Listing manually post-installed packages in Debian

The problem of getting rid of obsolete stuff is the same as the one that keeps C clinically alive, even if brain dead.

“We have always done it like this.”

“We have used X since decades, already!!!”

“X is everywhere, therefore we cannot switch.”

“I wrote an X script 30 years ago and it still works!”

“The new technology is not as mature as the decades old X, therefore it is not stable enough for production.”

I could go on and on with these examples, paraphrasing in what way obsolete technologies are apologetically defended and sometimes even praised.

Just replace “X” with any obsolete crap. Whether it is C, POSIX or certain UNIX tools, like e.g. the original Bourne shell, it’s always the same shit they use as arguments.

The mentality needs to change.

  • No infinite backward compatibility. It’s fine, if a script works for 10 years and after that much time breaks because of a major improvement in the scripting language used.
  • If something is inherently broken and the fundamental design of something obsolete is utterly broken and not adjusted to our reality, anymore, then it should not be “fixed”, it should be replaced by new software, written from scratch, which fixes the fundamental design issues, rather than “fixing” the old crap and making it even more complicated and less secure. X11 would be a great example of how a newly written software is desperately needed, as X11 just cannot be “fixed” without rewriting everything, anyway.
  • Design patterns, software architectural best practices & real life requirements change. For example, 30 years ago storage was a huge issue. So, there were tools inherently designed to be disk space efficient. Now, times changed. Disk space is, compared to other options of holding information available, the cheapest & easiest way of holding information available. Therefore, we now need tools which leverage more disk space for the sake of other improvements, which are more fit in our current time. Stop trying to fix the old crap. It’s a waste of time, because most of the time those pieces of software need to be entirely rewriten, anyway, if you want a real design change.
  • Stop clinging onto old shit. I keep some stuff in my life for nostalgic reasons, but when it comes to software, I do not understand how people can cling to that, just because they have used it for so damn long. First of all, doesn’t it get boring at some point? Secondly, just because that software was fine in the 90’s, does not mean it stays like that forever. I wish those same people would continue to use Linux 2.4 and Windows 98, instead of polluting newer versions with their age old ideas and design decisions… (C developers are infamous for clinging onto the language, just because they have used it for decades… What kind of software architectural argument is “because I have always done it like this”? It does not make any sense, whatsoever.)
1 Like

The other side of the coin is… we dont want to be always grabbing onto the latest design fad

“… change has a Siren voice, and the latest song is ever the most
readily sung…” Ronald Knox

1 Like

I’m not talking about the latest design. I am talking about the recent most successful and reasonable design. If it has been proven to be the best, as of now, since years, it’s time to accommodate to that.

Cant object to that… it is how we make progress. Develop something, test it , then use it.

What I ( and Knox) are saying is, chasing fashions is not progress. The clothing industry is a good example. Distrowatch is a bit of a fashion parade.

2 Likes