(TXT) View source
       
       # 2024-04-16 - SystemD vs SysV
       
       I've gone back and forth between using different operating systems
       over time.  For years i used Fedora at home and RHEL at work, so I
       had years of using systemd.  I never had any major problems with
       systemd per-se, but i can totally understand why other people do.
       It represents a major change.  It touches all aspects of the system.
       And it has a history of breaking things.
       
       When these problems are reported, it's not unusual to get responses
       like "fix your users."  The idea being that the users were using the
       system in the wrong way.  It was a "happy accident" if their wrong
       way ever worked at all.  I view this as a philosophical struggle
       that affects all architecture including technical architecture.
       
       On one hand you have the idea that it's too expensive to support
       every possible use case.  Thus the need to "fix your users."  
       
 (HTM) https://github.com/systemd/systemd/issues/2039
       
       On the other hand you have the idea that the whole reason the
       architecture exists at all is because of the users.  In other words,
       "the Internet is for end-users."
       
 (TXT) gopher://gopher.fnord.one/0/Mirrors/RFC/rfc8890.txt
       
       In the life cycle of an architecture, it is designed once and then
       used for a time.  When my grandparents hired a carpenter to build
       their kitchen, my grandmother asked for tall counters because she was
       tall.  The carpenter refused.  He explained that there is a standard
       counter height, and that deviating from the standard will cause
       problems down the road.  So he built a standard kitchen that my
       grandmother used for the rest of her life.
       
       SystemD vs SysV is more of a kitchen than a bike shed.  It sits at
       the heart of the operating system like the kitchen sits at the heart
       of the household.  The counter height is an implementation detail.
       whether it is short, standard, or tall is not inherently problematic.
       
       The problem with "fixing your users" is the hidden external costs.
       Whenever you build and use something, it costs more than you think.
       Ultimately, you are paying for it, not the architect.  If it doesn't
       fit you then you will pay the price of poor ergonomics all throughout
       its lifecycle.  Are these ergonomic problems a fair trade-off for the
       economies of scale promised by standardization?  That SHOULD be your
       decision to make.
       
       I wish i could go back in time to confront that architect.  If he
       had the gall to challenge me to find one thing i liked about the way
       HE wanted to build the counters, then i would challenge him to find
       more compliant clients.
       
       Because i value personal freedom so much, i don't intend to judge
       anyone for which software they choose to use, or how they choose to
       use it.
       
       tags: bencollver,retrocomputing,technical
       
       # Tags
       
 (DIR) bencollver
 (DIR) retrocomputing
 (DIR) technical