Doing a paper on OSS, would like input

So doing a paper on how OSS took over, both in phrase (as in most people say Open over Free) and in application for most major companies, governments, etc.

Specifically I am going to talk about Stallman philosophy vs Raymond’s, and why Raymond’s more pragmatic rhetoric won out (and how)

Going to talk about Red Hat, but would love to hear thoughts of where OSS won out over proprietary and examples of companies other than Red Hat that use OSS to great effect (as in making money or filling a niche).

This is a paper for normies, and its not super long to go very deep (only 17,000 words).

I’ll post some of my paper if anyone cares (when I get around to writing it)

4 Likes

Well for normies go into about how large portion of Google’s and Apple’s software is OSS. Firefox Open Broadcaster, Blender and VLC are mainstream so those are good topics. Especially the transition from Netscape to Firefox and ESR’s influence.

Talking about MS’s OSS initiatives would be a nice way to cap it.

Of course, the plucky Swede and his kernel will have to be mentioned.

2 Likes

A few quick thought watching this transition happen from Solaris/AiX/HPuX, etc… to Linux…

In the Chip DV world, this transition began largely as a function of what is now called devOPS people and/or individual engineers who had now been using linux for a few years on their own machines porting various compute intensive things to linux and discovering that very cheap linux boxes often performed as well or significantly better than solaris, HPuX, Aix boxes IFF (if and only if) you could get your software running on them which meant getting tool vendors to provide proprietary tools to do so).

I guessed at the time they were seeing the same things internally as many of the tool vendors started providing linux builds for their tools. Almost always RedHat based builds at the time (and now).

This was before the death of Sun/Solaris was obvious… but it didn’t take long before the overwhelming improvement in throughput is laughably low cost of the computes compared to Sun’s offerings caught he attention of higher-ups.

There was then and still is a wide variety of linux builds all requiring different glibc, glibc++ and various libraries. So, early on, RedHat was the “safe bet” as most proprietary tool vendors would at least provide support for that. The universe of distros is still complex enough that bigger companies want the “Safe bet” for compatibility and someone to call and yell at when things go wrong… so, the RedHat’s of the world fill that role.

Over the span of a few years, I watched Solaris go from a primary target build for tools to a second to linux. It was rapid in the chip design world. As a tool developer, I also watched SunOS/Solaris support drop off… man-pages went stale, problems developed in very expensive proprietary compilers, GCC began to overtake in performance, etc…

The dot-com bubble burst fundamentally changed the world in terms of budgets, so linux just accelerated form there as the low cost option even at Fortunexxx companies from what I saw…

At every stage and still up to now, the various licenses out there, and the sheer volume of them has been an obstacle. The lawyers don’t know what to make of them and just see infinite risk. So, they have to be dragged kicking and screaming to sign-off on compliance at companies where such things are taken seriously… .

There have been various “scares” in the industry when the courts got involved in GPL. Which is why the lawyers are and were scared. No one knew what to expect the courts to do with the various “viral” licenses and and the like. An assertion of a license is a known thing… a disclaimer of a license is a known thing. GPL brought a new type of idea of imposing restrictions on not just what you do with your ideas, but derivative ideas that linked either statically or dynamically.

There was FUDD circulating at times that any linking would trigger GPL, dynamic or static…

Now where we find ourselves is a lot of corporate lawyers will allow the use of many of the licenses out there (apache, mit, bsd, mozilla, etc… ) but specifically say, “NO GPL!”.

They will also insist that you link dynamically not statically.

The risk of your product being dragged into the public domain by virtue of linkage to GPL is taken very seriously and frankly an impediment to usage of software that uses this license to this day.

There are various companies like RedHat and Tolltech (bought by nokia and perhaps even someone else now owns Qt???) that try to provide a license “clearing house”, but the end result as far as I can see is that its still a mess…

The difference is that the economic forces driving re-use are compelling enough with Open Source that companies are willing to put up with the hassle. I suspect there are many smaller companies who play fast and loose with this stuff. Now and then you hear about company X getting caught doing so. Large publicly listed companies generally don’t play around with such things.

1 Like

You can’t ignore Android. So many embedded devices using it now.

3 Likes

Surprised you didn’t bring up TiVo and how it cause RMS to serg out and make GPL3

2 Likes

That was among the examples of companies getting caught “time to time”. I forgot about that one specifically.

Open source license compliance is not trivial even for the well intentioned. When you build against QT/chromium for example here are 30+ different licenses you have to evaluate each with different requirements for compliance, disclosure, etc… you pay QT a developer license, in theory to isolate yourself from this hassle, but they only get you part of the way there.

When you get into the corporate world where everything has a check box you run into a few problems:

  1. Do the lawyers understand the engineering sufficiently to achieve compliance
  2. Do the engineers understand the law sufficiently to achieve compliance.

I can tell you first hand, there is a lot of ambiguity, frustration and confusion with larger packages. I think a lot of people on the open source side of the equation think only of their individual thing and it’s license. Cobbling together a license disclosure and adequate disclosure of code that must be disclosed is a lot of work.

That’s ignoring the culture clash of trade secrets with open source. Engineers like science, so they are bad at secrets many times. So, having them interact with the outside world often makes the lawyers cringe to begin with.

1 Like

I wont, I used it as a source when I presented my paper idea (cuz I forgot a third source that day haha). It’s a great point for the paper though (I know I left it out in the op)

idk but I was going to add IOS in it. Though I think it’s a harder sell but apple’s IOS iirc is a modified UNIX derivative.

These are all good points peeps, thanks for the feed back (I forgot about TiVo and RMS @lesser lol that was gold) and thanks for the personal insight @cekim.

I would love to hear @admindev’s, and @rgk’s perspective of open source software too (specifically in the practical business side and not in philosophical terms).

I dont know how much iOS and OSX share in common.

Theres also the stuff sun did before oracle acquired them :troll:

1 Like

I’m on mobile, give me a few hours and I’ll write something up.

1 Like

Google. Google has profited off of the OSS capabilities for a long time. They spawned what is now known as Kubernetes, a container orchestration platform.

Before that they used Java, Python, and JavaScript frameworks for their platforms. I don’t know about licensing anything to make money but they made a fuck ton.

HashiCorp is another company that makes money off of open source. Their big hitters, Terraform, Vault, Vagrant, and Consul are all open source software (you can view it all on GitHub under HashiCorp). But they have enterprise and licensed variants. They went from $0 a year to several hundreds of millions in less than two years.

1 Like