometimes sometimes seven, but these, as you say, small numbers you can cut them on your fingers. And you keep multiplying out on the right, you have some more 2000 Yeah, different.
Noz Urbina /UC
Exactly. Exactly. Exactly. I which and which organization has two options today, either they don’t deliver them, even though there is a market need, or a user desire. Or they deliver them at excruciating cost. And it’s one of the other Yeah, yeah. So I mean, that’s one of the powers of x forms is that that these are variables in in your in your data that you change, and then all the rest are just constraints and calculations based on those and they will content and comes out to match what you’ve what you’ve just specified. So that’s one of the reasons why you can cut the the cost of production by such a huge amount because you’re you’re not having to worry about all those differences. So we’ve mentioned accent forms, kind of on the fly several times. Let’s let’s get back. All right, so I’ve heard of I’ve heard of a website, and I’ve heard of an app. So and I understand you know, Microsoft Word introduced X forms to me from the beginning. Well, who uses it? What is the what? Who, who uses it when what I asked someone to use it what I use it myself?
Right, so well, so let me back right up. Yeah. So the first programming languages were designed in the 50s, COBOL, Fortran Algol, Lisp, the Big Four and in the 50s, computers were immensely expensive millions. And the programmer was very, very cheap, sometimes almost free in comparison. And so the program said when I met you that, that they would, the company would give you some programmers, if you bought a system that was not bought the computers were were so expensive that most didn’t buy them, they leased them. And if you leased a computer you got you got a couple of programmers in the deal. And the reason that that was good was because all computers were different. So it was great to have somebody in house already who knew the operating code the operating system and could teach teach others and start writing programs for you in a world conservative before standards. Exactly, exactly. And so, so the economic relationship between computer and programmer was a computer was hugely expensive, the program was essentially free. And so therefore, it didn’t matter if programming took a long time as long as the computer wasn’t put to too much trouble. Over the years, of course, the cost of computers has plummeted and the price of programmers has risen so that in about the ages they actually crossed but without any sort of fanfare, so that now computers are almost free, and programmers are really expensive. But we’re still programming in programming languages that are direct descendants of those four that I mentioned, that they’ve still telling you, you’re still telling the computer step by step how to perform, because you’re assuming that you’re still assuming the computer’s expensive and the programmer is cheap, and it’s not that way around. So my much of my research has been based on trying to fix that relationship trying to design programming languages that don’t make that assumption and therefore try to save programming time and not computer time so that you might you let the computer do more of the work for you. So so that’s that’s one of the things behind x forms. Now x forms is rather than a procedural programming language is a declarative programming language. Now, the example I use of the difference between those two things is at school you learn how to add, multiply, divide, and subtract procedurally, so you actually learn how to do it, but at a given point, you’re told what the square root is that the square root of a number is a number that if you multiply it by itself gives you the original number back. You learn that concept. You understand what it is, you could recognize one, but you nobody comes from school with the ability to calculate a square root. Nobody knows how to do it. So that’s a declarative definition of square root. You know what it is you said what you want, but you haven’t said how to do it. So when we’re we’re designing a programming language where you decide much more describe the solution space rather than the the roots to get to the solution. It’s like, you give a picture of the cake that you want, rather than the recipe to bake it. So it’s much much shorter. It’s much easier to do. It uses it a little more computing power, but one big example said that it was only about 50%. more computing power, but they this was a huge project that they normally did with 30 people in five years and they managed to do in one year with 10 people. So so they’ve gone down from 150 person years to 10 person years. So you know, that was a cost that they you know, they the extra runtime wasn’t wasn’t a real problem for them because they’d save so much money in the interim. So I don’t want to I don’t want to interrupt that it’s yeah, I never in the audience for all these times. They’ve gone to technical departments or teams and said I want annex or I want to change this functionality or I want to change this webpage, etc. And had to wait eight months. For what what from the outside seems to be trivial changes like why is it taking you so long? And so I think this this getting to the root of building things, so that they are maintainable faster, they are adaptable faster. And as you said, like 100 times faster. That’s the kind of change that we need.
Yeah, exactly. Exactly. And so I mean, I personally believe that once the managers get wind of this fact that that that you can, you can produce the same stuff in 100 for the with the hundreds of the resources. I think that in the end, programming will, will go that way just because there’s so much advantage to it. And you know, and I have I have many examples of such such such such advantages. I mean, there was an insurance company that wanted an app. And so the manager said to the programmers, I want you to work out for me what resources you need to build this app and send it also to the ex forms person. And two days later, they came back in the programming guy said, well, I need another 30 days to work out what resources I need for the app. And the excellent guy had written the app already in those two days single handedly. So you know that it’s, it’s, it’s, you know, it’s two orders of magnitude. Better so, you know, the advantages just are fantastic.
Noz Urbina /UC 34:14
I’m so exciting, the excitement of adding that on top of structuring content so that when I it’s just, it’s it’s dizzying. I’m sure that there’s gonna be several people in the room who are listening to this podcast who are struggling to wrap their heads around doing things hundreds of times faster. Right, right. I didn’t I wasn’t going to read this because I wasn’t gonna raise this because we don’t see fully eye to eye on this topic. But you kind of dragged me into it kicking and screaming, because you’re reminding me so much about the topic. Of how we talk to and how we discuss artificial intelligences, right, in this sense, or I’ll say what I think the parallel is that when I’m training people, or I’m giving people methods for how to use an artificial intelligence, we talk about turning the problem around, rather than figuring out exactly what you want this, this application to do. You are trying to figure out the logic of the logic of what the result you want, how do you want the result to be what is it that you’re after? And rather than micromanaging? You’re you’re explaining what the solution looks like to things do you see any connection between these two and is my initial gut reaction of oh, well, if I could explain with my words, to an artificial intelligence then it could build me an extraordinary application could that you know would that allow even me as so many of the people on this on this listen to this podcast will not be programmers themselves, right? I do not want to become programmers did not want to get their hands dirty. But writing out some some clear instructions of what they’re looking for. Might be so do you envision a world where and I I do do what do you think about this vision of a world where the average business owner can user can actually be able to explain, explain their desires and actual applications will come out the other end? So my, my fundamental point of view is that if it works, you should do it. Right. I mean, that that I mean, that’s, that’s, that’s the essence of everything. And so, so saying to AI what you want, and then getting it is the same as saying that to a programmer, and you’ve just replaced the programmer with a program that does the same thing. So you know, if that floats your boat then by all means, that’s That’s great. My as you know, having been to my talk, my my, my my problems with AI is that firstly, there’s no I in there yet that that it’s it’s machine learning, but it’s not artificial intelligence. So that you can’t trust it and you know, as an example, I asked a machine intelligence and machine learning program, artificial intelligence, to write a bio for me, and it’s sort of it was okay, but I couldn’t accept it as it was because there were errors. So I had to I had to fix it up and that’s the same with programming that I wouldn’t trust. Well, to be fair, I wouldn’t trust a program written by anybody because, because, because we programs, programmers are used to the idea of debugging. And so that means you’re always going to get buggy programs, as we are seeing for instance in the UK at the moment, the post office scandal I don’t know if you follow that but but so quick little thing about the basically the post office was just using buggy software for ages and actually the citizen taxpayers were getting the stick.
Steven Pemberton 37:59
Well, yes. The post of the people at the post of the people running the post offices they were getting blamed for the mistakes in the software, that money was disappearing because of bugs in the code, but, but anyway, we’re getting away from the
Noz Urbina /UC 38:18
benefit applications have bugs. Yeah, absolutely. Everybody in the modern world knows that.
Steven Pemberton 38:23
That’s right. they will from AI as well, because there’s no intelligence there. You’ve described something, it’s going to copy code from elsewhere, because that’s all that machine learning does. And so you won’t be able to trust it any more than what you get from a programmer. But but as I say, you know, if that if that does it for you then do it but but my critique over AI is more about that it’s not intelligent and and that you can’t yet trust it. So
Noz Urbina /UC 38:54
I think that we we can overlap and and we’ve only got we’re, we’re coming up kind of to the end of the time, and I don’t want to launch into that debate. But I think where we can see where we do see the IDI very much is when we come to this reliability word. And I’m cooking an article in my head about this. I haven’t I’ve haven’t made it public yet. But the where you can use what we’re calling AI. Today is in the places where you can tolerate a lack of reproducibility and consistency. So the point is, we’ve been thinking about computers as as infallible calculating machines, you know, no matter how many times you ask the computer to for for a 10 digit number factorial or, you know, to multiply 252 numbers together every time and every computer will always get that right. And the i and then we just scale that idea. Okay, so I need these files moved from here to here. And every single time I want all the files from here moved from here to here. And if it doesn’t, we call that a bug. Now, we’ve introduced a type of computer application and I also kind of put big air quotes on that, which doesn’t do that. You know, it’s a roll. There’s a roll of a dice there. Sometimes it’ll just do something. You know, it’ll come up with some sort of weird variation of that because it dreamed it up or et cetera. But as you just said, kind of like we do. So rather than trying to employ these things, in places where you need computer reliability as we’ve, as we’ve come to be accustomed to it in the past 40 years, you need to employ these things in places where a little bit of a roll of the dice doesn’t disrupt your business process. Sometimes that there I’ve been thinking about it like creatives, you know, I don’t, I don’t hire my accountant to decorate, decorate my house and I don’t hire hire my interior designer to do my accounting for me, because there are different skill sets there. I need different things out of that type of resource. So this is kind of where I do agree that they’re there. We shouldn’t think about them as an artificial intelligence if what you mean by that is the benefits of a computer and the benefits of a human they do not have that. They absolutely do not have something that you can trust like a computer application, but with this sort of cookie eccentricity which is a you need to think about where can I use a little bit of cookie Centricity in my business where where am i They got useful and don’t expect, don’t expect the best of both worlds. Right? Right. Right. Right, which then raises the point, you know, to what extent do you want to do and have an AI do your programming but, you know, it might it might be just as bad or just as good as as a human. But don’t don’t expect it to be perfect by any means. You’re still going to have to check it out and and test it and like my biography that it wrote, you know, it just got things wrong. It met well, didn’t get things wrong, it made things up that just weren’t true. I saw my just as long as we’re here on this. I have been I spent the holidays trying to try to get a ChatGPT that will take the transcript of the video, pull out some salient quotes and put the times next to them. In capable, it just couldn’t could not do it. It would make up times it would put in phrases that weren’t quite there. It would start to do the first half and it will kind of get bored. It really does feel like working with a with an intern who just arrived maybe from college, but also maybe you know, you know high school high school diploma kind of in Right, right. Right. Right. It really feels that same kind of way. Yeah. Okay. So, there are like to sum up. So we’ve got lots of lots of great things on the horizon. You can you can check out where can people learn more about x forms? And if you just go to my homepage and I mean, that’s just a collection of all the things that I do but in including resources about x forms that where you can learn more. Did you do any? Do you have any kind of like a, like a best introductory webinar or something that people can put into?
Steven Pemberton 44:45
Yeah, yeah. If you if you go again, if you go to my homepage and select the X forms, Introduction to x forms, and you’ll find tutorials, you’ll find some videos. You’ll find a whole bunch of equipment, quick references and a whole bunch of resources. Great. Okay. So I recommend people to check that out. Maybe not for your own needs, but just understand the code. Yeah, what it’s about and maybe raise that in your organization. See if anybody else has looked into it. Maybe you can be an innovator, as always, with everything that I’m associated with. Probably not necessarily something that you’ll implement tomorrow but getting you ready for the next 10 to 20 years. I have to warn you that a lot of programmers don’t like the idea that they have to rethink how they program. And so if you ask a programmer to compare, they’ll often say oh, no, this is no good or this is not the sort of stuff that we do. And it’s true. Yeah, it’s too different. But I should also point out that just this week, I found out that the the FDA the drugs organization in the United States is using x forms, amongst amongst others. So also something that UK ambulance service was the National Health
Service. Yes. So so this was a project that disastrously expensive project that went completely wrong. And then one guy who worked on that project went away and rewrote it in three in three years, alone alone. Using X forms. Now it’s running in the National Health Service hospitals, and it’s being rolled out over in Ukraine at the moment. So, you know, it’s also good to know that it’s being put to good use as well. I
Noz Urbina /UC 45:15
mean, Ukrainians know their technology. Yeah. Yeah. Okay. So thank you so much. It’s a pleasure and an honor Steven, to have you on the show. Great to be here. My thank you for joining us and everyone, let us know if you have any other follow up questions for Steven. Oh, if people wanted to are you a networker LinkedIn kind of? Yeah. You’ll find me on LinkedIn. Yeah. Yeah, absolutely. Okay, fantastic. Thank you so much. And we will see you all next time. Fantastic. Cheers. Okay, good. It was pleasure. Fine. Wonderful. Yeah. Thank you very much. That was really, really fun. Thank you, Steven. Have you seen the IBC in the IT Crowd? The British? Yeah, of course. So in my head, my head I’ve been lovingly calling you an elder of the internet.
Steven Pemberton 46:10
That episode. No, I haven’t seen that episode. Oh, I shall immediately go and get it. And
Noz Urbina /UC 46:18
elders of the internet. All right. I’m gonna probably drop that gag in the intro.
Steven Pemberton 46:26
Thanks. Okay, good. Well, hope you get over the tiredness soon.
Noz Urbina /UC 46:29
Yes. Thank you so much. Thank you so much. Okay, bye