 Well, I, I recently was at a conference on emotions and someone asked me, well, what do you do. Well, I wasn't doing anything on emotion so I told her I'm doing early modern legislation but obviously that's the wrong answer. Because I got scrolled by a colleague for telling that person that because I do a lot of cool stuff with digital humanities and as soon as I say that I'm working on early modern legislation people start to try to find a way to the exits and they won't invite me to talk whereas my agenda is kind of full with giving workshops on transcribes. So, in a way research comes a lot sexier when you're working on transcribes or when you're using transcribes because well, that's something people are willing to get a lot of interest in and start understanding it. So, even though I can write some pretty interesting articles, I think, on how governments function and on legislation. That's not something that most people seem to be interesting in, although this one did get a few hits in three weeks time. Yeah, I wanted to share some experiences on using transcribers for a little bit over four years now. And I figured that I put them into 10 lessons, and that still allows me to show you some of the pictures of the stuff that I'm dealing with. So, when I work with early modern legislation, and when I started this endeavor of learning to work with digital humanities, I had no idea what I was getting into, which may have been a good thing. It was a long learning curve, but an interesting one and I started that with the KB National Library, where I was doing a fellowship, and I was working with a computer scientist and they also assigned me a project manager and I had no idea why I would need a project manager. But it turned out I spoke an entirely different language than my computer scientist colleague. So, the first question, or basically, the starting point is what are you going to do. And that may sound like a really easy question, but I wanted to segment a legal text, and then figure out how to automatically metadata them. What I didn't know. And that's bringing me to the second thing is that layout matters because the computer doesn't understand what layout is for me it's obvious where there is a title. If there's an exercise or then anything in between happens to be the body of the text, but apparently the computer is stupid, and I have to tell them everything, which I didn't know. So I started using transcribes after a couple of workshops by Louise Seaworth back then. And I was kind of enthusiastic about creating models and running them over my pages, which was nice. I could easily read them I could search for words, but then the computer didn't know my titles so layout really matters and it's important to pay attention to it and even though two years ago the computer couldn't. You could not do the text recognition first and then the layout analysis that has been improved by now. So that's a good thing. So that make life a lot easier but it's so important to think about whether it matters for your output and whether you need to pour time into it, because play out analysis is, in my opinion, the most boring part and you want to start with the boring part because then the good stuff still has to come. One of the things I learned, especially when giving workshops is that you need to keep your transcriptions in a loop. So even though we keep saying you need to have 30 pages of transcriptions that's a really cool thing, but it's also bloody annoying if you have a very hard handwriting. So I start with five pages and then create a very small model and hopefully it works a little bit on my next pages and I can save a little bit of time, and I keep trying to do that. Even though the model is pretty easy. It makes my life quite a lot easier in the long run because the 30 pages is something I really do not like having to do. But I also learned, especially from my computer scientist colleague and that programmer, and that project manager, but also within the Institute where I now work is that it's very important to talk to various disciplines because understanding what is going on helps you also think about that process and makes you adapt to it. Plus it makes you sound smart. And that also gets you invited sometimes. And then bridge between disciplines and what I tell my students these days is that that is really the key in our future jobs. No one's going to, well, I still have some hope that someone's going to hire me as a historian who knows a lot about legal history, but chances are quite limited. So in order to increase my chances, knowing a little bit about digital humanities might help. And again, if you just look at your research, it helps you to anticipate on what you might be expecting as outcomes or what you could try to meddle with. Yeah, well, kind of obvious teach what you preach. It's a fun thing to do workshops because if you talk to people and you get to teach them how to use a tool or anything about digital humanities. It's kind of helpful to also have that settling your own mind and have that trickle down in your system and not just okay I click this button and click this button but you have to think about that whole process. It really helps just also for your own understanding. So even though if you're using it and you're explaining it to a colleague that can be really helpful for your own thinking process. The fun part is obviously that people tend to have the idea of understanding that you are working on all hands writing, and that the computer can help you. So I did it with the people I was in the train last night, and they actually came up with the idea so I can read my grandfather's handwriting. Yeah, you can. But as soon as I told them that I normally work on 17th century handwriting of legislation they were like, Okay, so that's interesting. Also cool, especially if you're still at the early stages of your career publishing models and data counts as a publication. And when I did my dissertation I did a lot of transcriptions by hand. And that didn't count as publications, but now I'm training models and I'm learning the computer how to read a certain and, and that does count as a publication so that looks cool on my list. And then, well, you history may be a slow science because while writing book is taking your four years at least for dissertation and then it takes a couple of years more to rewrite the whole damn thing. And then, if you're working with the digital humanities, then using transcripts, then you have data sets, and you have models and as some of you heard with the workshop, we can still make that output and figure out smart ways of referencing that. So, that's cool. What you need to be aware of is that techniques changes so we just heard this morning that expert is going to get out of the system and we have to get used to light which, well, that's a bit traumatic for me to hear this morning. Given the number of webinars I gave on expert and how many times that has been viewed. And there are new options as well, like a named entity recognition which is very interesting that we can now adopt that within a model. And that also needs to be something that you put in a footnote that you created the model like two years ago and you couldn't do named entity recognition yet. So that's not incorporated in your model because otherwise you will get comments from your reviewers that, oh boy you didn't incorporate named entity recognition. So, you kind of need to be aware and be aware that that technique changes and that you need to account for the moment that you wrote or created a model, or created the data set and what you could not yet include, but something from hindsight you could now do. So that's something to be aware of and you might want to go back on your steps and change it later on. And then, yeah, it's worth investing your time in, because, well, when I was applying for the grant that I'm currently working my project on. I had to do this interview in front of the panel. I hope they are not in the digital crowd because I know they are not here, but I had this professor asking me why I didn't put transcripts in my timetable. Because I was going to take a lot of time to put everything into a computer program and then use that well guess what it saves a lot of time, but somehow I do manage to get my agenda filled anyway, so I'm not saving, I'm not entirely saving time I'm still doing more stuff that's also fun. But yeah, working with transcript is certainly made doing research, lots more interesting. And I'm not sure if you would have listened to me if I were giving a talk about the early modern legislation. So thank you for your attention.