Skip navigation
2006

 

BEA and IBM are doing a good marketing job of spinning their "strategy". BEA calls it a "Blended" strategy....IBM calls it Bluewashing.

 

Marketing spin aside, the strategy is "OSS Strip Mining" which is taking open source software built by a community and "Bluewashing" or "Blending" within proprietary, closed source offerings; forking/changing the open source code as needed in the process. The community does not benefit from this, but IBM and BEA shareholders absolutely benefit.

 

A perfect example is IBM WebSphere Community Edition. It is based on the Apache Geronimo open source project and is "Bluewashed" whereby Geronimo is strip-mined for whatever is useful, then combined with other ingredients, etc. While the resulting product is sometimes free, it is no longer "open source" by any stretch of the imagination. This again is why competing commercial vendors prefer BSD-style licensing. It enables this “stealing” from the community, or in marketing-speak “leveraging”. At JBoss, we prefer FSF licenses, such as LGPL, that prevent the strip-mining while at the same time allowing end users and ISVs to reuse our code. Our users deserve the stability and comfort of knowing that the distribution of JBoss they are using is the same one that everyone else is running.

 

Another commercial software "strategy" is proprietary software "Waste Dumping". Think BEA with Beehive. This is the opposite of strip-mining in that it entails the commercial vendors "donating" some technology to open source. Eager to be part of the open source wave, the vendor identifies some technology that is inferior or of limited value to them, and they dump (oops...sorry...they "donate") it into open source. BEA is probably the worst offender here. There has been so much Waste Dumping going on lately that we may very well need an "Open Source landfill" to deal with the cleanup of all of this waste and its damage to our environment. I’m sorry ladies and gentlemen, but we like our communities clean and not taken for fools.

 

JBoss has always been about pure open source. We started in OSS and we will die in OSS. It rubs me the wrong way when our competitors mischaracterize us, the real developers of FOSS, for not being OSS-enough according to their view of the world. The fact that our LGPL license prevents them from strip-mining us is the real problem. Professional OSS paired with subscription services is what customers want... Not bait and switch, not “Geronimo is for the third world” as I recently heard an IBM exec say at the MIT conference.

 

JBoss is based on the assumption that everyone deserves “world-class” software, fully-loaded with ALL the features. Professional OSS is our livelihood and is a proven model that is driven by customer value. It also drives shareholder value IF you construct your business model properly. Shifting from a software license revenue model to a subscription revenue model is PAINFUL for big vendors like BEA and IBM who need to protect their sacred cows so Wall Street doesn't devalue them.

 

Don’t believe the hype and the FUD. Pure open source has no gimmicks, and the users and customers will settle the score. At JBoss, what you see is what you get and what we support.

 

Peace, Love and Pure Open Source.

 

marcf

 

I already blogged about this topic once and promised to be on the prowl for good articles relating to one of my favorite hypotheses. That AO like mechanisms play a role in nature's evolution. Let me replicate here some of the abstract of "Quasi-species diversity determines pathognenesis through cooperative interactions in a viral population". (Vignuzzi, Stone, Arnold, Cameron, Andino; Nature Vol 439, 19 Jan 2006 page 344-347).

An RNA virus population does not consist of a single genotype; rather, it is an ensemble of related sequences, termed quasi-species. Quasispecies arise from rapid genomic evolution powered by the high mutation rate of RNA viral replication. Although a high mutation rate is dangerous for a virus because it results in nonviable individuals, it has been hypothesized that high mutation rates create a cloud of potentially beneficial mutation at the population level, which afford the viral quasispecies a greater probability to evolve and adapt to new environments and challenges during infection. Mathematical models predict that viral quasispecies are not simply a collection of diverse mutants but a group of interactive variants which together contribute to the characteristics of the populations. According to this view, viral populations, rather than individual variants, are the target of evolutionary selection

 

Basically the article is then an observation of this phenomena and therefore factual proof that quasipopulations are the way of evolution. I will try to translate some of this and draw links to software.

 

What really frustrates me about biology is that the lingo they use is even more obscure than the stuff we use in software. I mean we knock ourselves out with acronyms and create a language barrier to entry. People think we are talking in tongues when really it is not that complex. Not that the concepts are trivial but our communication is more complex than it needs to be. This is particularly true of the bio field. I find myself ignoring all the fancy greek names these guys throw around (too many to remember anyway) and I just try to map the concepts to visual things. BTW if you want something even simpler than a "dummies book" I recommend "A cartoon guide to genetics" which explains IN DRAWING AND FUNNY PLAIN ENGLISH what genetics is all about. It is a GREAT book that is very clear, provides a "rosetta stone" of the lingo they use to visuals and english, and frankly covers a lot of ground, I just enjoyed that book.

 

So the first point I want to focus on is the rate of RNA change. See DNA is a very stable structure. Not only chemically is it stable (dual chain stabilizes) but also there are many mechanisms encoded in it for its stability. Replication of it comes with a lot of error correction. See my previous blog entry on "non-mendelian inheritance in arabidopsis" where basically a "versioning system" (think CVS) has been demonstrated in plants DNA. Metabolic stress due to a wrong mutation in DNA with trigger a reversal of version. DNA is the long term storage of information and there are a lot of error correction algorythms built around it. DNA doesn't change fast, which is good for evolution and it is believed that complex lifeforms evolved only when DNA appeared as a long term storage device. For more on this pick up last week's NATURE where some french guy is at the forefront of a theory that RNA actually preceeded DNA in life.

 

see the point is that RNA actually has a HIGH RATE OF MUTATION. While evolution is something that seems to be encoded in DNA over millions of years. RNA changes on the scale of minutes. There are random mutations (ERRORS IN TRANSCRIPTION) that occur continuously as RNA replicates. The article referenced here talks about 2 point mutations per genome containing 50,000 bases as it replicates in a couple of hours. Multiply that be the billions of RNA strands that are replicating in your own body at this very minute and you get an idea of the bazaar that is going on. Viruses are transport mechanism (jar files in software) that get to cells and use the machinery to replicate. Retro-viruses in fact weave themselves in the host DNA (aspect weaving is the parallel).

 

So onto evolution. If the rate of mutation of RNA is too low, the virus dies. Why? It cannot invade new hosts. Viruses are for the most part benign and believed to have played a big role in evolution so forget for a second the pathologies. Pathologies (the rat dies) are a proof positive of the exponential development (succesful, if out of control replication) of the virus. Viruses treated for low rates of transformation do not create enough diversity (the quasi population they talk about) to actually survive. Don't change and you die.

 

On the other hand the same virus exposed to Ribavirin, a mutagen (generator of mutations) will kill the virus as well. Why? It changes too fast so there is really no continuity in the virus code. Change too fast and there is no lineage. This is called the "error threshold" of viral mutation that will actually kill a viral entity. Ribavirin is a drug used in medicine today.

 

bottom line? There is actually a band of RATE OF MUTATION that is selected by nature, blindly. The article claims there are mathematical models of this dynamic and I intend to check them out since it can't be more complex than exponential growth (rate of death, rate of replication). Replicate with too much fidelity and you die, replicate with not enough fidelity and you die, there is a fine tuned rate at which you are supposed to change and by definition of "selection" that rate is where we are at (otherwise we wouldn't be here, get it? :).

 

I am convinced RNA is natures advices and quasipopulations are natures aspects (collections of advices), retro-viruses are natures weaving mechanism. My question is whether there is a language. There is a debate right now about Intelligent Design and whether all this machinery we are discovering in biology is the product of an intelligent engineering or whether it is the product of chance. Unfortunately ID is pushed by religious factions for evangelical purposes and I don't believe the two should mix. The other unfortunate side of the coin is that most scientists have emotional responses and reject the ID hypothesis only on the basis that in fact it is mostly a religiously motivated hypothesis based on faith alone.

 

Here is what puzzling me. We humans are ALREADY doing ID. We are interacting with Genomes and splicing tomatoes and fruit genomes for our own purposes. In other words ID is a reality for humans. Imagine (for a second) that another race picks up some DNA we engineered in a couple million years, there will be ID (ours) in those DNA structures and they will fiercely debate the issue as well. Unfortunately this is a debate that overlaps with religious faith, which should be separated from scientific credos and inquiry, and in my mind the two approaches should not be mixed but are not incompatible if separated.

 

The counterpoint to ID is that the rate of mutations may be fast enough to explain the high degree of evolution and adaptability in nature. I always though that if DNA doesn't move fast enough then there is only a small chance that the combinatorials would result in such intricate designs. However RNA mutation rate is on another scale (order of 6 magnitude faster?) and its distribution mechanisms, through viruses, and weaving mechanisms, through retroviruses, could explain creation of our reality. Even if not, it is entirely possible that "we just got lucky" and our evolutionary path "found" DNA, just cause "we got lucky". In other words I do not think that even in the case the mathematical models show statistics that couldn't be computed (MIPS style) by nature in time to evolve to our structure, that we need to default to aliens or God rather than "blind luck".

 

anyway, all of this is about proving that AO has played a role in evolution, if after all with "intelligent design in software" (are we that intelligent in software? Hmmmm I wonder ;) we found that the best way to maintain software infrastructure was to adopt AO methodologies then it is entirely believable to me, that AO was "found" or "designed" into natures code mechanisms. For example on the basis of the viral research going on, people are investigating virus as distribution mechanisms for anti-dotes to chemical warfare. Just like we would use AO to patch JDBC drivers under DDOS attacks, we could use viruses to distribute the RNA that encodes a particular suppresor of a viral agent. Trivial. There is enough bits out there with the viruses, the RNA, the retro-weaving, the fast rate of change, to support this hypothesis. I will try and dig a bit in the mathematical models of growth see if there is any statical picture we can derive there.

 

It strikes me, as a final point, that nature's evolution, will recreate a lot of the same mathematics we know from quantum theory where quantum electrodynamics for example takes into account ALL THE POSSIBLE WORLD in its dynamic equations. The math got hairy there. Another field where this appears is in finance, where monte-carlo simulations, which amount to calculating the net present value of a tree of decisions by taking into account all variants and assigning probabilities and profit to them lets you make decision on the viability of projects (a standard procedure in basic corporate finance decision making) proves once again that "real" is in fact a product of "possibles". Nature's mathematics of viral evolution will probably result in the same type of statistical calculations where the equivalent of a project being done is "entity is alive" as opposed to dead.

 

Remember blind luck loves you too :)

 

marcf

Filter Blog

By date: