Lately I have discovered that when developers fork our repo on github here vanilla-rtb some try to modify existing core libraries , even though it's kinda clear from the intention of our project is to be included as a framework either as we suggest on the front page via gh-submodule or gh-subtree and later specialized outside of our framework with custom implementation according to concrete business needs.
I had to add 2 more pages last night in a hope that people start reading text if reading code in the examples is hard to follow.
I am reaching out to C++ and other programming communities in a hope to get clues what went wrong . Currently it's the only open-source framework capable to bid on every impression while receiving up to 50K QPS ( queries per second ) on it's endpoint. While industry standard is 100ms around trip , given the network latency and other overheads you are only given around 40ms to bid on impressions . A lot of things need to happen in your stack before you can return a valid bid or no-bid response. So I think vanilla-team did an excellent job delivering this light weight library with minimum dependencies on the third-party libraries . However , even though we bid within 20 microseconds we are not the most popular stack out of 3 available today. So, the question to community :
- What went wrong ?
- What needs to be done to improve popularity of this stack?
- What needs to be done to attract more contributors ?
- Why our stack attracts a lot of coders from China ?
If you are reading this question and you are Python, C++, Java , JavaScript programmer interested in contributing to our open-source project or have questions or suggestions on how to reorganize our project please respond , as we hope to hear from you ASAP!
Aucun commentaire:
Enregistrer un commentaire