Imagine that there is a legacy(maybe) code running with boost and we are suddenly saying that we no longer support boost and that you have to change to std. This will be so much pain for end users. So instead what we are doing is not making end users go through pain by reimplementing boost in terms of std and in return we will have to do a little hard-work. (By doing so we would have honed our skills and helped the community with stable chrono. Win-win for everyone.) Mr. Howard Hinnant might have envisioned that boost has served its purpose and its time for it to retire. Think of it in this way, after this task this library will require zero or very little maintenance as its using std. The only time maybe if there are changes to upstream std then we need to maintain boost which I think would be after so many years. Also I would like to remind that boost is superset of std (there might be something that might have slipped my eyes but I think we can take care of it) Also I have seen end users heavily modifying boost since its open source. So I think reimplementation of boost in terms of std will give them a clean understandable code which they can modify very easily. We can also mark the entire doc as deprecated (except the {boost}-{std} part). Boost was always about cutting edge libs for developers and end users. Thanks! To all for such a great discussion. If during discussions I might have hurt someone, sorry for that! I think we should brainstorm a little bit. Thanks! Atharva Veer