mkj
Hello; I'm Michael. Live in the Swedish countryside, work as a programmer. Own dogs, love the outdoors. Linux on desktop since circa 2000; Debian since Lenny. Interested in security/infosec and privacy. Licensed radio amateur. Write too little, read too much.
Personal account, not affiliated with anyone else or any organization. Thoughts, opinions, etc are my own.
Please excuse me for shaving characters off posts to stay under 500. For longer texts check out my blog, where I post occasionally.

@RoganDawes@infosec.exchange Or use strong passwords and a restricted set of accounts allowed to log in over SSH.
If your root password is password123!, then sure, having SSH publicly accessible and with password-only authentication is a bad idea. But if the only account that can log in over SSH is an unprivileged one with a password like gIrD9T4t.639IV1TJLNUN2ruVt6Igc0iUE3nKdt7l6eGl10trJ2lhhiVJTeiGr9? How would that be significantly worse than key pair authentication?
Since @media aural and @media speech are deprecated in CSS since at least Media Queries level 4:
What is the current preferred way to detect (preferably in CSS) when a user is using text-to-speech output so that document elements that have no relevance or can't be rendered aurally can be hidden/skipped?
I've looked through the MQ4 media features list and don't see anything relevant; but surely I can't be the only one who wants to do this?
- Textual interfaces are hard so let's make it non-deterministic in addition.