This week we had one class for group contribution, and another class lecturing on open source in the industry. We learned how open source organizations earn money in multiple ways. This addressed much of my doubt towards this field, and also provided me with insights into how familiar projects or organizations like RedHat work behind scene.

Open Source in the Industry

In this part, what impressed me most was SaaS (Software-as-a-Service). I have heard of this term before and I know that it is gaining popularity recent years. So when it is mentioned in the lecture, it caught my interest and I was glad to have such opportunity learning more about this concept.

Based on the lecture and the article “4 successful open source business models to consider,” I learned an example of SaaS, the WordPress.com. “WordPress is an open source content management system that has to be hosted on a web server, whereas WordPress.com is a privately held hosting service that runs using the WordPress CMS.” Basically, their word is to help people with less experience to build websites or blogs. In this case, customers don’t need to handle the technical issues of the WordPress CMS themselves. And since all the service in on a web server, customers don’t need to do configurations either, which saves a lot of energy.

I found that this could be a trend in the industry. For example, I notice that many companies are developing cloud services, saving users’ energy spent on setting up local environment. Many are also promoting products that provide extension for apps so that users with less experience can also build fascinating products using them. And of course, this kind of business models is not restricted to private projects, it is also suitable for open source projects in terms of gaining financial income. It is encouraging as on the one hand, it is a practical business model for open source organizations to make profits and keep the community alive, so that we can anticipate more “freely” shared knowledge and resources. On the other hand, it also marks a trend that people are trying to make users’ life easier. But also, while it makes people with less experience build products more easily, it also means that programmers may need to learn deeper and have more complicated skills in order to be competitive.

Group Project

By far, the most complicated PR I opened is for adding a new feature, enabling verbose in gaussian process regression, so that users can track the process of regression by specifying the value of this verbose. However, though it is towards an issue and hasn’t been done by anyone else, it has not been reviewed. The merged PRs are all about automatic parameter validation.

So it is a little bit disappointing that Sklearn maintainers are more willing to review simple PRs like automatic parameter validation, but in some extent ignore more complicated ones.

Another kind of disappointing thing is that recently, there is an issue about implementing some new functions that can enhance grid search and maintaining them by ourselves. But some maintainers are against it because the workload is heavy. However, I know that from users’ perspective, the enhancement is preferred. Some of my friends did convey concerns towards the old version when I talked with them about this. It seems that people are having different opinions about the priority and workload management. But we are still focusing on the issue and we hope more voices can be heard and some progress can be made.