PolicyPages; GilbertsProjectGoalProposal
I believe the previous article expresses what this project has become. This is what we already do. I do not propose change. Don't mess with success. I propose that we formally adopt what we have. We have been a successful research project. We have educated the public and our members about the issues. Delegating the details to our sister projects is a good idea and it's working. -- GilbertHerschberger (11 August 2001).
Bear in mind that these proposals stem from John Leuner agreeing with an e-mail that Peter Donald sent a year ago: that there's too much duplication of effort in the Project (its sister projects), and as a result, there should be either a stronger, more centralized Project, one with a single codebase to which the members contribute. That is, the Project should be oriented toward the production of JOS, rather than the discovery of it(s issues). This requires a decision be made for that codebase as regards its license, hence the discussion thereof.
Three courses of action immediately suggest themselves. Continue on as-is; let the Project continue on as is, but for some specific subproject, e.g. the JVM, unify development inasmuch as possible; or attempt to redirect the Project's focus to production. Action can begin on the second two any time someone 'takes point;' the difference between the two for established members is rather minimal. For new members, the primary difference (it seems to me) is one of presentation; with a single 'blessed' codebase (or set of codebases for different pieces) deciding which codebase (sister project) to work on is straightforward. The information for deciding between multiple sister projects is not readily available. In any course of action, additional work on the jos.org website would be useful -- the Info tab/section needs to be done, and the GettingStarted page should probably be updated. What information, and how to present it, remains a question for me. I'll think about it. -- ToddMiller (14 August 2001)
No, Todd. My proposal is a lot older than one year. It is timeless. From the beginning, this project had an identity crisis. It could not decide if it was an implementation project or a research project. Many times its members have complained and critisized this project because its focus is not on source code. It does not build a common codebase.
- My proposal does not stem from "John Leuner agreeing with an e-mail that Peter Donald sent a year ago". I have read all available messages back to the beginning of this project. It has always been an issue. It has always been un-decided. Our members may not know how to think well of a project that produces research and documentation, rather than source code.
- I do not believe that "there's too much duplication of effort" in this project or its sister projects. Rather, healthy competition is a good thing. Competition is a fundamental point of open source software. Competition is most healthy in a public forum, such as provided by this project. May the best implementations win.
- I do not believe that there should be "a single codebase to which the members contribute". One codebase must have one software license. More than one software license is needed; therefore, more than one codebase is needed.
Again and again, its members wonder why this project does not produce more source code. After three years, you might expect an implementation project to have at least some source code, right? But this project is not an implementation project.
- JimMorrison and ToddMiller created the JJOSProject, our first sister project. This project has created working source code. It boots from both Etherboot and GRUB.
- JohnLeuner created teaseme, another sister project. That project is an implementation project; it has created working source code. It re-uses the Linux kernel.
- And I created the CjOSProject, another sister project. That project is an implementation project; it has working source code. It re-uses any foreign operating system.
Our message is inconsistent. Our own members are confused. Members critical of the JOS Project have written articles about what it should be, what it could be, what it might be. Rarely did they write about what it is. Unfortunately, these articles have been allowed to stand un-opposed. I am trying to correct that misconception by explaining what the project is. I am not proposing change. I am establishing a starting point.
The JOS Project is a great project. It is successful. It is valuable. It is a cornerstone for building a Java-based operating system. It is beneficial to the community...
Once again, the same question has been asked: where is the source code? We have to do more to help others understand that, while source code is important, this has never been done before. Discovering and working through the tough issues is more important when it has never been done before.
- Some say that the most successful open source projects have always re-created something that has been done before. Linux re-created UNIX. XFree86 re-created X. Kissme, teaseme and JJOS re-created a Java virtual machine. Kore, Kaffe and classpath re-created the standard Java class libraries.
- Some say that the JOS Project was intended to recreate JavaOS. But has been abandoned because JavaOS was a flop.
Meanwhile, the JOS Project has collected thousands of links to reference material. If you want to build a Java-based operating system, you should start here. We have written almost 2000 articles on issues related to building a Java-based operating system. If you want to learn more or if you have something you'd like to share, you should start here.
-- GilbertHerschberger (16 August 2001)
|
Content of these pages are owned and copyrighted by the poster.
|
|
Hosted by:
|
|