there is physical distance between main and backup site. that means there are physical limitations how fast transaction can be done if you require master/master configuration (you are limited by speed of signal propagation). to ensure both sides to have same data (sync replication) you need 2x time distance. that means high latency. and stock exchange wants to have low latency as much as possible. so only design is to have main / DR (DR being async copy).

Accept that your puny 5 senses will be perfectly simulated ... soon. then Reality goes by by .. tech evolution is going forward, and humans are mostly standing in one place. many years ago audio recording was "proof" and is then faked, then still image was "proof" and Photoshop enabled to make any you want, now artificial video is created good enough and no longer "proof". we just need simulated smell, taste and touch and we are done :)

i cannot see reason why GPL-ed sound editing desktop application must know 'user.age' in any way/shape/form in first place ?!?

tried btrfs on linux server with 50+ locally attached disks. it worked, but feel somewhat slow. caused few panics. tested for half year. then wiped disks, installed FreeBSD+zfs. then forget about it, since it worked with expected performances and absolute stability.

then you would need to store this agreements for some time, and have someone who can search them when needed. they want to get out that market and clean things behind. probably cheaper to recycle them than ship them+manage papirology.

"The Warrior's bland acronym, MMI, obscures the true horror of this monstrosity. Its inventors promise a new era of genius, but meanwhile unscrupulous power brokers use its forcible installation to violate the sanctity of unwilling human minds." MMI - Mind/Machine Interface

simply add possession of HAM radio&License as negative to SocialScore. problem solved :)

Main memory of today's computers cannot be considered "private" & "secure" enough as Intel IME and similar garbage can directly read from it. Assumption that you have total control of memory is false.

also, Google have own OS (android) which reducing Apple smartphone business income, and Google is buying custom build servers (not hyper pricey IBM gear).

from Draft: The IP addresses and ports can change in the middle of a connection. ad/js/css blocking will be much more difficult, as browser probably dont have proper API for plugins to kill those. blocking will probably need to move to some proxy (which need to be done from scratch) also browser should be put in separate linux namespace to avoid ability to go-around proxy. iptables support for this protocol woud be much more complicated as there is now mixing of data/tranasport/security features ... IMHO .. garbage (just to gain few ms in latency)

nope. you are not only one. i'm using fvwm since late '90is :) and dont have plans to change that. i tryed few others and returned back. power of window manager + no othern clutter. also no eye candy who waste resources.

first on my mind list:

- parallel read/write to multiple disk in RAID5 in btrfs (speed!, specially when multiple reading/writing userspace processes on btrfs which contains 10+ disks)

- pull back multi-path mgmt at FibreChannel driver level in kernel (WTF userspace need to know that scsi device have multiple path, userspace just get device and use it, rest is kernel work to handle, IMHO)

- throw out interface teaming (bonding is more that good, long time used in production, with VLAN tagging. 0 problems :)

p.s. you ask for Linux, not GNU Linux. Right? :)

dd if=/dev/zero of=/dev/sda bs=512 count=1 faster, not real damage, much to learn during recover :)