Heather Meeker: All right. Well, today, I am talking to Roger Dannenberg, who is a professor of community of Computer Science, Art and Music. What a wonderful group of things to be a professor of, by the way at Carnegie Mellon University and he is also a Trumpeter, a Composer and a Computer Scientist, of course, and he's the co-creator of Audacity, which is an extremely popular open source digital audio editor and recording application. Roger, welcome to FOSSDA. I'm so happy that you're willing to chat with us for a while today. So I'm going to just dive in and say what was it that brought you to working with an open to creating an open source project? How did you make the decision to do that?

Roger Dannenberg: Well, the project started basically from research needs that that we have that we were working on actually on query by humming, which was getting computers to recognize tunes that people might sing or hum and search a database for them and part of that is just recognizing pitch and tunes and we were working on some algorithms to improve that. And of course algorithms fail and then when they fail, you want to see what happened and we were trying to visualize audio and the analyses that we did and figure out what was going wrong and we just felt like it would be really great to have a tool that could help us display audio data in different formats and deal with different audio files.

Roger Dannenberg: This comes up over and over again in audio and music research and at the time there, the tools that were there, just, you know, didn't seem very good. They were either very proprietary and we had problems, you know, moving from one computer to another or they were designed for a special purpose and we were doing research and wanted something more general purpose.

Roger Dannenberg: And so, you know, maybe it was foolish to think that well, we could just start from scratch and build our own but that's what we did. And yeah, and so the actually, you know, I'm not sure that we ever certainly I never made an initial decision that "Oh, I want to build an open source audio editor distributed in the world and it'll become the most popular." Yeah. that It'll be the primary editor of its kind in the world that just never crossed our minds. In fact, even, you know, at that time which was around 2000, you know, most people in the world didn't have the kinds of computers or the kinds of network connections and everything else that you would need to even distribute something like this. So, you know, it grew out of some, some very personal goals and needs in a lab. And it just it kind of happened to be useful for some other things and make its way gradually out into the world and develop into what it is now.

Heather Meeker: Is there a decision point that you recall where you decided? This is going to be open source, and we're going to collect, you know input from a lot of people. How did that decision come about?

Roger Dannenberg: Yeah, a lot of from Dominic Mazzoni who was a grad student at the time and Dominic did did more of the work than anyone else. It was mainly the two of us and I was contributing some kind of low-level audio drivers that I had built and and worked on some symbolic display as opposed to audio waveform display, but Dominic really put together the basic waveform display and interaction. And it was it was Dominic that suggested in the very beginning that if if we are going to build something to display data , it really wouldn't be much additional work to cut copy paste and right, you know, and write it back out to disk. So you could, we could integrate some basic editing functions and I thought, well, yeah, it makes sense. That's a great idea, but don't spend too much time on it. So that worked out and and then it was really Dominic that that had some particular ideas about, for example, using the canoe license.

Roger Dannenberg: And, you know, sort of how things would be organized and made into a open source distribution. I had previously done a lot of software and made it available and not never really worried much about licensing or any of these details. So I was happy also to get the software out there, especially since we incorporated these editing functions and we thought it would be helpful at least other researchers.

Roger Dannenberg: And so, you know that and then the idea of a community and bringing people in really came about after the initial work. We, you know, it was just us and we were using the software and released it the software. And, but then later other people sort of started using it and wanted to make improvements and most of that actually happened interacting with he was he was the one that sort of pushed it out there and he was the one that people got in contact with and and I had a lot of other other duties teaching and researching and doing doing my stuff back at Carnegie Mellon.

Roger Dannenberg: So I was, you know, it kind of just evolved that way and then because I was still a user and I have written some of the software and I was helping to do some maintenance at at some point, you know, much later in the game. I kind of came back into the group and started working much more with the group and, with audacity.

Heather Meeker: So you said you'd developed other software before. Do you perceive a difference? Given this was an open source project from sort of other projects you had done. Like, did that change the character of the experience for you? Or what? What was it about? It being open source that that made any difference.

Roger Dannenberg: Well, I think maybe open sources is really an outgrowth and maybe I just see it this way, but maybe it really is an outgrowth of the academic world. Where in academia we do research and you know, we get the rewards are not so much money, but recognition. That people want to want to be recognized for being, for making contributions to whatever field, they do research or public or writing in. And so, we publish, but we don't, we don't publish, you know, and signed book deals. Well, actually, we do sign book deals, but terrible deals, right? We're lucky. We're lucky to get anyone to publish anything because dry academic stuff and, similarly with journal articles, like, you know, in some sense, we're professional writers, but we don't get paid by the publishers. Well, I don't know, if everybody knows this or not, but if you didn't know, it might be interesting that, you know, these thousands and thousands of scientific papers that get, published the author's, typically get zero and, and, the publishers. Many of the publishers are nonprofit organizations, like communications for the associate. I mean, the association for computing machinery, the ACM, or IEEE. These are nonprofit organizations.

Roger Dannenberg: So, you know, the point is just to get the literature out there and and share it and and then our, you know, hiring and promotion is kind of based on peer review and our recognition in the world. So, we don't get anything directly from publishing and and I think software is at least when I was going through school and, you know, decades ago software was kind of in the same category as publishing. People would write software, but academic publishing people would write software and systems and and they put it out there and see if other people will use it and that was just kind of like a form of experimental science to see what works and what doesn't and so it seems it was almost unnatural to me, to think about selling software and, I do actually have a few patents.

Roger Dannenberg: The first patent that I got was, essentially a software patent, which a lot of people said "oh, you can't do that" and some people said it was wrong. I felt like it was, I mean, this particular patent, we won't go into it, but it was unless you want, it was you know for a very specific invention and a very specific application, that seemed to have some real commercial use and the patent was actually key in enabling it to be commercialized. So, but other than that, you know, most of the stuff we do, we don't like, we never expected anyone to pay for this audio editor.

Roger Dannenberg: There were plenty of audio editor products out there that were available, but we wanted to make something free that would help people do the kind of work that we were doing and we were very fortunate to have some support through other means, so we could afford to do it and make it free.

Heather Meeker: Yeah. Well, so for anyone who's watching and isn't aware of it, I mean, Audacity is basically on, I mean, I would think on the top of every musician and you know, audio editor and everybody uses this tool. So clearly it evolved from something that was basically a research project to something that had, you know, an incredible like mass audience and users everywhere. So as the project changed from what it was originally to what it was eventually. Like how did how did the experience of participating in the project? Did it change for you at all? When it like took a step up and got so popular?

Roger Dannenberg: Yeah, I think I mean I learned, I certainly learned some things from this. And one thing, one thing I learned just from observation was, is that Audacity is very simple to use. And I mean, some people would say no, it's not really that simple but compared to the kind of software that people are using in studios that, you know, really take some work to get into and to get things done. Audacity was deliberately made to be a kind of what-you-see-is-what-you-get editor with direct manipulation and very simple menu commands and all of that stuff was there and on purpose, you know, we talked about it. And again, I have to give Dominic credit for wanting it to be that way because I think other developers including me just trying to you know add features I needed to do this.

Roger Dannenberg: Let's do it like this and then you know, if you do that too many times you end up with this monster that nobody understands. Yeah, and so Dominic was always kind of the back pressure saying yeah, we could do that. But you know that's going to be confusing. What if we did it like this? Or what if you know, or maybe maybe we just shouldn't do that at all. You can do that, you know with your own private extension or do it with another piece of software or something and kind of that attitude of, this has to be really simple, so that anybody can just download it and start using it right away. Really, it was a huge contribution. And I think that's also something, you know, maybe that open-source can do much better than commercial ventures, which is that and I've learned this from the commercial side that products have to have features and products in order to market products, at least the belief is that you have to have a long feature list and you have to in order to keep generating revenues you have to add new features and the list, get longer and longer and longer and the interface is get more and more complicated.

Roger Dannenberg: And what we can do with open source is safe, like we just don't care about selling, you know a new, improved version with more features to people. We just want to make the most useful thing for people to use right now and if they like it, maybe they'll upgrade in a few years. And when they upgrade, you know, we'll get a penny from advertising or something but it just doesn't matter. And so we can really focus on what's important for users.

Heather Meeker: Yeah, that's that's great. So was there anything about the Open Source process? Participation, you know participated in running a project and so forth, anything that you wish you'd done differently, or that was a problem or a challenge and that you figured out how to solve or what were the downsides and how did you and the team react to those?

Roger Dannenberg: Well, I think Audacity, you know, as you said it was incredibly successful and I don't think anyone could hope to do something so successful, you know, by design it just right. I've, in fact, I've tried to do multiple other things by design, if they haven't worked out nearly as well, so..

Heather Meeker: And if someone came to you and said, we're going to make this program that's going to be wildly successful. How do we do that in? People would have just said? No, that's not possible.

Roger Dannenberg: Yeah, so but some issues that have come up. I would say that the main issues are getting people to work together. I'm working and I haven't really touched it in a long time. And I went back to sort of see look for other versions and see if there are any bug fixes or things that uou know, may have happened on the side without me knowing about it and sure enough I found like a hundred different repos on GitHub that had clones of my software and various things. So, one thing that's, you know, that I've learned that's difficult, is to really build a community of people that are working together to support the some thing and I'm not really sure how that worked on Audacity.

Roger Dannenberg: Maybe it was just you know, the luck of having having the right people kind of sign up and also just having some very generous people that wanted to say "well I'm you know, I'm not just going to take Audacity and add some features and rebranded or something. I'm going to I'm going to work on. I'm going to join the Audacity project. I'm going to contribute to this one project and make it good." And, having so getting people to actually contribute instead of just, you know, maybe take the software and send you email saying you should fix this. You should fix that not putting any time into it. That's, I think that's a real problem.

Roger Dannenberg: So another problem is when you do have people that are contributing and very generous, you, well, at least with Audacity, there was a team of people that developed. So it was kind of like an inside, you know, it's as if you've got the inside group that has kind of all the power and all, you know, carries all the all the decisions and you have the outside group, which you also hope is there, you know, making contributions, but you know, the governance of very loose volunteer organization is really difficult because I think if it is completely open and everybody has an equal vote, then it's very easy to get distracted. You know, someone will come in and just make a lot of noise and maybe not have much to contribute but it can be a real distraction. On the other hand, if you have sort of a core team that, that team can you know, it runs the risks of alienating other people or being disconnected from other people. And, you know, I think Audacity's had some very generous people, you know, we've often referred to Dominic as the benevolent dictator, which is a good description. You know, he really put his heart into it and did a lot to keep things going and made really good decisions all along. So, but how do you how do you do that? Or how would you repeat that? It's I think those are some really difficult problems.

Heather Meeker: Yeah, I think leadership matters, no matter what you're doing, right? And a big project, a lot of what lots of volunteers can be very challenging right? Definitely. Well, so anything else you want to add? Before we wrap it up a, anything you want to that you're looking at in the future or anything you want to say about what your experiences have been?

Roger Dannenberg: Yeah, I would I guess I would say that. I mean one thing that came out of that, you know, that I learned from this project, that might not be obvious to people thinking about open source is that there are a lot of possibilities for generating income and one of those is, you know, some of some of the Audacity developers, did some consulting on the side. So they would, for example, adapt Audacity for to be packaged with a product or are bundled with a crime or they would, you know, work with researchers and get some money to add some feature for them or something like that.

Roger Dannenberg: Another possibility is that, you know, when you go and download software, you think of it? Oh, well great. This is free. And you might wonder if you thought about it, you know, who's paying the researcher, who's paying the electricity, and for those servers like if we're downloading a million copies of Audacity a month, that's actually pretty expensive to provide that much service. You can't just, you know, run that from home and so you might think, oh, well, you know how could an open source project even afford to do that. And part of the answer is that depending on where downloads take place, most downloading services run ads. And so the downloaders will actually pay for the privilege of providing the servers to download your software. It's not a lot of money. So, and maybe that's a good thing because if nobody's ever going to really get rich off of it, then it just kind of can help grease the wheels and you know, pay for an occasional machine or allow you to contract out something that really needs to get done that, nobody wants to do for pay for a graphics designer or you know, there's it gives you some flexibility in a mostly volunteer effort, and so anyway, those you know kind of the business and money side is something we don't normally like to talk about but are and it gets tricky. But it is real and I think it's part of the story thats interesting.

Heather Meeker: Yeah, very interesting. Well, thank you so much for talking to me today is really interesting to hear about the project and your participation in it and And thank you for all the work you did because I love the project myself particularly. So thanks so much for participating and have a great day.

Roger Dannenberg: Yeah, thank you. This has been great.

Heather Meeker: Okay, bye-bye.

Roger Dannenberg

Summary