LWN's guide to 2024

Posted Jan 3, 2024 11:29 UTC (Wed) by paulj (subscriber, #341)
In reply to: LWN's guide to 2024 by elenril
Parent article: LWN's guide to 2024

Well, the whole point of SerenityOS is to implement a useable desktop OS, from scratch. Just for the hell of it, just so the developers can learn and/or get a sense of accomplishment.

If that's their hobby, why not.

LWN's guide to 2024

Posted Jan 3, 2024 11:38 UTC (Wed) by elenril (subscriber, #165899) [Link] (8 responses)

I'm not saying they shouldn't do it, I'm saying that such an approach is unlikely to produce something more useful than say TempleOS. Of course anyone is entitled to have fun in any way the like, as long as it's not hurting anyone.

LWN's guide to 2024

Posted Jan 3, 2024 12:06 UTC (Wed) by paulj (subscriber, #341) [Link]

At a minimum a group of developers have learned a _lot_ about developing across a very broad spectrum of computer software. From x86 emulation to implement a valgrind-like debug tool, to filesystems, all kinds of formats (audio, graphics, sound, etc., etc.), how to display stuff.

It's seriously seriously impressive what they've achieved. Hats off to them. If that makes them happy, good for them.

Have a read of the birthday posts, that review accomplishments each year, it's just amazing. Starting with (and this is /1/ guy at this point): https://www.serenityos.org/happy/1st/

LWN's guide to 2024

Posted Jan 3, 2024 12:50 UTC (Wed) by Wol (subscriber, #4433) [Link] (6 responses)

> unlikely to produce something more useful than say TempleOS.

But isn't that exactly the attitude that produced Linux? And going back in time, it produced WordPerfect, OpenQM/Scarlet, Unix itself, the list goes on ...

(Of course, a lot of stuff - maybe most of it - that was produced we never hear of today ...)

Cheers,
Wol

LWN's guide to 2024

Posted Jan 3, 2024 13:04 UTC (Wed) by paulj (subscriber, #341) [Link]

And who knows what will come out of SerenityOS? Even if SerenityOS stays a happy playground, where young devs can jump in and learn-by-implementing whatever they find interesting, if even /1/ of the /many/ little components they are developing ends up becoming important to the wider world, that's a huge bonus - on top of what they've achieved personally for themselves (learning, joy through the sheer accomplisment of bringing up a whole OS from scratch, etc.).

E.g., Ladybird could be quite useful to the wider world one day. Just that'd be worth it. Imagine if someone had stopped the KHTML devs cause Mozilla existed!!

Maybe Jakt - new systems language with (hoped) much more user-friendly memory safety (via refcounting) - ends up being useful.

Maybe some other pieces.

LWN's guide to 2024

Posted Jan 3, 2024 13:22 UTC (Wed) by elenril (subscriber, #165899) [Link] (4 responses)

>But isn't that exactly the attitude that produced Linux?
No, Linus set out to write an OS kernel. He did not also write his own compiler, text editor, or libc as a part of it.

LWN's guide to 2024

Posted Jan 3, 2024 14:47 UTC (Wed) by Wol (subscriber, #4433) [Link] (2 responses)

To quote Linus (from memory, not verbatim)

THIS IS A TOY, NOT MEANT TO BE ANYTHING BIG.

Cheers,
Wol

LWN's guide to 2024

Posted Jan 3, 2024 16:19 UTC (Wed) by elenril (subscriber, #165899) [Link] (1 responses)

That is entirely orthogonal to my point, which is that trying to do too many things at once usually results in not doing any of them particularly well.

LWN's guide to 2024

Posted Jan 3, 2024 21:20 UTC (Wed) by Wol (subscriber, #4433) [Link]

> trying to do too many things at once usually results in not doing any of them particularly well.

(a) Who cares? and

(b) People who do this "Just for Fun" are usually clever enough to be "Jack of all trades, MASTER OF MOST".

If they weren't clever, they wouldn't *want* to do it.

Cheers,
Wol

LWN's guide to 2024

Posted Jan 4, 2024 1:22 UTC (Thu) by JoeBuck (subscriber, #2330) [Link]

Yes, Linux succeeded because GNU, BSD, and X had already done the work of building a free Unix-compatible userland and development tools. The kernel was the last piece (and an extremely important one). But without all of that work, it would not have been immediately useful.