A program is free software if its users have certain crucial freedoms. Roughly speaking, they are: the freedom to run the program, the freedom to study and change the source, the freedom to redistribute the source and binaries, and the freedom to publish improved versions. (See http://www.gnu.org/philosophy/free-sw.html.) Whether any given program is free software depends solely on the meaning of its license.
Whether the program can be used in the Free World, used by people who mean to live in freedom, is a more complex question. This is not determined by the program's own license, because no program works in isolation. Every program depends on other programs. For instance, a program needs to be compiled or interpreted, so it depends on a compiler or interpreter. If compiled into byte code, it depends on a byte code interpreter. Moreover, it needs libraries in order to run, and it may also invoke other separate programs that run in other processes. All of these programs are dependencies. Dependencies may be necessary for the program to run at all, or they may be necessary only for certain features. Either way, all or part of the program cannot operate without the dependencies.
If some of a program's dependencies are non-free, this means that all or part of the program is unable to run in an entirely free system--it is unusable in the Free World. Sure, we could redistribute the program and have copies on our machines, but that's not much good if it won't run. That program is free software, but it is effectively shackled by its non-free dependencies.
This problem can occur in any kind of software, in any language. For instance, a free program that only runs on Microsoft Windows is clearly useless in the Free World. But software that runs on GNU/Linux can also be useless if it depends on other non-free software. In the past, Motif (before we had LessTif) and Qt (before its developers made it free software) were major causes of this problem. Most 3D video cards work fully only with non-free drivers, which also cause this problem. But the major source of this problem today is Java, because people who write free software often feel Java is sexy. Blinded by their attraction to the language, they overlook the issue of dependencies, and they fall into the Java Trap.
Sun's implementation of Java is non-free. Blackdown is also non-free; it is an adaptation of Sun's proprietary code. The standard Java libraries are non-free also. We do have free implementations of Java, such as the GNU Java Compiler and GNU Classpath, but they don't support all the features yet. We are still catching up.
If you develop a Java program on Sun's Java platform, you are liable to use Sun-only features without even noticing. By the time you find this out, you may have been using them for months, and redoing the work could take more months. You might say, "It's too much work to start over." Then your program will have fallen into the Java Trap; it will be unusable in the Free World.
The reliable way to avoid the Java Trap is to have only a free implementation of Java on your system. Then if you use a Java feature or library that free software does not yet support, you will find out straightaway, and you can rewrite that code immediately.
Sun continues to develop additional "standard" Java libraries, and nearly all of them are non-free; in many cases, even library's specification is a trade secret, and Sun's latest license for these specifications prohibits release of anything less than a full implementation of the specification. (See http://jcp.org/aboutJava/communityprocess/JSPA2.pdf and http://jcp.org/aboutJava/communityprocess/final/jsr129/j2me_pb-1_0-fr-spec-license.html, for examples.
Fortunately, that specification license does permit releasing an implementation as free software; others who receive the library can be allowed to change it and are not required to adhere to the specification. But the requirement has the effect of prohibiting the use of a collaborative development model to produce the free implementation. Use of that model would entail publishing incomplete versions, which those who have read the spec are not allowed to do.
In the early days of the Free Software Movement, it was impossible to avoid depending on non-free programs. Before we had the GNU C compiler, every C program (free or not) depended on a non-free C compiler. Before we had the GNU C library, every program depended on a non-free C library. Before we had Linux, the first free kernel, every program depended on a non-free kernel. Before we had Bash, every shell script had to be interpreted by a non-free shell. It was inevitable that our first programs would initially be hampered by these dependencies, but we accepted this because our plan included rescuing them subsequently. Our overall goal, a self-hosting GNU operating system, included free replacements for all those dependencies; if we reached the goal, all our programs would be rescued. Thus it happened: with the GNU/Linux system, we can now run these programs on free platforms.
The situation is different today. We now have powerful free operating systems and many free programming tools. Whatever job you want to do, you can do it on a free platform; there is no need to accept a non-free dependency even temporarily. The main reason people fall into the trap today is because they are not thinking about it. The easiest solution to the problem of the Java Trap is to teach people not to fall into it.
To keep your Java code safe from the Java Trap, install a free Java development environment and use it. More generally, whatever language you use, keep your eyes open, and check the free status of programs your code depends on. The easiest way to verify that program is free is by looking for it in the Free Software Directory (http://www.fsf.org/directory). If a program is not in the directory, you can check its license(s) against the list of free software licenses (http://www.gnu.org/licenses/license-list.html).
We are trying to rescue the trapped Java programs, so if you like the Java language, we invite you to help in developing GNU Classpath. Trying your programs with the the GJC Compiler and GNU Classpath, and reporting any problems you encounter in classes already implemented, is also useful. However, finishing GNU Classpath will take time; if more non-free libraries continue to be added, we may never have all the latest ones. So please don't put your free software in shackles. When you write an application program today, write it to run on free facilities from the start.
Copyright 2004 Richard Stallman
Verbatim copying and distribution of this entire article are permitted
worldwide without royalty in any medium provided this notice is preserved.
Note: Comments are owned by the poster. We are not responsible for their content.
once again with this semantic merry-go-round. i'm sick of hearing this version.
yes, you have the "freedom" of choice to choose non-freedom. but what you end up with is less freedom than you'd have had you chosen a free alternative.
duh.
Let's not. The real Java Trap isn't that Java programs are running on non-free runtimes, it's that they're Java programs. Java is a closed system, it interoperates poorly at best with other languages and environments, and it spreads the use of interfaces that are dependent on the increasingly cumbersome Java runtime.
I would applaud the possibility that Microsoft has finally driven a stake in Java's heart if not for the fact that Microsoft's replacement isn't likely to remain even as open as Java.
Ron Cain's Small-C compiler was free (as in beer) and widely used in the 8-bit world before GCC was a twinkle in RMS' eye, there were even free (as in beer) operating systems written for it back when GCC was still "entrapped" by the need to run on proprietary hardware and systems.
Free-as-in-fsf software is still free-as-in-beer.
Yes, this is true. But this does not mean "Free-as-in-beer is the same as Free-as-in-fsf." You are arguing the opposite point as the original poster.
again with the anonymous slagging of slashdot. i think it's great.
Slashdot can be great without this is as bad as slashdot being anything but piquant and precise commentary.
well, luser, you could try asking a whole ass-load of IBM developers who spend time on the kernel. and a bunch of the guys Sun pays to develop OpenOffice/StarOffice. Or the netscape/mozilla guys whom AOL used to employ.
the vast majority of people who program for a living do NOT write shrink-wrapped commodity software like OS's and office suites. they write code for their companies, in virtually every business market that exists.
if you're writing proprietary shrink-wrapped commodity software, yeah, you might think about a career change. that doesn't mean it's destroying the profession of writing software - it just means that certain kinds of software will no longer be developed in the proprietary paradigm.
What exactly are the economics of software? Almost all of the cost is up front, and after the product is good enough to use the distribution costs are negligible and the maintenance costs aren't much higher (ideally approaching zero).
If a piece of software is worth the cost of production to the person who programmed it or to their employer who ultimately paid for the work, then all benefits derived from that work after the sunk costs are recovered is gravy. Giving that software away costs pocket lint, but the benefit that another person/company who uses that given software may be substantial. All benefit others get from using the software is also gravy. Gravy gravy gravy.
If a work of software's value does not exceed its production cost to the producers, then maybe it is worth some fraction of that production cost. That fraction is the reciprocal to the number of times the producers must sell it to break even. The risk of that software's failure is relative to that sales target. What if you can't sell it? What if someone else is duplicating your effort and doing it for less? However, if the value of the software is nearly universal, then it is unlikely anyone can produce software that costs as much as the target market is willing to pay as a whole. This is the rockstar fantasy of non-free software. The bigger the software market, the less significant the costs (ie. distribution and production) are. The risks go up with the gravy. Gravy risk; Gravy gravy gravy risk risk risk.
So, if there is economic gain, does that mean the software is ethical? Why are we throwing ethics into the debate when it is really an economic question of quality and efficiency?
the free software community has offered to give back the most valuable thing of all - more code, via open source development. sun turned them down.
Este texto foi traduzido para o pt_BR em <A HREF="http://www.propus.com.br/news/40" TITLE="propus.com.br">Propus</a propus.com.br>
There is a free java
Posted by: Galik on April 12, 2004 06:37 PM#