Three decades ago, the British evolutionary biologist Richard Dawkins proposed the concept of a “meme” – that is, a unit of cultural transmission that exists independently of one person, and therefore, can spread around throughout a culture.

Some memes are obvious. President Bush referred to himself last April as “the decider,” that is, what we used to call a decision-maker. At first, other people used that term in mocking references to Mr. Bush, but now the word has caught on.

I’m watching a meme evolve right now: the Novell Sweatshirt meme.

In mid-1996, Novell introduced a new bundle of its NetWare network operating system with a suite of server applications, like a Web server. It called this version, “IntranetWare.” During the launch press tour, Novell gave me a sweatshirt with the IntranetWare logo.

The dark green sweatshirt was exactly the sort I like: medium-weight cotton/polyester fleece material, mildly elastic cuffs, and a cardigan-style zipper. There’s no hood — I can’t stand hooded sweatshirts. The IntranetWare logo is small and discrete, embroidered in black thread, so it’s not obvious that I’m wearing vendorware.

I wear my comfy Novell sweatshirt often, indoors and out; it’s my favorite item of clothing, frankly. Many of my friends and colleagues have seen it over the past decade.

A few years ago, my son bought a light gray, zippered, hoodless fleecy sweatshirt. And he calls it his Novell sweatshirt, because it’s just like mine.

Yesterday, my extended family was down shopping at REI, a sporting-goods and clothing store in San Carlos, Calif. They were having a big after-Christmas sale. A relative said, “Oh, look, they have a nice selection of Novell sweatshirts like yours,” and bought a charcoal-colored zippered, hoodless fleece sweatshirt from Columbia Sportsware (pictured). He’d heard the term from my son.

I bought a new dark blue Novell sweatshirt too. You can’t fight a meme, especially when it’s on sale for only US$29.83.

Microsoft has just announced the 2007 dates for its biannual Professional Developer Conference. It’ll be held Oct. 2-5 at the Los Angeles Convention Center.

Microsoft PDC is my favorite of their three big technical events; it’s where you get the real deep-dives into their upcoming technology, and their road maps. I’ll be there… I try never to miss a PDC.

I’m not sure if I’ll make it out to Microsoft Tech-Ed this year. Usually I go, sometimes I don’t, especially in a year when there’s a PDC. The conference, set for Jun. 4-8 in Orlando, is more practical than PDC, but is also much broader. Tech-Ed’s audience includes developers, DBAs and sys admins. So, you’ll find lots of classes about managing Exchange Server or managing servers with Microsoft Operations Manager, which doesn’t interest me.

The third big event is Microsoft WinHEC, which is a hardware design conference; that’s where Microsoft lays out its reference platforms. WinHEC will be held this year on May 14-17 in Los Angeles. WinHEC is an important conference, but I don’t always make it there.

Note that in 2007, Microsoft had planned to have both TechEd and WinHEC in New Orleans, but moved the events because of uncertainly about whether New Orleans’ still-recovering infrastructure would be able to support such large conference.

For a quick reference to significant U.S. conferences for software development managers, bookmark the BZ Media Industry Events Calendar.

If you’re going to EclipseCon, be sure to catch my presentation, “Working with the IT Press.” This “long talk” session has just been accepted by the conference program committee.

“Working with the IT Press” is an evolving presentation that I’ve been delivering for more than a decade at various conferences and trade shows, and in shorter form one-on-one with small company owners.

My goal is to help small organizations (such as software companies or open-source projects) to know the ground rules of working the media: how to pitch us on news stories, how to handle an interview, how to interact with writers who are doing product reviews, and so-on. Some of those rules are counter-intuitive.

Big companies, and their professional public-relations agencies, always have an advantage, in that they know how to work the press. This session tries to level the playing field. Hope to see you there. EclipseCon 2007 is March 5-8, in Santa Clara, Calif.

Thank you, Sun (and Sun’s PR team), for sending me one of Sun’s SPOT (Small Programmable Object Technology) kits. It arrived earlier this week.

The SPOT kit contains a couple of battery-powered devices, each with sensors and a wireless transceiver, plus a base station and a Java SDK. You can see all the parts at Sun’s Web site.

What am I going to do with it? My son and I have a few ideas, which we’ll be playing with. We probably start with something really simple, like using it to make an indoor/outdoor temperature display (put one SPOT inside, one outside) as our “Hello, World” application. After that, who knows?

Andrew Binstock has written an excellent blog entry about the pros/cons of porting an existing C/C++ project from Visual Studio to Eclipse. His post was inspired by a tutorial on developerWorks by Priyadarshini Sampath (pictured), Ramakrishnan Kannan and Karthik Subbian, all of IBM.

You should read Andrew’s post, but his conclusion agrees with my own belief that a developer who is solely targeting Windows (native Win32 code, or managed .NET code, or ASP.NET code) and who is using Microsoft’s compilers is probably better off using Visual Studio, because of the tight coupling between the IDE and Windows, and also because Visual Studio is smarter about working with the Microsoft’s compilers.

For Windows developers, the strength of Eclipse is that it’s inherently multiplatform. If multiplatform development is a consideration, or if you’re planning to use any non-Microsoft elements in your toolchain, such as gcc and gdb, Eclipse is the tool for you. The developerWorks article does admit, however,

“The growing popularity, versatility, and open source nature of Eclipse motivates many to embrace Eclipse as the development platform of the future. Nevertheless, porting Windows applications to use open source development tools like GCC, GDB, or GCC/GDB for Windows providing functionalities similar to Windows SDK is a nontrivial task today.”

Therefore: If multiplatform development is not a consideration, and you’re using the Microsoft compilers, then you’re probably better off with Visual Studio.

See CNN and AutoWeek.

Send your letters to:

Ford Motor Company
Customer Relationship Center
P.O. Box 6248
Dearborn, MI 48126

A study by International Data Corp. — a respected IT market research firm — predicts that within its first year of shipment, Windows Vista will be installed on more than 90 million computers worldwide, and that in the United States alone, Vista-related employment will reach 18% of all IT jobs during that period.

The study, “The Economic Impact of Microsoft Windows Vista in the United States,” was commissioned by and paid for by Microsoft.

The study further estimates that there will be 157,000 new jobs (emphasis theirs) and US$70 billion in revenue to companies in the U.S. IT industry, with two million IT professionals and industry employees working with Vista in 2007.

The latest version of the Java Platform, Standard Edition, was formally released yesterday (Dec. 11). It would be fair to call this iteration evolutionary, not revolutionary. I can’t think of any “must-have” additions to the platform, but overall, there are a number of incremental upgrades wrapped into the package.

As BEA’s Bill Roth wrote, “First, XML processing, via JAXP and JAXB are now standard.” Also, Java 6 standardized the annotation processing API, and improves support for dynamic languages. That’s potential huge, given the interest in running languages like the JCP’s own Groovy, as well as JavaScript and Ruby, within the JVM.

There are a number of other enhancements as well to JSR-270, which is the umbrella Java service request for Java SE 6, which had been code-named “Mustang.” A big one, but which hasn’t been discussed much, is JDBC 4.0, which makes it easier for developers to associate SQL queries with class representations.

Also very significant (although not part of the JSR) is that Sun is including Apache Derby, a Java database, as part of its own JDK for Java SE 6. This move will cement Derby’s status as the de facto standard Java database.

Alex Handy talks about some other changes, such as its integration with Solaris’ DTrace functionality, in an SD Times news story.

Again: Java 6 isn’t revolutionary, but in regards to the database and scripting language support, it’s significant.

The IBM Rational tools keep going and going… they’ve undergone changes in just about every dimension, over the past five years. They’ve been purchased by IBM. They’ve been rebuilt on Eclipse, instead of the old VisualAge code base. The latest version, called IBM Rational Software Delivery Platform 7.0, debuted yesterday (Dec. 5).

As you’d expect, the big focus of the update (code-named Caspian) is driving SOA-based development. SOA is the acronym-of-choice within the big account that IBM likes to serve with its software tools, its WebSphere app and portal servers, its DB2 database, its Tivoli management suites, and of course, with IBM Global Services.

Beyond the emphasis on SOA, the biggest improvements are in Eclipse support. SDP 7.0 now uses the latest 3.2 version of the Eclipse platform, including the multilanguage IDE. While the focus of the SDP suite is on Java EE-based development on WebSphere, some of the tools (specifically, the testing suites) also work on .NET code as well.

But make no mistake: IBM is continuing to go after big shops that do formal UML-based modeling and who like rigid processes like the Rational Unified Process. The SDP isn’t targeting small or mid-sized development shops using Visual Studio or Borland, or those who like to “roll their own” Eclipse toolchains.

If DoDAF and RUP aren’t in your daily vocabulary, and you have small teams working on agile projects, IBM’s SDP probably isn’t for you. But if you do care about DoDAF and RUP, you have hundreds or thousands of developers on some projects, business processing modeling is at the top of your agenda, and your pockets are deep, then you should call IBM right away to talk about upgrade pricing and migration services.

Santa Barbara, Calif. – Following CEO Dan O’Dowd’s talk about the need for a secure network platform, Green Hills Software announced that it is launching a secure network platform.

The Green Hills Platform for Secure Networking (to use its formal name) includes the company’s INTEGRITY real-time operating system, an IPv4/IPv6 networking stack, a file system, IPSec, IKE, a secure Web server that uses SSL/TLS, a SSH client, a RADIUS client, and SNMP. (The software is being officially announced on Wed., Dec. 6th.)

The target market for the GHPfSN (to make up an acronym) is networked devices, such as LAN/WAN infrastructure, media servers/gateways, bank systems, handheld devices, and medical equipment. The company cites recently published buffer-overflow flaws in Cisco’s IOS (the operating system inside its switches and routers) as explaining why GHPfSN is needed.

What makes GHPfSN secure, according to Green Hills, is that INTEGRITY isolates and separates the kernel from applications using protected address spaces for applications, system services, system data, and hardware device drivers. So, for example, the IPv4/IPv6 stack is in its own secure address space, so that even if it’s compromised, it doesn’t provide an entry point into the kernel, and from there into the rest of the device.

Other Green Hills news…

The company has updated its INTEGRITY real-time operating system. Version 10 adds support for symmetric multiprocessing, as well as a real-time debugging agent tied into Green Hills’ EventAnalyzer software. There are also new security features, including a new way of allocating memory resources, and a new device-driver model. INTEGRITY 10 will be available in the first quarter of 2007, according to Green Hills.

As a technology analyst, I’m expected to be fair, unbiased, impartial, open-minded, objective and neutral at all times.

And I am, of course.

Even so, I must confess some amount of pleasure in the blow dealt to SCO in its perennial lawsuits against IBM and Novell this week, when (as reported on Groklaw) U.S. District Court judge Dale Kimball affirmed previous court rulings that dismissed many of SCO’s claims against IBM. To quote from the documents (many thanks to Groklaw’s Pamela Jones for making this available),

Having thoroughly reviewed and considered the briefing related to IBM’s Motion to Limit SCO’s Claims, the briefing related to SCO’s objections, the underlying previous discovery orders, and the arguments made at the October 24 hearing, the court finds that, even under a de novo standard of review, the Magistrate Judge’s June 28, 2006 Order is correct. The court finds that SCO failed to comply with the court’s previous discovery-related Orders and Rule 26(e), that SCO acted willfully, that SCO’s conduct has resulted in prejudice to IBM, and that this result–the inability of SCO to use the evidence at issue to prove its claims– should come as no surprise to SCO. In addition, the court finds that neither particularized findings on an item-by-item basis nor an evidentiary hearing is required to make these determinations. The court, therefore, affirms and adopts the Magistrate Judge’s June 28, 2006 Order in its entirety.

In other words, SCO hasn’t provided any substantive evidence (as required by the court) of IBM’s improper use of its intellectual property, and therefore, many of SCO’s claims are summarily dismissed, buh-bye.

Given that SCO has behaved appallingly throughout all this protracted mess, it’s hard for even the most impartial observer to watch the company’s setbacks without applauding. (Of course, partial observers, like those who work for Microsoft, are undoubtedly dismayed by this court order.)

This week, SCO’s stock (NASDAQ:SCOX) dropped to from US$2.43 per share to $1.22 per share — a fall of just about 50 percent. Too bad, so sad. I like Forbes’ headline: “Investors Abandon SCO.”

The XQuery 1.0 specification — in progress since the Mesolithic era — reached a new milestone last week: it’s now a Proposed Recommendation from the W3C. (Well, no, it’s not really dating back 8000 years, but XQuery has been in progress since early 2003.)

XQuery is important because it’s essentially the XML-based analog to SQL. In other words, XQuery is a query language that applications can use to talk to servers, to extract data from documents, and do data-centric document transformation. It’s essential for data-driven service oriented architecture.

In the database realm, XQuery will be a replacement for XSLT 1.0, the Extensible Stylesheet Language Transformation facility. XSLT was really designed to assist with the presentation of XML documents, such as for output to a Web browser or exporting as PDF. Even so, XSLT has been shoehorned into doing just about everything to do with XML data transfer, in many cases inefficiently or inelegantly.

Industry support for XQuery is high, and the preliminary versions have been widely implemented in a number of commercial tools. Expect the W3C to swift approve the proposed draft, and make XQuery 1.0 into a full Recommendation (i.e., a standard). A lot of people are waiting anxiously for it.

Even so, the first rev of the specification isn’t a full analog to SQL; rather, it’s a data-retrieval subset of what we should expect XQuery to mature into. Thus, you can’t use XQuery 1.0 to submit data into a database: It’s pull, rather than push. For those capabilities, wait for XQuery 2.0.

Also on Nov. 22, the W3C upgraded XSLT 2.0 and XPath 2.0 to Proposed Recommendations. These are also important upgrades, and these three specs together vastly extend our ability to work with XML programmatically, and will drive data-centric SOA implementations.

A fond memory from my early IT career, in the late 1970s, was working with APL — A Programming Language — on the IBM mainframe.

Our data center had one LA36 DECwriter II terminal that had the special APL character set, and a small number of enthusiasts used to fight over who got to use it. The losers had to use standard DECwriter II terminals or glass TTYs (from Televideo or Perkin-Elmer) and type in the verbose equivalents of the symbols. I have no idea why we had APL on our System/370; it was never used for anything but fooling around.

APL was launched on Dec. 1, 1966, and had been invented by Kenneth Iverson, an IBM researcher. Amazingly, IBM still sells APL tools today, 40 years later, though they’re richly priced at US$1,741 for one workstation license with 12 months maintenance. Dr. Iverson passed away in Oct. 2004.

APL is an interpreted language designed for working with numerical arrays. You can perform complex calculations with only a few characters; designing algorithms to be concise (and often unreadable) is a marvelous talent. Sadly, I didn’t spend enough time with APL to develop much skill. My “real” work involved PL/1, COBOL and FORTRAN, and noodling with APL was essentially a hobby.

For a quick overview of APL, I’ll refer you to the Wikipedia (which is where the keyboard image above comes from). However, that article merely scratches the surface, and doesn’t show the beauty and elegance of APL programming.

In the early 1980s, I evaluated an APL interpreter for DOS, called APL*Plus, from STSC. It was a bizarre implementation, and unstable. Alas, I haven’t worked with the language since.

What the heck are Microsoft and Novell up to? At the beginning of this month, the companies announced a multi-part agreement, wherein they’d work together to improve interoperability between SUSE Linux and Windows. That’s all well and good. But there’s more to the agreement, much more, that is not well and good.

* Microsoft agreed to redistribute SUSE Linux Enterprise to its customer base – to run inside virtual machine under Windows Server 2003. That’s bizarre, but I can see where it might appear to be a win-win. Novell gets some Microsoft endorsement, and Microsoft can sell server licenses. However, statements like “Novell will also make running royalty payments based on a percentage of its revenues from open source products” are worrisome and suspicious.

* There’s also a patent cross-licensing agreement in the deal, which Novell says was added at Microsoft’s request. The gist of that agreement was presented as that Novell and Microsoft wouldn’t not to sue each others’ customers over patent infringement. That’s even more bizarre: Why would Novell think that Microsoft would be able to sue Novell’s customers? Why would Microsoft think that Novell would be able to sue Microsoft’s customers?

Predictably, the blogosphere went nuts over the agreement (sparked by inflammatory comments by Red Hat, the other big Linux distributor, and its JBoss subsidiary). Is Novell selling out? (Answer: Yes.) Is Microsoft launching a new attempt to crush Linux? (Answer: Yes.) However, it all would have died down in a few days… had not Microsoft itself showed its cards.

In the press conference, blogged by the Seattle Post-Intelligencer, Microsoft CEO Steve Ballmer (pictured, sitting with Novell CEO Ron Hovsepian) asserted that Linux contains Microsoft intellectual property: “But to the degree that people are going to deploy Linux, we want Suse Linux to have the highest percent share of that, because only a customer who has Suse Linux actually has paid properly for the use of intellectual property from Microsoft.”

Novell immediately cried foul, denying that SUSE Linux contains Microsoft intellectual property, and insisting that in making a cross-licensing agreement, it wasn’t implicitly acknowledging any infringement. Microsoft is sticking to its guns, however:

“Microsoft and Novell have agreed to disagree on whether certain open source offerings infringe Microsoft patents and whether certain Microsoft offerings infringe Novell patents…. We at Microsoft respect Novell’s point of view on the patent issue, even while we respectfully take a different view. Novell is absolutely right in stating that it did not admit or acknowledge any patent problems as part of entering into the patent collaboration agreement. At Microsoft we undertook our own analysis of our patent portfolio and concluded that it was necessary and important to create a patent covenant for customers of these products.”

Two things are clear. First, Novell appears to have signed an agreement that it didn’t read carefully. Second, we’re going to hear a lot more about this… from Microsoft. Probably in court documents.