On Thu, May 26, 2016 at 5:19 PM, Jim Hall <jh...@freedos.org> wrote: > > I'll add a few comments here: > > Stallman holds a very strict view on "free software" because he > created GNU and the Free Software Foundation for ethical reasons. > Stallman didn't like that companies and organizations were no longer > sharing the source code to their programs, and were "locking out" > users from modifying these programs to fix bugs or to make the > software do something different. Stallman wanted all software to be > free as in "freedom."
Stallman started out working for the MIT AI lab. They were doing a lot of stuff in Lisp. It was the days before personal computers. Back then, Lisp Machines and Symbolics were selling dedicated hardware for running Lisp. Symbolics made some decisions Stallman disagreed with that prevented the sort of code sharing he liked, and when a new Symbolics release came out, he'd reverse engineer it and contribute the changes to Lisp Machines. The FSF and the GPL were Stallman's efforts to enforce the sort of code sharing he cared about. The GPL *requires* you to make source available for code you issue under it. The issue it encounters is that it's *viral*: if your code links against GPLed code, it too becomes GPLed. There are folks writing open source code who won't go near the GPL in consequence. They want their code *used*, and see the GPL as a barrier to adoption. > Raymond came at the topic from a different direction. He was helping > companies implement free software - either to use in-house or to > include in a product. Raymond kept running into problems with the term > "free." Companies would see "free software" and assume they would have > to give everything away at no cost. "Free" software was confusing. So > Raymond created the term "open source software" to clarify that the > source code was something you could see and even modify, but you > didn't necessarily have to give it away for free. Raymond was a director of several corporations involved in open source efforts, like VA Linux and Red Hat. It was the days when the dot COM boom was in swing, the Internet was a new paradigm, and stock prices would ascend continually without details like actual revenues and profits required. Eric once predicted he would be immensely wealthy when the dust settled on IPOs of companies where he was a director. And he *was* briefly worth about $48 million after IPO, but values dropped precipitously shortly after as the dot COM bubble burst. Eric had software tracking and charting the downward trend of his net worth and posting it for the public. He has a bit stashed away from those days, but is by no means wealthy. His wife Kathy is a lawyer, and has been mainly responsible for paying the bills, leaving Eric free to concentrate on non-paying open source projects. (Eric *is* being paid to write code these days - a consortium of the outfits that operate stratum 1 NTP servers are funding an effort to update and harden NTP server code, because it's become an attack vector. Eric is architectural lead on the new code. He estimates he's reduced the existing code base by half, and expects it to be two thirds by the time they are finished. The core code was heavily encrusted with special case code to support obscure hardware and protocols no longer used. He feels that reducing code size reduces attack surface, and notes that his code is not subject to 9 of 11 recently documented NTP server code vulnerabilities, because he's removed the stuff those creating those vulnerabilities.) > (Lots of companies make money by selling open source software. Red Hat > is one such example.) *NO!* Red Hat does *not* make money *selling* open source *software*. *Nobody* does. They make their money selling *support*. Red Hat's principal product is Red Hat Enterprise Linux. You can get the code for free, and the CentOS distribution is the open source version. It is identical to RHEL save for branding. You can get it and run it and not pay a cent to anyone. I've been a point man pushing open source in companies I've worked for. The first question corporate users ask is "Who do we call if it breaks?" They like commercial software because they can *get* support on products they've paid for. Red Hat makes it's living being who you call. If you run CentOS, and decide supporting it needs more resources than you have, you can apply an RPM that changes the branding from CentOS to RHEL, then call Red Hat and say "I'm running RHEL, and I'd like a support contract. What are my options?" Canonical, Ltd, Ubuntu's parent, is playing in the same space, and offers supported commercial versions of Ubuntu Server. Once again, the software itself is free, and can be downloaded and run with no charge. You pay for support. As a rule, people do not *buy* software you can get for free in open source form. There is nothing in an open source license that says you *can't* charge money for it, but good luck getting anyone to pay. You make money on *support*, not sales. > But Raymond also had a more permissive view to what could be > considered "open source software." Stallman didn't like that, and > thought Raymond was missing the point. They have never gotten along. It's a little more complicated than that. I've met Stallman. I've known Raymond from *before* he coined the term "open source". Eric has been a major contributor to Gnu efforts, and a good chunk of Gnu Emacs is his code. He's always willing to talk to Stallman. Whether Stallman is willing to talk to *him* is an imponderable, possibly connected with phases of the moon. One of Eric's recent efforts was converting the Gnu Emacs repository from CVS to git, with Stallman's enthusiastic support. (Eric has written tools to largely automate such conversions, and recently built The Great Beast of Malvern - a high end workstation with 64GB RAM - to be a conversion platform. It can convert multi-gigabyte repositories holding millions of lines of code in hours, instead of days or weeks. His goal is to eliminate CVS in our lifetime.) Open source developers have pretty much all migrated to git these days, and Emacs being under CVS was a barrier to entry by new contributors. There were 30 years worth of development history involved that needed preservation. (Preserving code diffs isn't that hard. Preserving *comments* on what was changed and why is a lot trickier, due to architectural differences on what comments are attached to.) > So while many people use "open source software" and "free software" > interchangeably, remember that Stallman does not think they are > interchangable. > > (I've somewhat simplified Stallman's and Raymond's views, above.) Open source folks tend to talk about "Free as in freedom", and "Free as in beer". For the developers and communicants of the open source religion, the critical part is "Free as in freedom". You can get the source to what you run. For the vast majority of end users *running* open source code, the critical part is "Free as in beer." They don't have to *pay* for it. The majority of *users* of open source products don't *need* the source, and couldn't use it if they had it. They aren't programmers, wouldn't understand the code, couldn't fix bugs or make enhancements, and couldn't reproduce the build environment and build a duplicate of the binary they got on their own machine. The just want free-as-in-beer code that does what they need and they don't have to pay for. > I usually describe "open source software" and "free software" as a > Venn diagram: All free software is open source software (all free > software requires that you make the source code available to others). > But not all open source software is free software (sometimes the > license for "open source software" doesn't meet Stallman's strict > requirements). That's Stallman's problem. The question is how much anyone else *cares* what he thinks. My impression is that increasingly few do. Every open source license I'm aware of assumes you will make source offered under it available. If you aren't willing to release source, you don't *use* an open source license. The differences lie elsewhere, like how the code may be reused. And every license I'm aware of other than public domain has the expectation that if you fork it to produce a closed source commercial variant, you will negotiate a closed source license with the original author permitting you to do so. You may not simply make a closed source fork. You *go* closed source commercial because you plan to *sell* the software you will create, and you cannot successfully *sell* stuff offered as open source. If you do that, the original author will generally expect a cut of the take, because you are making money off his code. > These days, I think "open source software" and "free software" are > pretty much the same. I use the terms interchangeably. The question becomes whether you mean Free as in Freedom, or Free as in Beer. :-) > To avoid running into problems, my preference is to include open > source software with FreeDOS. I don't necessarily draw the line at > "free software" the same way that Stallman does. Even a few years ago, > I wanted FreeDOS to meet Stallman's Free System Distribution > Guidelines (see Rugxulo's link), but these days it doesn't matter to > me. I don't care if it's "open source" or some GNU license. If the > program works, is open source, and let's others use it, I'm happy with > it. If you *only* wanted software in a FreeDOS distro that met Stallman's requirements, you might have problems actually making a distro. Too much of what you might like to include won't be under a GPL license. If you relax your licensing requirements, things become easier. A FreeDOS distro should consist of code issued under a license that lets it provide source as well as binaries. Precisely what licenses are used is a detail. Licensing incompatibilities affect whether code from one project can be reused in another. That's the developer's problem. It doesn't affect whether FreeDOS can offer it, or whether users can run it. The more practical question is whether a FreeDOS distro should *include* the source for what it offers along with the binaries and documentation. Offhand, I'd say it *shouldn't*. The source should be *available*, and the distro should state where and how to get it. But as mentioned above, most users don't need and can't use source. They just want binaries and docs. Why swell a distro with stuff the end user will simply discard or ignore? (And since the whole world doesn't have fast broadband, reducing the delivered size of a distro reduces the bandwidth required to get it., which may be a factor for FreeDOS users. A distro might exist as two parts - part 1 is binaries and docs. Part 2 is source for them. Everyone inclined to use FreeDOS gets part 1. The few who need source get part 2.) > Jim ______ Dennis ------------------------------------------------------------------------------ What NetFlow Analyzer can do for you? Monitors network bandwidth and traffic patterns at an interface-level. Reveals which users, apps, and protocols are consuming the most bandwidth. Provides multi-vendor support for NetFlow, J-Flow, sFlow and other flows. Make informed decisions using capacity planning reports. https://ad.doubleclick.net/ddm/clk/305295220;132659582;e _______________________________________________ Freedos-user mailing list Freedos-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/freedos-user
