Yeah! I agree with a lot of Christian's points. I have a Nord Micro Modular, and I'm getting ready to sell it soon. It has a great mid-90's mid-fi sound, but it's such a pain to use nowadays (no USB, tightrope "will my next OS update break the editor" compatibility, having to remember which patch is assigned to which number, no power switch, etc.) that it doesn't make sense to have unless I specifically needed it for live shows. Plus, the Axoloti makes for a pretty solid, more modern replacement at a much lower cost (http://www.axoloti.com/), especially with the Music Thing Modular control surface (http://modularaddict.com/musicthing-axocontrol-kit). The Jim Clark book is actually meant as an extension to the Rob Hordijk book: https://rhordijk.home.xs4all.nl/G2Pages/index.htm If you haven't read that one, it's a better place to start! I haven't read the Roland Kuit books before. The pricing is out of reach for me (and my students). Plus, I think the Nord Modular can be a pedagogical nightmare. Take a look at the "Logic Matrix Loop" image at the bottom of this page: http://rolandkuit.com/Laboratory.html That's not to dismiss Roland's massive accomplishment at all. It seems like an absolutely incredible resource! I just think that the Nord Modular's interface and sense of scale leads to massively confusing patches once you get anywhere past the basics. I've opened up a few of my patches from a month or two back and find them to be completely inscrutable. Same goes for aspects of the Jim Clark book, despite how much I love it. Look at some of these logic patches: http://www.cim.mcgill.ca/~clark/nordmodularbook/nm_logic.html#logic_functions At points, learning a Nord Modular feels more like Computer Science than Music. He's putting together patches that remind me of CPU architecture lectures. I find Eurorack modules to be the most intuitive way to learn, but they're just so expensive and it's really easy to make a costly mistake when designing a system. I'm hoping that more schools will start to develop small teaching systems so that you can have that actual physical experience of re-routing a patch. I think the absolute best way to learn, though, is simply working with other musicians! I learned a lot from Richard Devine when he visited my school, and we spent a few of the days developing patches. One of the most important things that I learned about putting together generative patches is that I massively overthought the setup process. There's this feeling of wanting to plan everything out in advance, and it ends up leaving you with a blank page. The best thing to do is to start with the most stereotypical patches ever: 4 on the floor kick drum, basic arpeggiator, boring ADSR-based subtractive, etc. Then, manipulate or replace one element at a time. Even something as simple as massively dropping the tempo can take you into great territories. For instance, I showed Richard a Machinedrum patch that I had been working on that hadn't quite come together. He turned a knob, dropped the tempo down from 120 to 20 BPM, and suddenly the patch was magic. I've tried to show this "start simple and evolve" technique in this patch: https://www.native-instruments.com/en/reaktor-community/reaktor-user-library/entry/show/9675/ Which brings me to another great generative strategy: stop tempo syncing things, and use slow modulation. You'd be surprised at how much complexity you can get out of two slow, unsynced LFOs! The Wavetable LFO that I ported from Sandy Small's oscillator is terrific for this sort of thing.