Open Source Movement dot Org

It looks as if the Apache Software Foundation will be working with Sun on Project Harmony which is an implementation of J2SE 5.

More surprisingly, it looks like Sun Microsystems is happy about it. They have been taking a huge hit from Linux servers replacing Solaris servers. With that in mind, I think Sun is welcoming assistance in making of the test suites, compilers and other parts of the J2SE 5 system while they concentrate on the most important aspects, the core architectural changes.

This is a big boon for the Apache Jakarta set of software, which all rely on Java. When stable version of Harmony comes out, these projects will be able to run inside of an open source runtime environment and they will be able to distribute, for instance, Tomcat, completely with an Apache license.

The new kid on the block of Open Source business models - the stack model - is gaining VC acceptance. The money that’s been on the sidelines is moving into the game. Both SpikeSource and SourceLabs are a less risky, easier to sell, open source solution.

There have been other stacks of software sold such as Red Hat’s server software, and smaller, loosely federated stacks that target developers for the most-part. However, the current stacks lean on proprietary software extensions built into the stacks (as a module on top of mostly open source cores) for revenue. The stack model companies mentioned in this article seem to be going after the enterprise systems by leaning more on packaging, service, support, and certification.

The merit based OS community fosters an environment in which software is only released when it is ready. The problem is that only that one small piece is ready, not the entire framework. Developers who review a lot of Open Source projects envision many the fragmented OS projects converging into a framework, however, they realize that a good bit of the cost is putting all these pieces together for a customer.

This is where the stack business model steps in. Configured, enterprise-ready, supportable stacks of open source with minimal proprietary tie-in, unlike the Linux distribution model that was popular from 2000-2004.

Stacks are also important for developers because it gives them a large base of OS projects on which they can now build on top of as a whole. This allows for further freedoms both in the OS realm and commercial realm of stacks without the hassle of putting all the pieces together.

These stacks are built on top of a loose framework referred to widely as LAMP, which stands for Linux Apache MySQL [PHP, Python]. Ironically these stacks are also available on Windows also. SpikeSource calls that stack the WAMP Stack.

Mr. Lessig has recently posted on his blog about a new report from the OECD, which consists of 30 member countries. The US joined a few months after the organization was established in December 1960.

The panel that produced the new report:

The panel was designed to analyze and discuss changing digital broadband content value chains and business models and help identify new challenges and issues facing the development and delivery of digital content.

The first part of the summary states some very basic assumptions about the fluctuations in the marketplace surrounding media distribution, whether that’s CDs and DVDs or the Sopranos on HBO through access subscriptions:

Disruptive technologies, and broadband in particular, are challenging established business models while creating important development opportunities…

The relationships between content originators and final users are changing, intermediaries are being created or replaced, and attitudes to content ownership and acquisition are changing.

The “Early Impacts” section of the document on page 4 is the most interesting for me because it hilights BitTorrent and P2P:

Users are challenging established ownership and distribution arrangements, whether through P2P networks or open access/open archive publishing conventions, or through new mass distribution and inter-community trading.

Although it doesn’t go into the details, the reason BitTorrent got such a high billing so soon after it’s technology debut is because suddenly new models based on RSS and BitTorrent can effectively produce a “Tivo Season Pass” like feature with little to no cost of distribution.

The point of the entire paper to me is that there is obviously a demand for “easy / deep access” to video content and a need to modify current business models to deal with it.

I’m somewhat taken aback by the news out of Redmond about IronPython. They are now distributing their own flavor Python that is tuned to work with their bytecode engine and .Net framework.

I remember turning in a Computer Science assignment in Python instead of using C++ once back in college. I discovered it after doing research for a previous or concurrent CS course which was a survey on Programming Languages. I didn’t see the logic in the syntax requirements of C++ at the time and Python offered a clean alternative and fit the requirements (mostly) for my professor. I then used it on a web search engine for a trade mag that had music concerts listings which got quite a bit of traffic just after launch. I went on to use Python many other times during my career, all the while admiring it’s readability.

My next interesting run-in with Python was discovering it inside of a very early version of Microsoft Merchant Server the predecessor to Site Server. I found out that the code was from a company that Microsoft had acquired.

This makes the recent IronPython announcement still a precedent setting event. It is the most significant involvement Microsoft has had to-date with the Open Source movement. Thus, with cautious optimism, I congradulate Guido Van Rossum for his contribution to the movement, it is obviously making an impact on the industry. Oh and thanks for the help, especially with http upload with Netscape 3. I’m very cautious about having important projects like Python forked by large companies. Microsoft is testing out it’s first major attempt at working with an existing codebase.

Constant whining about the GPL from large companies eventually gets reverberated throughout the mediocre media without any thought as to the context of the use.

Most organizations simply use software on the application-level. They run copy of Linux as an operating system, or launch Gaim as an IM client. Still other common uses are to use the software to build larger systems such as networks. In general, most companies do not build software. These complaints that I see aired against the GPL are largely only valid when you are a software company that is building products out of software that is already GPL’ed. Outside of that context it is simply a red herring. In so doing, they obfuscate the very real benefits of GPL’ed codebases such as the Linux kernel. They will work fine side-by-side with proprietary software.

Most organizations do not differentiate themselves by software, they manufacture planes, trains, and automobiles. They either deal in hard-goods or soft-goods that are not software. I say to the majority of companies out there that are not large software companies that own their own Operating System - you have plenty to gain from GPL’ed code and it will not invade your business practices or make you pledge your firstborn.

For a lot of projects the GPL is crucial because the developer (software artist) that released the code did so for the sake of his community, not for it to disappear into the codebase of Microsoft or Sun as soon as it is made available. This will continue to be a truth, Sun should learn to deal with it appropriately and with one clear voice. This is the modern-day equivalent to a painter only wanting their work on public display. They have no interest in a museum procuring it or entering the work into the marketplace. It is their right to do so, and a company that knows software and Open Source should know that.