Slashdot Mirror
ESR Advocates Proprietary Software
mvdwege writes "Apparently, Eric Raymond has decided that proprietary software is now a good thing, according to The Register. I must say it is rather revealing how easily he is willing to compromise on this particular freedom. Is his earlier vocal proclamation of the importance of freedom (still visible on his homepage) mere posturing? And if so, how about his vocal support of other freedoms?"
ESR, Eric S. Raymond, is not associated with "FOSS". FOSS is a term used when one wants to give credit to both the Free Software and Open Source movements without favoring either. ESR is a proponent of the Open Source movement and one of the people who started the Open Source Initiative over a decade after the GNU Project and the Free Software movement had been going.
The Free Software movement advocates exclusively for free software because only free software respects users software freedoms (the freedoms to run, inspect, share, and modify software). The Free Software movement examines these issues in terms of ethics, speaks to all computer users, and takes a far broader view than the Open Source movement which never discusses user's freedoms and examines these issues in terms of a developmental process that is chiefly aimed at businesses.
The OSI has given a remarkably disrespectful view of the differences between the two movements, reducing the difference to "ideological tub-thumping" in their FAQ. The Free Software Foundation has a far more informative and respectful view in an essay on the differences between the two movements.
Digital Citizen
His belief is predicated upon the notion that the move to 64 bit computing means people are about to make choices about the next generation of operating systems.
Then this writing fails. 64-bit OSs as overrated and overhyped. The move from 16- to 32-bit was dramatic. A lot of people, including those around here, seem to naively believe the move from 32- to 64-bit will be a similar event. It will not. 64-bit will be meaningful for some servers and some other high end applications, for the rest there will be no appreciable immediate benefit. *IF* Joe User gets all excited about 64-bit in the near term it will probably be due to a successful Micorosoft marketing campaign designed to artificially create an upgrade cycle. Barring this there will be a slow migration to 64-bit as Apple and Microsoft make the 64-bit versions of their OSs the default version, not an optional upgrade. In other words Joe User will get 64-bit when he happens to buy some distant new machine (4-5 years ?). The near term upgrades and build-to-order options will be a minority. I'll do it, I'm a programmer, I want my code to be 32/64-bit clean.
POSIX and Windows are both operating systems that use hardware memory management to seperate otherwise unsafe (written in liberal mid-level languages like C) processes from one another, using a security model based largely upon user ownership. If a large number of jobs have to be done that require communications between the different programs performing the different jobs, a single program - running in its own memory space - is generally written that manages all the jobs, rather than the jobs being split out one-per-program.
Programs are loaded from files on disk, from a simplified file system that uses hierarchical name management and that's based upon arbitrary length binary files. From the point of view of the user, there are running programs and files.
This is essentially classic 1970s computer technology. In terms of where it became the standard for system design, that's roughly when it dates to.
There have been serious attempts since then to reform that model, but generally they haven't gone anywhere. The 1980s was full of simplified systems where all processes ran in the same memory space - Windows, Mac OS 6+, AmigaOS, Sinclair QDOS, etc. At the other end of the spectrum, there were many abortive attempts to break up operating systems into more simplified units protected from one another. None of these design changes are present in Windows, Mac OS X, or GNU/Linux, either the simplified or the microkernel strategies (with good reasons for both.)
The 1990s saw the beginning of managed code. While this probably does represent the future of computing, we're not seeing it yet. As yet, managed code only exists in mainstream operating systems running as high level processes at the same level as other ordinary user applications. Mac OS X, Windows, and GNU/Linux do not use managed code, they merely support it.
File systems have been reformed several times. DEC VMS supported native rich file types with record indexing. This has yet to appear anywhere else. Mac OS introduced forks and added creator and type information to the file system. While present for legacy reasons, Apple has deprecated support; Microsoft technically supports file forks in NT but has made no effort to use them or encourage their use; GNU/Linux has only recently started to support additional metadata, and the feature is barely used. Systems like Smalltalk, NewtonOS and PalmOS blur the differences between files and other objects. Achieving minor success on PDAs, their approaches have yet to really have any serious impact.
Essentially, Mac OS X, Windows, and GNU/Linux, are the latest, most optimal, versions of what you saw in the seventies. That's not a terrible thing, but 64 bits gives us the opportunity to rethink why we're programming the same way we did thirty years ago. In particular, the combination of managed code and the massive 64 bit address spaces gives us a chance to revisit the question of how we can most efficiently prevent operating system and application components from treading upon one another, and how we can keep the system secure.
You are not alone. This is not normal. None of this is normal.
It's fair to say that Linux might well not even exist without the work that RMS and his cohorts did in the first years of the FSF's existence.
It's also fair to say that it's NOT true that if RMS hadn't done what he did that someone else would have. It is not to be taken for granted by anyone that without RMS & FSF, sooner or later we would have ended up in essentially the same place we are today.
I know what it's like to have to get a company's permission to write software on their computers, and to pay them a LOT of money for the 'privilege'. NOT FUN. RMS has changed all of our lives in a way that we can only understand by knowing the history and by sitting back for bit and actually thinking about it.
I can't say that for ESR. All he ever did for me was threaten me for using his US service mark 'open source' on my web site, a service mark he didn't actually have. I find it easy to ignore him.