Computer Vision News - May 2024

Computer Vision News 28 internet is what led to its success. I feel like the same can be said about robotics. We’re not going to make a lot of progress, and particularly small organizations aren’t going to make a lot of progress unless we have a diverse and rich, open-source ecosystem. What do you think we need to get there? Generally, at least with respect to open source, and particularly with respect to ROS, we have an awareness problem. I don’t think people understand it exists. I personally think the thing with software is in the early days, you had to write everything, and then somewhere in the ’80s and ’90s, we kind of said, ‘Oh, we should write libraries and make our lives easier and inside of those libraries, we’ll put the things we use every day.’ ROS ends up being where we put all the things we use for robotics, and you can take from that body of work what you want and leave what you don’t want. I always see people having this discussion of it’s either open source or no open source, and it’s more like we’re going to take this set of things from over here, and we’re going to build this set of things over here because we have this particular set of requirements. When you have binary thinking where it’s like this or the other, it’s not generally productive. The other thing to remember after that is that it’s like a national park. You should visit it, you should go do things there, but you should also leave it better than you found it. If you can pick up some trash or clean something up along the way, we’d really appreciate it. Do you not mind cleaning other people’s trash? Every software job involves a little bit of taking out the trash. [she laughs] It’s like the young students that come in, and they think that all they’re ever going to do is write greenfield code. The reality is you’ll be cleaning up people’s things forever, and that’s part of the job. I feel like most of us should spend a little bit of time leaving the open-source world better than we found it. My readers are aware of this position because it is similar to Yann LeCun's, whom I interviewed 6 months ago. Do you think it is the Women in Computer Vision