Open Standards


Red Hat’s CEO Jim Whitehurst kicked off his third year at Red Hat with a State of the Union address.  In his post, Jim discussed Red Hat’s efforts within the Java community:

“Late last year the Java Community Process (JCP) reached a significant milestone when they approved the specification for the next generation of Enterprise Java; JavaTM Platform, Enterprise Edition 6 (Java EE 6). We believe that the approval of this specification starts a new chapter in the story of Java and we are proud to have contributed and acted in a leadership role in the formation of this standard which aims to make enterprise Java easier to use and more appealing to more developers, while still maintaining the benefits of open standards.”

Craig Muzilla, Vice President of Middleware at Red Hat picked up on the Java thread and wrote a nice post ahead of Oracle’s roadmap presentation on Wednesday.  While many will be following Apple’s every move on Wednesday, those of us in the Java community will be listening to Oracle’s plans for Sun products, including Java.

“As Oracle Chief Executive Larry Ellison said shortly after the acquisition announcement  in April of last year, Java is “the single most important software asset we have ever acquired.”

We agree with Mr. Ellison’s statement; Java is one of the most important technologies developed and adopted during the past twenty years. It has fostered significant innovation throughout the IT industry and has enabled businesses and governments to operate with greater efficiency and effectiveness. Java is larger than any single company; we are all part of Java, customers and vendors alike.

We encourage Oracle to fulfill their original proposal and establish an independent governance process for the JCP (Java Community Process). And, finally, we encourage Oracle to continue the tradition of making the technology easily accessible, to vendors and customers alike, to secure its broad adoption and continued strength in the market.”

Craig points out that Oracle was amongst several vendors, including IBM and Red Hat, calling for Sun to make the Java process more open and less susceptible to any one vendor’s influence or control.  While I’d be surprised if Oracle announced an independent governance process for the JCP on Wednesday, I don’t think Oracle will act to damage the Java ecosystem.  Customers and developers have invested too much in Java products over the past decade on the basis of their investment being protected through a multi-vendor community.  The customer backlash would far outstrip any perceived competitive benefit of tightening control over the JCP.  Oracle’s far too smart of a vendor to risk that or encourage non-standard Java usage as we’ve seen with Android.

Follow me on twitter at: SavioRodrigues

PS: I should state: “The postings on this site are my own and don’t necessarily represent IBM’s positions, strategies or opinions.”

Last week I wrote about the benefits of open standards versus open source. I argued that open standards provide greater protection against vendor lock-in than open source alone. I was reminded of this conclusion when reading Peter-Paul Koch’s analysis of WebKit implementations. Thanks to Palm’s Dion Almaer for pointing out the analysis.

Readers know WebKit as the open source web browser engine used by several mobile and PC web browsers including Apple’s Safari, Google’s Chrome, Palm’s WebOS and the Web Browser for Android. In fact, Wikipedia lists 19 browsers that are based on the open source WebKit browser engine. As you read on, keep in mind that there is no standard that vendors using WebKit must adhere to or claim certification against. A WebKit based browser is, well, whatever the vendor wants it to be.

When Koch tested WebKit browser versions on twenty seven tests, he found:

  • Out of 19 tested WebKits, no two are exactly the same.
  • The best WebKit available is Safari 4; the worst is S60v3.
  • The Android G1 and G2 WebKits score rather badly; it’s the worst mobile WebKit except for S60.
  • Regressions are fairly common: iPhone 3.1, Android G2, and S60v5 all (partially) dropped support for something their predecessors did support.
  • The closest relation of a desktop WebKit to a mobile WebKit is between Safari 3.0 and S60v5. I’m now fairly certain S60v5 is actually based on Safari 3.0. Unfortunately this is the single example of such a close relation.

Koch’s testing highlights two truths. First, pity the mobile web application developer whose manager or customer expects that the application will work on multiple mobile browsers that are “built on WebKit”. Second, open source does not make it easier for customers, or their developers, to transition applications from, for example, building for the Google Android platform to, for example, Plam WebOS.

Imagine if there were a WebKit standard and a compliance test suite that vendors had to certify against to use the “WebKit” brand. Customers and developers would gain protection against vendor lock-in that open standards deliver to a much higher degree than open source alone. I’m not naive enough to think that open standards equals “write once, run anywhere”. But even if a WebKit open standard could drive a 50% improvement in compatibility across WebKit-based browsers, that would be something to write home about.

Follow me on twitter at: SavioRodrigues

PS: I should state: “The postings on this site are my own and don’t necessarily represent IBM’s positions, strategies or opinions.”

Reading Michael Tiemann’s notes for his Open World Forum speech I can’t help but think that he’s attributing a point of value to open source that is much more aligned to open standards based software.

Tiemann writes:

“Lately, I have seen an increasingly important and new implication even beyond what Corey Doctorow has said about free when it comes to software: when the cost of acquisition goes to zero, the cost of retirement (aka “exit cost”) cannot be ignored.”

Tiemann goes on to suggest that the cost of exit for open source software is lower than proprietary software. That can sometimes be true. But open source is no panacea for exit costs. Open standards, while no panacea either, are orders of magnitude more helpful at reducing exit costs than open source by itself.

Let’s say I’m using an open source administration and management product, or an open source content management system and want to migrate to an alternative. How easy will that migration be? Compare with migrating in either direction from GlassFish Application Server to Apache Geronimo. The latter will be easier because both products are standards compliant with the JEE specifications. It goes without saying that open source isn’t a necessary condition to being standards compliant. I’d argue that the cost of exit for standards based software, open source or proprietary, is on average, equal or lower than the cost of exit of open source software around which there are no overarching standards.  I don’t have hard data to prove this point.  However, I do know that CIOs making middleware purchase decisions are keenly interested in reducing vendor lock-in through open standards.

I’m not alone in disagreeing with Tiemann’s views on software exit costs. Open source business intelligence vendor Pentaho’s James Dixon writes in response to Tiemann:

“…true that it might sometimes be easier to get your data out of an open system but most proprietary vendors have migration tools to help move data into their system from a competitors – and of course pre-sales engineers to help out.”

Again, while standards aren’t guarantees of seamless migrations and lower exit costs, standards do however ease the burden. If only open standards were given the same spotlight that open source receives.

Follow me on twitter at: SavioRodrigues

PS: I should state: “The postings on this site are my own and don’t necessarily represent IBM’s positions, strategies or opinions.”

Mike Dolan believes that we need an open API for the various App Stores that are or will be on the market soon.  He writes:

“I can’t help but think with Apple, Sun, Nokia and others with “AppStores” live and working or clones in the works, is it time to have 1 set of open, “AppStore” APIs?”

I couldn’t agree more, especially from the ISV point of view.

Most ISVs will wish to submit their applications to multiple stores.  For instance, an ISV with an application supporting Linux would want the product listed on the Novell, Red Hat and Ubuntu application stores.  Reducing duplicate, or worse, conflicting work, while packaging and submitting the application to these three stores would definitely be a good thing.

The App Store standard would be less about lock-in, since the ISV isn’t locked into any of the App Stores they’ve submitted their application to.  The standard would be about driving consistency, to varying degrees, across application packaging, submission, installation, update, etc. process.

Redmonk’s Stephen O’Grady has written about the need for an App Store for the Enterprise:

“I would love for Ubuntu, our server platform at RedMonk, to connect me back to qualified, rated community resources capable of working on the various packages available in the repository…To the extent that I would pay them for it.”

As enterprise platform vendors, at the operating system, middleware and application levels, begin to expand their App Store capabilities, the wrong path forward will be for each vendor to invest in building the store platform themselves.  The App Store isn’t going to be differentiating technology, so why invest in it individually?  A better approach is for one of these vendors to open source their existing store platform code, with the goal of building open standards based on that implementation.  Then the cost of maintaining and evolving the App Store platform code would be shared across multiple vendors.

What do you think?

Follow me on twitter at: SavioRodrigues

Rod had an amusing and interesting session at EclipseCON titled The Future of Enterprise Java. In the session he spoke about innovation in general and sources of Enterprise Java innovation in particular (more on this later in this post).

He discussed how the JCP is like the Russian Commissar who was tasked with doing what is “in the best interest” of the people, but sometimes gets in the way (of innovation). Rod went on to highlight how political motivations in the JCP hurt customers. He gave several examples including, most recently, Sun attempting to reinvent a sub-optimal alternative to OSGi (with JSR-277) when OSGi exists as a standard already. Apparently this is happening even though IBM, Oracle and BEA would rather utilize OSGi within the JEE spec. Here’s the deck from Rod.

Rod talked about the three sources of innovation in Enterprise Java:

  • The Cathedral (proprietary vendors)
  • The Bazaar (OSS vendors)
  • Commissar (the JCP)

Rod challenged the notion that the Bazaar model is the best or only way forward. Rod claimed (near quotes):

“The bazaar model encourages competition in implementation, but may not produce innovation.

….

The cathedral model is more likely to produce innovation. Remember that Eclipse began as a cathedral project from IBM.

….

Now, the combination of the bazaar model and the cathedral model drive innovation to a much higher degree than either would alone.”

Until now, I’ve always considered the cathedral vs. bazaar models to be mutually exclusive choices. But upon reflection, Rod used the Cathedral & Bazaar metaphors to state what I’ve long believed about the future of the software market. Namely, that the combined use of the OSS model and the proprietary software model is the future. I’ve seen (and lived through) proof that proprietary vendors are learning from OSS vendors. Are OSS vendors doing the same with lessons from proprietary vendors? Or have OSS proponents led OSS vendors to believe that there is nothing to be learned from “the past”. I hope not.

First, I’m an IBMer. These are my views. IBM does not necessarily agree with them.

Second, I think most of you will agree that I’m relatively friendly towards Microsoft when they may a positive OSS move. I am not anti-Microsoft just because it’s cool to love Apple these days.

With those two items out of the way, I couldn’t help but shake my head when reading about Microsoft’s Oslo project for “model-centric apps”.

An InfoWorld article describes Oslo:

“With Oslo, Microsoft is making investments aligned with a vision to simplify the effort needed to build, deploy, and manage composite applications within and across organizations. The effort builds on model-driven and service-enabled principles and extends SOA beyond the firewall.”

Another description comes from Tim Rayburn’s blog (via Google):

“Oslo is not a product, but rather an over-arching project which will contain within it updates to many products and services including BizTalk Server, Visual Studio, and more. “

“..extends SOA beyond the firewall”…oh, you mean to companies that aren’t Microsoft shops? Great, tell me more! I read and re-read the IW article & Tim’s blog to see if there is any mention of non-Microsoft technologies. You know, SOA and especially Composite Applications, are supposed to be about heterogeneous environments. I didn’t find a thing that leads me to believe that Oslo has much to do with interoperability. So, “extends SOA beyond the firewall” should really say “extends SOA beyond the firewall from one 100% Microsoft shop to another 100% Microsoft shop”

Next, for those of us who think standards are generally good ideas…IW sets us straight:

“Microsoft’s approach is not about Unified Modeling Language (UML), a technology of which Microsoft has not been a big supporter…A modeling language is part of Oslo, but Microsoft is building a modeling language, a set of tools, and a unified repository.”

Anywho, it’s all vaporware and likely will be well beyond 2010 (Massimo is being overly generous when he suggests 2009):

“It’s pretty ambitious, and as a matter of fact, I believe we’re not going to see anything concrete until well into 2008 and possibly 2009,” said analyst Massimo Pezzini, vice president of the application platform strategies group at Gartner.

Microsoft, open you eyes…your “SOA” customers care about more than .NET and all customers benefit from open standards. This is not a new news.

I just read the JS post about the continuing rise of OpenOffice.org.

Why do vendors like Sun, IBM (officially now), Google, etc. support OO.o? Well, providing customer choice and openness are two of the key reasons quoted. These are definitely great reasons that we all want to see OO.o succeed.

Most vendors backing OO.o compete with Microsoft in markets other than Office Productivity software. So, what about the competitive benefits for vendors supporting OO.o? Let’s face it, $1 diverted from MS Office could become $0.70 diverted from the budgets of Microsoft’s other business units.

Here’s a look at Microsoft’s 2006 FY financials by reporting division:

– MS Office is inside of “Information Worker”
– MS Windows Client operating systems are inside of “Client”
– MS Windows Server operating systems are inside of “Server & Tools”
– MS SQL Server is inside of “Server & Tools”
– MS Ad revenue is inside of “MSN”

When you look at the operating margins of Microsoft’s business units, it’s quite easy to see why attacking MS Windows profits with Linux and MS Office profits with OO.o is the strategy of choice for vendors that compete with MS in other markets.

PS: I should state: “The postings on this site are my own and don’t necessarily represent IBM’s positions, strategies or opinions.”

Next Page »