Linux at 20, some personal memories (2011) | Hacker News Reader

Linux at 20, some personal memories (2011)

liw.fi

113 pointsadamnemecek17 comments

> I told Linus about rn, and he liked it too. We both spent way too much time reading newsgroups. Neither of us posted anything that year, though. There were big warnings against wasting other people's time, and we took them to heart.

Oh, the good old days. And yes, I see the irony.

It's cool that Linux started as a terminal emulator so that Linux could read Usenet from home: a terminal emulator and modem interface were within the grasp of a bright college student in the mid-90s.

What worries me is that the necessary software stack to do anything interest today is so heavyweight that we'll never see another great hobbyist OS. You don't just have to support WiFi and/or Ethernet, IPv6 and/or IPv4, TCP, parse HTML & display HTML (that's all doable, more-or-less, by someone bright and willing to compromise): you have to have graphics, you need a DOM, you need JavaScript, you need all the fiddly browser APIs, you need CSS — because the MVP for a web terminal is _everything_.

The MVP for an old fashioned terminal was 'read & write to modem, display text on screen.' You could cheat and not even offer paging at first!

Or you could go the TempleOS route and pick what you want to support because they're The Right Choice.

My focus lately is on a VPS over ssh: http://akkartik.name/post/mu. Since that article we're trying to add OpenBSD to that list, as a pretty accessible yet full-featured stack to hack on. Stephen Malina has good instructions here for getting started: https://github.com/akkartik/mu/wiki/Installing-OpenBSD-on-a-...

That's true in any field though. Yet people are still able build their own cars, planes, design/fabricate their own chips and so on. I think its primarily because our tooling has also advanced allowing you to do stuff in your garage that previously could only be done at some factory. To take your example, all of those things are not really that hard. Mainly because a lot of those things are established standards (using the word loosely here, as with anything to do with the W3C) with existing implementations whose source is available. Its up to you to decide how much of the standard you would have to implement for a MVP.

I dunno if that's true. There's plenty of projects doing baremetal onto Raspberry Pi or other ARM SoC boards... stuff I've personally put together started to venture into OS territory. Once you have POSIX, porting existing software isn't that bad.

And there are web browsers like NetSurf that can run on very simple or strange or stripped down systems. I've seen NetSurf running on one of my 8mhz mc68000 Atari ST systems.

> Once you have POSIX, porting existing software isn't that bad.

Yeah, but the really exciting stuff won't be POSIX. I definitely don't think it'll run C (well, of course someone will eventually port a C compiler to it, but the performance might not be great).

I think a really exciting new OS would try to do something radically different, not just reimplement what's been done before (which, to be fair, is what Linux does and why Linux is successful: if Linus had implemented Plan 9 instead we'd never have heard of him).

FWIW, NetSurf definitely sounds interesting, but a) it's still written in C and b) it doesn't appear to support JavaScript in production yet. I personally don't care, but all the hip kids these days keep on telling me that the only way for someone to read an article online is to download & execute hundreds of megabytes of obfuscated executable code of dubious origin.

When introducing Linux with my enterprise customers at the time, the chorus of "are you insane", "this will never work", and "this will never last" was deafening....

Had the same. The technical director of the company I worked for was saying that it was all going to be commercial windows and x86 in 2020. I said unix, open source and ARM. I think we had a £20 bet on it.

I had a manager at the IBM Global Services company I was contracting at breathing down my neck for even using Linux on small cheap one-off tasks on old hardware, as late as 1998. The argument was we should be using a licensed SCO Unix. Mind blowing.

> "For a while, Linux did not quite do everything right with networking, so it was banned from the university network. That gave Linus a lot of motivation to fix it, of course"

Sounds like there's another interesting story there regarding why it got banned. Anyone any wiser?

Most likely it was doing something naughty in it's network stack.

Yeah but that could have been any number of things.

I had worked that much out for myself. What with it being explicitly stated in the quote that the network stack wasn't working correctly and had to be fixed. :P

The bit I was interested in was what was Linux doing wrong and how was the fault detected? Did it take anything else down with it? Was the fix something straightforward (ie it was a silly oversight that caused the problem) or something complicated?

It can be interesting to read stories about tech going wrong. Sometimes because the results can be amusing in retrospect (less so at the time of the problem) and sometimes because it's just nice to read that problems and mistakes can happen to other people too.

Just a guess, but I vaguely recall something with linux in the early 90's. Because no other end devices (PC's and the like) would advertise things like RIP, many of the routers on campus and corporate networks would accept route announcements from any device. They assumed any device that could do these things were legit by default.

So, a misconfigured Linux box could hork up the network pretty quickly. Of course, that's just a guess.

The article is far more interesting than the disclaimer at the beginning suggests. BTW, I had it read to be by a speech synthesizer at sort of high speed while fully conscious.

Maybe add an (2013) to the post title.

Reply on news.ycombinator.com