Joel Spolsky wrote a good argument to the contrary: https://www.joelonsoftware.com/2000/04/06/things-you-should-...
Any good examples of large projects where a rewrite from scratch has been good?
This is ridiculous. Maybe that's applicable in a very short timescale, but definitely not to the protocol which appeared in 1984.
There is a thing called progress, people invent quite a lot through the years: type systems, patterns, design ideas. Not to mention that hardware has changed quite a lot through the years: X11 was done in the age of terminals and absence of hardware acceleration.
IRIX, OpenGL.
They invented it.
Technical debt i.e. "code that should be refactored, because it was originally implemented in a hurry"? Yes, technically you're right and deleting all of the code is a way to also get rid of the code that should be refactored.
The same way whole-limb amputation is a very effective way to remove nail polish.
More like fundamental design decisions and API's which you can't refactor without breaking compatibility.
Maybe it's time for people to stop whining in the internet and start writing code?
X.org code is there, support it, contribute to it, improve it if you want to.
> infamous heartbleed bug
Yeah. Because real open source is like that: millions are arguing in forums, few write code.
And that's the problem Wayland is trying to solve BTW, by removing quite a lot of obscure legacy which only a few properly understand.
Writing things from scratch is usually good, it allows:
1) To follow modern practices so modern devs can understand the code and contribute
2) To use modern languages and technologies which are inherently more safe and secure (due to better type systems, linters, language design)
3) To get rid of technical debt