Hacker News new | past | comments | ask | show | jobs | submit login
GNU/Linux Distribution Timeline (futurist.se)
91 points by chaimkut on Mar 24, 2014 | hide | past | web | favorite | 57 comments

I maintain a specific timeline for Debian if anyone is interested: http://timeline.debian.net

That's great. Made me go back to to my early debian days with Potato CDs from the bookstore and a translation of "Debian GNU/Linux: Guide to Installation and Usage". I Could'nt update to woody because it was too costly over the dialup line, so i bought the Woody CD set together with a Friend. Every time i had to install something i had to walk to his house and get the CDs, or pay lots of money (for me as a teenager at that time) to download the files.

Maybe you could add another Frame with Versions of Popular Packages? I'm Thinking KDE, Gnome, GCC, mayor scripting Languages Browsers and Editors?

> Maybe you could add another Frame with Versions of Popular Packages? I'm Thinking KDE, Gnome, GCC, mayor scripting Languages Browsers and Editors?

That's interesting - to give more context to particular eras? I can certainly "place" myself in life when using GNOME 2.8 :)

yeah, to give more context. icons of the programs at that time would also be great.

Sent you a couple of fixes.

Oh - how? I haven't received an email/PR/blah.

Apparently I didn't send the PR, sorry! I've done it now.

I wish the dots were scaled by some metric indicating popularity. Such a metric is hard to come up with. Maybe total number of commits.

Way awesome though.

You already see some kind of popularity – just count the number of forks.

But would you say Slackware is more popular these days than OpenSuse?

I think number of forks is not a very precise metric.

Forks are also an indicator that people weren't happy with the origin.

It kind of depends on the forks. Many are just a different set of default installed packages plus a new wallpaper.

Or that people were really happy with the origin..

That's like saying countries that based their "constitutions" off of another countries did it because they don't like that countries constitution. What's more likely is that they really like it but just wanted to tweak something.

Perhaps number of downloads? Number of mirrors?

The page hit count on http://distrowatch.com/ is sometimes used as a popularity proxy. What do you reckon?

That it's horrible as a popularity index and should never be used as such. The number of visits is very low and the population which visits the site is very selective.

Instead use data from highly popular sites like wikipedia: http://stats.wikimedia.org/wikimedia/squids/SquidReportOpera...

I don't think that's really accurate enough either. For example that page says iOS 1 is the most popular version of the OS, way ahead of iOS 7.

This is beautiful. I'd like posters for my walls. Does anyone else know of similar diagrams suitable for printing?

It's a nice svg so I guess you couold turn it into a PDF with a converter.

SVG to PDF in google?

Outdated and misleading.

For instance, Korora is based on Fedora, not in Gentoo.

This is awesome, reminds me of this pngs I've been saving for a while (don't remember the source, but is probably written on the pngs somewhere)



This is from 2012 and missing the (2012) part in the title...

Agreed. Two things that I know are out of date:

No SteamOS (Based off Debian)

Backtrack was rebuilt and is now Kali

I'm sure there is other stuff.

Such a waste of time and manpower creating all these forks. I wish folks spent time scratching their itch by writing great software for Linux instead of trying to reinvent the same wheel. And yet folks will continue to fork platform trying to solve problem that doesn't exist.

Since someone intentionally created each fork, doing so was demonstrably not a "waste of time and manpower". If people perceive a problem, it demonstrably exists; whether or not other people also experience the same problem is scarcely relevant.

Many of the forks weren't/aren't serious anyway and the majority died out pretty soon, they shouldn't be seen as competition or wasted time. If you just want to experiment and don't want to pay attention to some community guidelines etc., then forking is better. It is exactly this freedom I love about Linux or open source in general.

A distro with enough options to satisfy the creators of every fork would be an unholy mess. 1,000 separate contributors are very good at adding every option under the sun to a piece of software. But reducing choices to the point where you can just burn a CD and "install Debian" is anti-parallelizable.

I think 10 different distributions are fine. 1000 distributions are not. Fedora is good for Gnome, Ubuntu is good for Unity, OpenSUSE is good for KDE, Arch Linux is good for minimalism, etc. All these distributions are good at something radically different which is fine but I want to use a different theme, let's create a new distribution or some driver didn't work on my system, so I am going to create a new distribution is hardly a valid reason.

For me freedom means two things (1) contribute to up stream project (2) write free software.

Like is not happening in every other platform.

Yeah, I don't know if Mac or Windows have 1000 forks. Do they?

The chart clearly shows that there is a lot of proliferation, which is confusing to a lot of new users. Imagine what would be possible if all these fine developers would focus on two or three distributions.

Ps. I am proud that I was a paid co-developer of the very first distribution listed ;).

>Imagine what would be possible if all these fine developers would focus on two or three distributions.

You don't think the lessons learned by those developers' experimentation weren't shared and applied in the decision-making process for more prominent distros? In a way, creating a distro is like contributing to all the other distros at the same time. What makes them fine developers is their drive to innovate and try new things so that one day the status quo can improve. Who knows: if all those developers worked on the same few distributions, they may have crashed and burned.

You don't think the lessons learned by those developers' experimentation weren't shared and applied in the decision-making process for more prominent distros?

Sometimes. But there is also a lot of 'let's take Ubuntu and change the default desktop/wallpaper/DVD package set/...'.

Also, it's not as if developers in major distributions such as Debian don't have room to experiment with different approaches. For instance, in Debian, the default init system discussion rages. The most promising candidates are already available in Debian itself.

Well, for any of us old enough, this is the revival of UNIX wars.

Not quite. The UNIX wars were basically "Hardware" wars. Basically, you bought your workstation from Sun, or HP, or SGI, or IBM or whoever, and you got the version of UNIX that was shipped for that hardware. You could compile your own BSD (/usr/ucb) or GNU (/usr/local) utils if you didn't want to pay a king's ransom for a compiler, but basically, the idea of running any type of software on any hardware, a la Linux or OpenBSD, hadn't really caught on.

I know it's cool to be all blase about this, but this is different.

That is only part of the picture.

Many of us bear scars from trying to write portable software across commercial UNIX systems, while discovering that POSIX only covered part of the required features, and even the POSIX ones had implementation defined behaviours.

Not to mention the wonders of each system having their own UI toolkits, besides each XWindows flavour.

I did it between AIX, HP-UX, Solaris, Xenix, DG/UX.

You were a paid developer on MCC?

I put this chart as the example on https://en.wikipedia.org/wiki/List_of_software_forks

What tool was used to create this graph? Does anyone know?

I often find the need to create such a diagram.

I am hoping the answer is not Photoshop.

The comments in the SVG file indicate that it was made with Gnuclad: https://launchpad.net/gnuclad.

Wasn't RedHat originally a Slackware derivative?

What about all the distributions that don't include GNU, or those that only include a tiny proportion of GNU code, is there a chart that includes those?

What about all the distributions that don't include Linux, or those that only include the kernel but has nothing else in common. What about distributions that has more than one kernel.

So instead, lets call it distributions! Except, distributions of what?

Maybe we should call it distributions of Unix compatible operative systems. Except, Unix is not a standard so then one have to ask, which unix version and what does compatible means?

Maybe we should call it distributions of POSIX compatible operative systems. Except, then we got the problem that most distributions are not 100% compatible.

Maybe we should call it distributions of partial POSIX compatible operative systems. Except, that is not a specific definition, so now you can't graph it.

I am sorry, I do not think there is a correct name we can use. I also do not think this is a discussion relevant to the article.

The only essential GNU software is GCC, and even GCC can be replaced with Clang for many applications nowadays. Let's not call it GNU/Linux - it's just silly.

They don't actually exist - it's just hot air from the anti-GNU crowd.

Did you mean this comment to be sarcastic? Android and a lot of embedded Linux distributions are not GNU/Linux.

Android doesn't have glibc, and it's the most popular consumer computing platform.

It's the most popular platform for mobile devices; the most popular consumer computing platform is still Windows.

Got any links for that? Everything I've read for the last couple of years says mobile/tablet has overtaken desktop dramatically.

What have you been looking at: sales figures or install base numbers?

Usage on consumer web.

Link? Stats linked elsewhere in this thread show Windows ash having about a 45% share of useragents visiting Wikipedia, with iOS having less than half that.

I couldn't find kylin linux.

Cool :)

Where does the source data come from?

The same one: https://commons.wikimedia.org/wiki/File:Linux_Distribution_T...

Not able to find source though. Anyone?

Edit: More info http://futurist.se/gldt/, still no source though.

The tar.bz there has a CSV file which looks pretty sourcy to me.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact