Unix has been bad before

May 31, 2015

These days it's popular to complain about the terrible state of software on modern Linux machines, with their tangle of opaque DBus services, weird Gnome (or KDE) software, and the requirement for all sorts of undocumented daemons to do anything. I've written a fair amount of entries like this myself. But make no mistake, Linux is not uniquely bad here and is not some terrible descent from a previous state of Unix desktop grace.

As I've alluded to before, the reality is that all of the old time Unix workstation vendors did all sorts of similarly terrible things themselves, back in the days when they were ongoing forces. No Unix desktop has ever been a neat and beautiful thing under the hood; all of them have been ugly and generally opaque conglomerations of wacky ideas. Sometimes these ideas spilled over into broader 'server' software and caused the expected heartburn in sysadmins there.

To the extent that the Unixes of the past were less terrible than the present, my view is that this is largely because old time Unix vendors were constrained by more limited hardware and software environments. Given modern RAM, CPUs, and graphics hardware and current software capabilities, they probably would have done things that are at least as bad as Linux systems are doing today. Instead, having only limited RAM and CPU power necessarily limited their ability to do really bad things (at least usually).

(One of the reasons that modern Linux stuff is better than it could otherwise be is that at least some of the people creating it have learned from the past and are thereby avoiding at least some of the mistakes people have already made.)

Also, while most of the terrible things have been confined to desktop Unix, not all of them were. Server Unix has seen its own share of past bad mistakes from various Unix vendors. Fortunately they tended to be smaller mistakes, if only because a lot of vendor effort was poured into desktops (well, most of the time; let's not talk about how the initial SunOS 4 releases ran on servers).

The large scale lesson I take from all of this is that Unix (as a whole) can and will recover from things that turn out to be mistakes. Sometimes it's a rocky road that's no fun during things, but we get there eventually.

Written on 31 May 2015.
« My view of setting up sane web server application delegations
The problem with 'what is your data worth?' »

Page tools: View Source, Add Comment.
Search:
Login: Password:
Atom Syndication: Recent Comments.

Last modified: Sun May 31 21:53:41 2015
This dinky wiki is brought to you by the Insane Hackers Guild, Python sub-branch.