On Thursday, 28 December 2017 at 20:03:06 UTC, Joakim wrote:
The canonical example is how Apple doesn't compete on
feature/spec checklists but on an integrated experience that
just works better. That may be tougher to market for tech
users, but it is increasingly what people want, even many
programmers.
Spot on.
Even markets which are occupied by tech users, like BSD/Linux
servers goes into this direction. Systemd is a prime example. It
allows modern functionality and integration which in today's
world is a must. And it still not provies enough. Devices are
hotplugged all the time, they go in different power states, they
respond asynch to system wide, net wide, and soon, world wide
events ... the user mode paradigm of Unix is woefully inadequate
and unprepared for this world. It needs serious adjustments.
(arguably even kernels on free unix like OSes are unprepared, no
modern LPC comparable to what can be found in NT derived kernels
for example)
Yet the users do not seen it this way. A big part of Linux users
went so far as to make life threats towards systemd engineers.
They hate any attempt to modernization and integration. They do
prefer the autoexec.bat like init systems spawning fragile
scaffoldings of scripts. Too much choice begets fragmentation,
both from a technical point of view and a social one.
Fragmentation begets tribes, and tribes kill each other.
Likewise, in a company , it is in my opinion better to keep the
languages used across the board as limited as possible, instead
of an everlasting search for best tool for each of the problem
you face. Just see if you can solve it with the tools you already
have, don't try to build a scaffolding of techs you barely
unde5rstand and control because the are the "best"
