LWN: Comments on "25 Years of Linux — so far"

25 Years of Linux and the World Wide Web.

anselm — Wed, 07 Sep 2016 08:48:10 +0000

Which would have been a bit silly given that one could write a basic Gopher server in half an afternoon and it would take perhaps three screens' worth of code.

I think one of the main advantages of the WWW over Gopher – at least with the advent of the likes of NCSA Mosaic – was that it made page-embedded hyperlinks and graphics a thing (although technically you could argue that these were features of HTML and it would have been perfectly possible to serve HTML over Gopher, which later versions of the Gopher protocol in fact allowed). Other drawbacks of Gopher included not having an equivalent to URLs, and only a very limited selection of allowable content types. In the end, Gopher was really too primitive to win.

25 Years of Linux — so far

mathstuf — Wed, 07 Sep 2016 01:16:55 +0000

Huh. And now we have plugins which do that (ignore front-page, skipping JS completely or block subsets based on URLs, etc.). Indeed it does make for some…interesting renders, but on the whole, I like it better.

25 Years of Linux — so far

flussence — Wed, 07 Sep 2016 01:12:47 +0000

Well, it was missing most of CSS 3. And it didn't live to see HTML 5 finally standardise the parsing of tag soup. And I think I had to file a bug about it sending malformed *HTTP headers* once.

It's pretty dire on today's web, I'll admit (even NetSurf's doing better). But it was pretty good back when Opera couldn't put CSS opacity and text anti-aliasing on a page at the same time, Firefox injected parts of its browser UI into the page's own DOM, and IE... existed...

25 Years of Linux and the World Wide Web.

chfisher — Tue, 06 Sep 2016 21:49:21 +0000

Part of the reason the WWW took off instead of Gopher is that CERN gave the WWW stuff away for free, whereas the Univ of Minnesota (authors of Gopher) attempted to monetize it by selling Gopher server licenses. Oops

25 Years of Linux — so far

zlynx — Tue, 06 Sep 2016 19:45:36 +0000

As I recall it, KHTML got that speed by simply ignoring large chunks of HTML and CSS specs that would have slowed it down. KHTML's page rendering was often -- interesting.

25 Years of Linux — so far

flussence — Tue, 06 Sep 2016 17:06:16 +0000

KHTML doesn't get anywhere near as much credit as it deserves. I'll sorely miss having a browser that had an option for middle-click autoscrolling, *and* could do said scrolling at 60fps. On a uniprocessor machine ten years ago!

(Okay, maybe Opera could've done the same at the time, but I couldn't stand that non-native UI.)

25 Years of Linux and the World Wide Web.

zlynx — Tue, 06 Sep 2016 16:40:41 +0000

A lot of the other "wonderful" hypertext ideas require either a single big server or cooperation between web servers.

Bidirectional links with consistent metadata don't just happen.

But with HTTP I can just create a link to anyone's web page. Whether they like it or not. And they can rename that web page whether I like it or not.

Other, better systems require far too many coordinated moving parts which is why they fail.

25 Years of Linux and the World Wide Web.

anselm — Tue, 06 Sep 2016 09:58:20 +0000

The World Wide Web is to distributed information retrieval systems essentially what MS-DOS (and its successors) are to operating systems. It works reasonably well for what it set out to do, it is very widespread, many people seem to like it, and by now, even though its shortcomings are well-known and there have been additions to the additions that were made to attempt to mitigate them, it is quite impossible to get rid of. It managed to get where it is now not because it was technically excellent or otherwise compelling but mostly because it was there at the right time when somebody (or somebodies) important was looking for something like it, and the other stuff that existed (remember Gopher?) for whatever reason wasn't interesting enough.

I remember attending a conference in the 1990s where a guy from a university in, I think, Austria gave a long talk which essentially amounted to whining that everyone was using this terrible WWW and nobody was interested in the much more clever and powerful system that his research group had invented. Didn't make a big impact, though.

25 Years of Linux and the World Wide Web.

tialaramex — Tue, 06 Sep 2016 00:10:17 +0000

Exactly, Tim _shipped something that worked_

I was studying and later working at a major hypermedia lab in the 1990s. Without a doubt some of the research work was on things that the World Wide Web still doesn't really do today. But unlike our research work, Tim's half-arsed "Web" had users out in the wild. A research project would scope out a small problem, solve some of it, and write up a paper. If you were lucky some of the code making it work lasted beyond the time taken to get published, and if you were _very_ lucky that code was actually available from an FTP site somewhere. But by the time I arrived even as an undergraduate, there were more pages on Tim's World Wide Web than in any of these research systems, and by the time I became a postgrad they were writing up the new research systems _on_ Tim's World Wide Web, and at that point it becomes obvious that you're basically just wanking. Even if Tim's system was much worse in a whole variety of ways than the prototype research projects (it was), Tim's system existed and the prototypes were just prototypes. By the time I ceased to be a postgraduate student Tim had an office in the building where I had studied and his name was highlighted above all the clever people who'd been designing "better" hypermedia systems while the World Wide Web was taking over the world.

25 Years of Linux and the World Wide Web.

nix — Mon, 05 Sep 2016 21:33:12 +0000

Tim Berners Lee's only real invention with the Web was the hyperlink.

Ted Nelson (the coiner of the term 'hypertext') was talking about unbreakable bidirectional hyperlinks (and not just point-to-point hyperlinks, but links to and from *regions*, with optional annotations and other attributes) decades before.

But Tim actually got it working on a large scale. :)

25 Years of Linux — so far

pboddie — Sun, 04 Sep 2016 14:10:29 +0000

The community was singularly unsuccessful at creating a proper web browser for Linux until the collapse of Netscape jump-started the development of the tool we now call Firefox.

That's rather unfair on those who developed KHTML - the basis of the decent-enough Konqueror - which then evolved to become WebKit, making possible the tools we now call Chrome, Chromium and (if it is still around because I don't follow Apple) Safari.

25 Years of Linux — so far

Wol — Fri, 02 Sep 2016 11:28:36 +0000

Plus infinity!!!

Sorry for the noise, but yes, she did an awesome job!

Cheers,
Wol

25 Years of Linux — so far

geek — Thu, 01 Sep 2016 22:38:18 +0000

yes, indeed! who could have imagined the great job Pamela did on groklaw?

25 Years of Linux and the World Wide Web.

Wol — Thu, 01 Sep 2016 05:06:47 +0000

> I shudder to think where we would be without these two inventions of August of 91.

In a similar place to where we are now?

Tim Berners Lee's only real invention with the Web was the hyperlink. That's not to decry his hard work - but he basically built on what was already there.

And instead of Linux we would have the BSDs. That world nearly happened ...

Really, it's the people that matter, not the inventions. Linus, with his excellent project management skills and readiness to accept contributions from anyone, and TBL with his open Scientific "giving" frame of mind.

Oh - and the lack, back then, of all these attempts to stifle creativity by patenting mathematics (software).

Cheers,
Wol

employer/employee copyright

giraffedata — Tue, 30 Aug 2016 02:27:21 +0000

I believe those are exactly the points I made. We made up copyright law and arbitrarily decided that by default it belongs to the employer (I see "naturally" could be taken another way; I just meant that with the law in place and no further action, the employer has the copyright), so for the employee to have it, the employer has to do something to give it up.

That's the opposite of the impression a reader would get from the phrasing, "employees should demand to keep their copyrights in Linux." The word "keep" implies that the copyright starts off belonging to the employee and the employee should insist on keeping it that way. If we want to give the correct impression (unless, again, we're trying to make a political point about truly natural rights being different from the ones created by law), a better wording would be, "employees should demand to own the copyright in their work on Linux."

employer/employee copyright

pabs — Mon, 29 Aug 2016 10:08:01 +0000

I don't think there is anything "natural" about copyright, nor about employers owning it. As well as copyright transfers, the "employer owns copyright" rule can be overridden by the employment contract in at least the US and the UK:

http://www.copyright.gov/title17/92chap2.html#201
https://www.gov.uk/guidance/ownership-of-copyright-works#...

employer/employee copyright

giraffedata — Sun, 28 Aug 2016 02:42:12 +0000

... reasons that Linux developers should demand to keep their copyrights in Linux instead of giving them to their employers.

Maybe you're trying to make a political point with this wording, but it's misleading. In the US, when someone pays someone else to write code, the copyright naturally belongs to the employer and the employer would have to give it to the developer for the developer to have it. Copyrights are man-made rights and the people who created them defined who's rights they would be.

25 Years of Linux — so far

halla — Fri, 26 Aug 2016 19:15:00 +0000

Yes... We're currently having a Krita sprint in Deventer, and one of the contributors who is doing a ton of great coding work... She just realized she was one year old when Krita started :-). But it's good to have a good age spread, in this project we've had twelve year old people contributing code, while I'm definitely getting geezerish!

25 Years of Linux — so far

sjj — Fri, 26 Aug 2016 18:20:38 +0000

It is kind of trippy to work with people for whom Linux and the Internet have *always* existed.

25 Years of Linux — so far

robbe — Fri, 26 Aug 2016 11:55:36 +0000

> As we have seen with projects like OpenSSL or GnuPG, it's not just
> the obscure projects that fall by the wayside; important
> infrastructure can also go without support.

I don’t think the amount of resources poured into these projects is that much lower than it was 10 or 20 years ago. It’s just that /many/ more people, companies, and projects rely on them.

This is not the only place where the „crumbling infrastructure“ problem exists.

25 Years of Linux and the World Wide Web.

mattrose — Thu, 25 Aug 2016 15:08:51 +0000

About two weeks ago, the 25th anniversary of Tim Berners-Lee's announcement of the WWW happened. I would argue that though nobody knew it at the time, the events of August of 1991 was to change the world for the better more than any other single month in history. I shudder to think where we would be without these two inventions of August of 91.

25 Years of Linux — so far

JFlorian — Thu, 25 Aug 2016 13:17:57 +0000

That sounds like a great suggestion for a future series of articles. I'd love to learn more about such endeavors. Important history is being made here (IMHO) and who better to capture that for future generations than our beloved LWN.

25 Years of Linux — so far

gnu — Thu, 25 Aug 2016 08:08:51 +0000

Excellent point, jdub! There are so many of them who can't be found via `git blame' but worked tirelessly on various projects.

25 Years of Linux — so far

jdub — Thu, 25 Aug 2016 04:01:14 +0000

> In general, a lot of the outcomes that people feared have not come to be.

In an otherwise excellent piece, I felt that this section skipped an important point: Many of those outcomes did not come to pass thanks to the non-technical contributions of some pretty surprising people.

Activists and researchers, lawyers and managers… there are so many people who found this world through shared politics, principles, and social connections, or were simply reassigned to a new job and got excited. Some of them have contributed more than they'll ever let on (corporate politics), did invaluable work we couldn't fully understand (pro bono legal expertise), or simply gave their staff time to get involved and do things right (enlightened management).

If you have ever worked on a major FLOSS project, you'll know people like this. You won't find them in git blame, you may not even find them in mailing list archives, but their contributions are still felt today.

25 Years of Linux — so far

pabs — Thu, 25 Aug 2016 03:17:02 +0000

There are other funding models for open source and we are starting to see some organisations move towards them but it would be interesting to see how a Linus crowd-funding campaign could go.

Thanks for highlighting the reasons that Linux developers should demand to keep their copyrights in Linux instead of giving them to their employers. Anyone who is doing that might want to talk to SFC about having their copyrights in Linux enforced.

https://sfconservancy.org/copyleft-compliance/

25 Years of Linux — so far

felixfix — Thu, 25 Aug 2016 01:43:36 +0000

I patched a binary card deck ... once. Only got one bit wrong out of 960. Took a while to find it. Ugh.

25 Years of Linux — so far

neilbrown — Thu, 25 Aug 2016 00:52:08 +0000

> each of which had its own problems and none of which could be fixed by its users.

... excepting those of us driven enough to create binary patches... It was fun to do that *once*.

12 Years of Linux ONLY — so far (for me)

pr1268 — Wed, 24 Aug 2016 22:44:45 +0000

Well, then that means it's been twelve years since I kicked MS Windows to the curb, wiped the drive, and went Linux only. (I had been dual-booting since about 1998 or so.)

I think it's a nifty coincidence that I made this "major lifestyle change" on the anniversary of Linus' newsgroup posting. ;-)

25 Years of Linux — so far

Depereo — Wed, 24 Aug 2016 22:24:00 +0000

It's still exciting, and I'm looking forward to Rust on RISC-V, and Microsoft naming a kernel release "Hurr durr I'ma sheep".

I wasn't working on the kernel 25 years ago (maybe working on learning my multiplication tables), but GNU/Linux is still a passion project for me. If I get paid to work with the tools and systems that excite and motivate me, all the better - I hope that I can still do meaningful and useful work (that makes its way upstream) without having to do it afterhours.

A quarter-century of growth and success is a very impressive achievement for a software project, and as I get more and more experience I appreciate the challenges it has overcome more and more.