/tech/ - Technology★

Nothing. Most of our problems are just a consequence of the fall of the west, which is more important than anything technology related.

I'm mad plan9 didn't replace unix/POSIX, though.

UNIX' main design principles still dominate OS design though. Every OS, even Windows, makes a clear distinction between the kernel and the components that run on top

>>1006093

>implying privilege separation doesn't predate unix by a decade.

>implying privilege separation is a part of the unix philosophy at all. (even the fake version proposed by the authors)

The real unix philosophy:

A small program is more desirable than a program that is functional or correct.

A shoddy job is perfectly acceptable.

When faced with a choice, cop out.

From this you naturally get the more well known aspects of the unix philosophy:

Structured data is bad data.

Functions should take state as input and change state as output, they should return nothing of use.

Functions should vomit their state without debug information into a file should a exception occur.

There should be as little a standard as possible so that each program can make their own unique mode of interacting.

All changes should be permanent so that users learn the hard way the proper way to use our tools.

Code is documentation, so why would we write anything else.

I could go on but it's likely not good for my mental health.

>>1006093

not templeos

>>1006127

bait

>>1006087

The fall of the west is just a consequence of

>>1006082 (OP)

>checks openbsd

>checks netbsd

>checks plan9

Kill yourself, Lennart.

>What's it like knowing that the Unix philosophy is dead?

You'll be long dead before the Unix philosophy is you lazy nigger faggot OP.

This is not a thread.

>>1006180

I wish.

>>1006207

Plan9 is literally someone looking at how shit >>1006127 is and making the least competent, least significant revision upon it, specifically this line is transformed:

>Functions should take state as input and change state as output, they should return nothing of use.

<Functions should take state in the form of a file as input, and change state in the form of a file as output, they should return nothing of use.

but what it's trying to fix is Unix's hatred of structured data and functions not taking arguments and returning results by dumping a string into a file. Imagine the right solution, programs being functions which map functions to a common data structure and print this information using the structural data (either a arbitrary dimension array, or a arbitrary complexity tree). You wouldn't have to have a parser of string input in every program, you could easily call these functions within your functions removing the need for a shell like language, if you don't have a shell language but rather a REPL you can get debug information out of these functions in the standard manner, by passing the structure of the data you remove the need to have more than one specialised procedures for the purpose of outputting human readable data (a good example is ls rather than find). Now imagine that code is data, code can be serialised and sent over the wire, code can be saved and stored for later, etc, etc. Doesn't clean drinking water sound nice anon?

Feels good man. Unix isn't important, users who have freedom is more important. To that end, we have GNU.

Defining the unix philosophy is a poor joke, and 'unix philosophy' is a buzzphrase

AT&T unix was a hack job, just like most corporate kludges.

BSDs were designed with significantly less organisation and yet performed better.

All following POSIX copies ejected AT&T style systems and began copying BSD where they could, save for extremely long term installations.