This work is licensed under the Creative Commons Attribution Noncommercial ShareAlike 4.0 Licence (CC-by-nc-sa) (or at your option a greater version of it).
Copyright © 2018 Shlomi Fish
This work is licensed under the Creative Commons Attribution Noncommercial ShareAlike 4.0 Licence (CC-by-nc-sa) (or at your option a greater version of it).
TODO:
Note: there is a similar essay on software disenchantment and probably many other similar ones.
Essay: reflections on how software quality has improved/worsened.
Follow-up on http://www.shlomifish.org/philosophy/computers/high-quality-software/rev2/
Follow-up on “Optimizing Code for Speed” Wikibook
Follow-up on http://perl-begin.org/tutorials/bad-elements/
Follow-up on http://www.shlomifish.org/lecture/C-and-CPP/bad-elements/
Follow-up on FAQ - “This site loads so quickly. What is your secret?”
Freecell Solver speedups ( https://github.com/shlomif/fc-solve/tree/master/fc-solve/benchmarks ), new features ( http://fc-solve.shlomifish.org/features.html ), tests suite, CI, etc.
Cost of code complexity and readability
Learned about CI from https://perlhacks.com/2012/03/you-must-hate-version-control-systems/
Since the essays were written, some apps improved while others got worse.
Electron-based apps:
Software becoming slower and more resource hungry:
Windows Update — slow:
Reportedly still breaks the system sometimes
Gets invoked at inconvenient times
Like at computer shutdown
Red Hat’s dnf
kinda slow
hopefully will improve
Mostly Positive: VLC
Easy to use interface
Popular among many Windows and macOS users.
Mostly Positive: Emscripten / asm.js / WebAssembly
Mostly Positive: pypy
Fast
KDE/Plasma’s update from 3 to 4 was poorly done, and so was 4→5.
Seems like most people except for the KDE core developers do not appreciate the emphasis on the so-called “Activities” instead of on virtual workspaces.
Mostly Positive: cmake
Not perfect but less errorprone and faster than GNU Autotools
Reflections on “Simple” in the context of software:
Simple may imply lacking features or being hard to use.
Hi! Good article. I should note that based on what I wrote here — Prof. Knuth may have been wrong because many small optimisations eventually can add to a significant improvement. Perhaps sometimes an optimisation is not worth it because the program is already fast enough, but I feel that we have accumulated too much junk bloatware recently (see my notes ).
Regarding the “it’s good there is more work for me”, this is the broken window fallacy — and it is also echoed in the “How to become a hacker (not ‘cracker’) howto”. You likely have better things to do than to fix others’ past failures.
This work is copyright by Shlomi Fish and licensed under the Creative Commons Attribution-Non Commercial-ShareAlike licence version 4.0 (CC-by-nc-sa) (or any later version). See my interpretation of it.