As I posted on February 18, I was a little surprised that the Byzantine nature of the open source software (OSS) culture had made it into a major metropolitan daily newspaper, the Boston Globe, on the op-ed page. It is less surprising to find it on the Business page but I think it is more than coincidence to find OSS in the same daily paper twice within 10 days (subscription might be required except today February 27).I think we might have a semi-sophisticated PR campaign afoot (what some might call “spin”) and the Boston Globe is the kickoff point because it’s the hometown paper of the Free Software Foundation (FSF). Of course, I still think there was a shooter on the grassy knoll.

If there is a PR campaign rolling out timed to the upcoming belated release of the FSF’s new GNU General Public License, Version 3, I hope the facts are reflected better than they are in this morning’s Globe. (Note that the Globe is part of the New York Times’ empire so this story may be picked up around the world, not to mention the ease of access via Google.) In the Globe’s defense, OSS, licensing and patent catfights is no doubt a tough issue to write about in the space restraints of a daily metro paper; they must also have to make some interesting assumptions about the technical level of their readers.

Whatever (PR campaign, coincidence) the story is full of one-sided “Microhate” right off of the FSF’s and associated web sites. Some of the problems with its slant include:

-- Linux is not made up of hundreds of pieces of software produced by the GNU project (the FSF-linked development “community”) as the story states. The Globe also says that Red Hat and Novell have a special service/product relationship with FSF and GNU. (IBM, HP, Oracle, etc., etc. always get left out of these stories because it ruins the big guy vs. little guy drama.)
1. As described here on January 24, Linux is an operating system kernel that has had many homes over the years and which is now sort of “resident” at the Linux Foundation . If you’ve been writing about high-tech since around 1990 like I have, you will find the Linux Foundation interestingly reminiscent of the Open Group, originally X/Open and the Open Software Foundation (OSF), to the extent that Unix was “resident” at OSF.
2. It’s hard to pinpoint where an OSS project “resides” but Linux does not relate to the FSF any more than it does to Fedora, Ubuntu, etc, all similar OSS communities such as GNU. In fact, even GNU tries to explain the relationship on its web site.
3. The real relationship between GNU and Linux is that the prime-mover developer of Linux chose the GNU license when he released it formally in the 1990s. If he had chosen any of a half-dozen other possibilities available at the time, the FSF would be a footnote in a computer history book.

-- What FSF is spinning (or the Globe is not getting) is that this “Microsoft-Novell vs. the good guys thing” is all about legalities, including intra-OSS infighting.
1. The last sentence of the story brings out the FSF’s real position. FSF is “anti-open-choice” in technology selection apparently because of a communistic aversion to intellectual property rights although I think that WTO-like position evolved in the FSF over time because it sounded altruistic.
2. A story that I like (just because it is a good story; I have no way of knowing its accuracy) is that the whole “free software” thing really started because someone kicked Richard Stallman out of an MIT lab where he had literally taken up residence in the early 1980s.

-- The rewriting of the 1991-vintage GPL has been ongoing for some time (and needed for a lot longer than that). It did not arise out of the Microsoft-Novell deal.
1. The GPL3 process is behind the originally announced schedule because many of the OSS movement leaders question the specific wording and others even question the need and/or intent.
2. The new wording only matters if someone other than GNU uses the new license on a new piece of software (see my February 5 post). Many OSS projects, including Mozilla, Berkeley, and Apache, use other licenses and the new GPL wording has no effect on any software licensed under the current GPL2 version.
3. MySQL, the leading open source database provider, just changed its licensing terms so that its “products” will not automatically be covered when the new license is released (see 12/22/06 entry in Arno’s blog).

-- “Renowned” open source programmer Jeremy Allison might have quit Novell because of the deal, as the Globe says he says, but he had been there two years already
1. The fact that he worked at at least 10 companies in the last 15 years (SGI, Vantive, Whistle Communications, IBM (when it acquired Whistle), Cygnus, Sun, VA Linux (‘downsized I think), HP, Novell, and now Google) and seems to switch regularly every two years might have had something do with it.
2. This of course only changes who signs his pay check. Like most “volunteer” OSS programmers, he is paid by a large IT supplier.
3. I don’t care that that’s the case but it must be taken into account when analyzing the OSS movement.

I am not qualified to comment on the legalities of the Novell/Microsoft patent agreements but I’m blogging so I will anyways:
1. They sound to me like the standard cross-licensing agreements in which many IT suppliers participate.
2. Only FSF really says that this November 2006 deal increases the threat level of patent infringement suits. The threat is longstanding; Ballmer has been talking about this for years

When I took up this blog I said I would treat contentious OSS isses case by case. I am as critical of Microsoft as anyone. But this will be the biggest "case" in OSS during March as the FSF GPL is formally released. What do you think?

Posted by dennisb in OSS Culture | Permalink | Comments (0) | TrackBacks (0)

One of the major benchmarks of open source software (OSS) acceptance will be the applications catalog. That works two ways:
-- open source applications
-- any applications that run on OSS platforms.
I don't see it on either McKesson's or Red Hat's home pages yet but Business Wire is reporting the following:

ATLANTA--(BUSINESS WIRE)--McKesson has joined with Red Hat (NYSE:RHT), the world's leading provider of open source solutions, to introduce the Red Hat Enterprise Healthcare Platform, a cost-effective open source information technology (IT) solution with services designed to meet the mission-critical demands of healthcare.

“The Red Hat solution offers our customers a reliable, affordable platform for delivering safe, high-quality patient care using McKesson’s clinical applications,” said Michael J. Simpson, chief technology officer for McKesson Provider Technologies. “The introduction of a high-value, open platform designed specifically for the needs of healthcare IT represents a major step forward in encouraging the use of open source technologies instead of closed, proprietary technologies that are costly to acquire, maintain and scale.”

The Red Hat Enterprise Healthcare Platform packages the Red Hat suite of open source products and services, including Red Hat Enterprise Linux updates via the Red Hat Network. It also includes dedicated customer support and services, as well as other open source technologies such as JBoss Enterprise Middleware. This complete solution provides superior levels of stability and performance as well as predictable maintenance cycles.

This is the classic HBO & Company application (company acquired by McKesson about five years ago) that was an important piece of Digital Equipment's move into the IT mainstream during the 1980s. More to follow

Posted by dennisb in OSS Business Issue | Permalink | Comments (0) | TrackBacks (0)

I have written frequently on my own web site (see link in panel to right) about the way IT leaders have co-opted the open-source movement in order to provide open choice. In fact, a subset of my research appeared on ebizQ in January and that's how I ended up writing this blog. Now I notice that LWN.net has really peeled back the details on this trend, particularly as it relates to the Linux kernel.

As you read the article, you can see the various comprehensive methodologies and perspectives taken by "corbet" in the analysis. The bottom line, after Red Hat and "truly volunteer," the big contributors to Linux 2.6.20 are--no surprise--IBM, Google, Oracle, HP, and the Linux Foundation (sponsored by the same big companies; see my January 24 post under OSS Culture, also link from panel to right) along with Novell, the University of Aberdeen, Nokia, Sony, Intel, and SGI. The actual rank differs depending on what is being measured but the names are pretty common across the measurements.

The article calculates that, despite conventional wisdom about guys in their attics pounding out code, no more than 33% of the latest Linux kernel was produced in that manner (and this revision is 700,000 lines larger than the last). In fact, that calculation assumes--for argument's sake--that all the unknowns are volunteers but it is just as likely that all the unknowns split in the same proportion as the knowns, which would mean only about 10%-15% of Linux is now being written by this generation's Linuses.

Posted by dennisb in OSS Development | Permalink | Comments (0) | TrackBacks (0)

As promised, I followed up last Thursday's blog (Feb 15) about the new Open Solutions Alliance (OSA) by "meeting" with Barry Klawans, OSA spokesperson and CTO at JasperSoft. As noted Jaspersoft is one of the founding members. I am happy to say there is meat to go with motherhood and apple pie I mentioned last week.

As I mentioned they are paying attention to the real applications (as opposed to tools and infrastructure) that the open source software (OSS) movement needs to take off. Their mantra is more about interoperability than my post last week would have indicated and they are espousing vendor neutral in terms of "come on, come all." I had surmised that "vendor neutral" might mean they would be developing a framework like PeopleTools among themselves. Barry admits that would be nice to have but wants someone to come along with a product to donate to the community or the desire to take a framework on as a project. Most likely they will work with pieces of various frameworks rather than adopting only one such as Red Hat JBoss' JEMS or the enhanced LAMP stack likely to emerge from various Apache efforts.

From an integration software user point of view (based on the subject of ebizQ; of course if you are an integration software user, you use other stuff that needs to be integrated), this group has identified exactly what the open source application market needs to break out. The idea that there might someday be an OSS equivalent of SAP does not make sense because it goes against the whole OSS-based open choice philosophy: one from column A, one from column B, etc. But the OSS movement needs an OSS Netweaver (or Fusion or whatever Microsoft is calling Project Green these days). If you know anyone that is working on such a framework, let me know and I'll chase them down. I am already following JBPM and a few others.

(More on Jaspersoft in a later post. It seems to be a good example of how people, proprietary products and OSS projects come together to move the ball down field. Barry mentioned how they made use of OSS software to develop their OSS software in the same way that Progress had mentioned Tuesday (see Feb 20 post) that it used OSS software to develop its traditional products (not sure if I am going to adopt the term closed source)).

Posted by dennisb in OSS Development | Permalink | Comments (0) | TrackBacks (0)

Apache says its ActiveMQ effort has graduated from the Incubator to become a top level project (TLP). The Apache ActiveMQ, termed a so-called enterprise service bus (ESB) in some press articles, is a distributed messaging system that will support clients in Java (JMS), C++, Ruby, Python and other languages. Apache ActiveMQ is said to fully support JMS 1.1 and J2EE 1.4.

(I say so-called ESB for two reasons:
-- There is some question among those purists such as me whether Apache ActiveMQ is really an ESB
-- An ESB is just a type of MOM (message-oriented middleware) to me anyways)

Still, moving up the stack is an important step for Apache, which cut its teeth and claimed its fame on the web server of the same name. Now, it is working at additional levels of modern middleware (most notably with the IBM Gluecode application server development team as part of the Geronimo project). ESB and/or MQ-level activity can be thought of as even "higher level" than a web server or an application server. I don't want to characterize it in a way that implies that technical success might be harder to accomplish but I would say corporate acceptance might be harder. The heritage MQ brand's longheld meaning is one of "bullet proofness" in enterprises. And with any level of ESB-like functionality (whatever us purists say), we are beginning to see OSS supporting applications that enterprises really consider mission critical. The involvement of a key Amazon developer might help allay any CIO fears.

In terms of commercial backing (see my February 16 post on the taxonomy of OSS and the importance of commerical backing), Logic Blaze and its Fuse project are the only one mentioned on the Apache.org site at this time. Optaros also say it supports ActiveMQ via its SOA practice (and Optaros is a purist like me, making a distinction between an MQ-like message broker and a true ESB). If you know of any sources of ActiveMQ support, let me know via the comments link below.

In terms of OSS development efforts, the ActiveMQ project competes with the Celtix ESB effort out of CORBA-expert Iona (which, to further illustrate the intertwining of the OSS culture, sprung out of Apache "competitor" ObjectWeb). There is also a Mule effort.

In terms of competing closed source products, almost all the leading software suppliers have this covered both ways. Progress' SONIC ESB is the most well known of the commerical suppliers because its Dave Chappell literally wrote the book (and I would say Progress' SonicMQ is acually more the direct competitor). But IBM, Oracle, SAP, Microsoft, BEA, TIBCO, Software AG and many more (including small pure-play providers such as Fiorano) provide ESBs and older generation MQ-like products as separately priced products or embedded elsewhere int their middleware offering--or most likely both. Progress' execs made the point to me in researching this subject that the big win for the commerical suppliers is the way in which they use OSS as components (Axis was cited as an example) and give back to the community with quality control on those components. Progress also participates by providing dedicated development resources; they have a guy on the Apache Synapse project for example.

Apache ActiveMQ is released under the Apache License, Version 2.0.

I opine that although the free download is as simple as any other OSS action, users are really getting into areas where documentation, training, and most important, previous experience are critical. Neither an ESB nor an old fashioned MOM is something you play around with in the sandbox for an afternoon even if it's free.

Posted by dennisb in OSS Development | Permalink | Comments (5) | TrackBacks (0)

I can't get away from the subject of open source software (OSS) even on my lazy day, Sunday. Today's Boston Globe editorial pages have an opinion column titled Free Beer purportedly explaining open source software in the context of a bunch of Scandanavian kids that have started an OSS-like movement for their home brew. I think you have to subscribe to the Globe to read the article unless you read it today Feb 18 (and even then you have to register) but the author, Attorney Henry Lanman, writes for Slate and other publications so the article or one like it will probably show up elsewhere. In the spirit of the Free Software Foundation (FSF), I could just copy and paste it here--given my acknowledgement of the Globe--but I believe that to be a violation of the Globe's intellectual property (IP) rights as publisher.

The article covers issues I have written on at my IT Investment Research website but which I have avoided so far on ebizQ: the not so trivial but eventually inconsequential difference between the FSF's ideals and the broader OSS movement's philosophy. Not unexpectedly, an article written for a daily metropolitan newspaper still misses some of the nuances:
-- Even the FSF makes the point that its message is not about "free software" as in the example that Quicken is often given out free by banks to entice you to use their banking services (or in the way the dastardly Microsoft provides Office Live free to try to get small businesspeople to use its business applications); that is the point behind the FSF mantra "think of free speech not free beer."
-- The "free software" idea in the FSF sense (meaning--among other things--that others can take the source code and modify it and redistribute it freely as long as they don't add restrictions) predates Richard Stallman and the FSF; the idea is more from the early 60s than early 80s and could even be said to date to the 1955 founding of the IBM COMMON users group
-- It's not the case that developers first started asserting IP rights in the early 80s; it's just that IP rights did not mean much until then because before that time, the software was pretty much locked into whatever hardware it was written on; the former Data General started asserting its IP rights in its software almost from its founding in 1968 and lost a great deal of money in the 1980s when those rights were not upheld in Federal Appeals Court and failed to get a hearing at the Supreme Court level (given changes in the law and subsequent court proceedings, DG probably would win the lawsuit today)
-- Though the OSS culture and commerical backers propagandize that the OSS development model results in higher quality programs, there is no empirical evidence of that (and it defies logic); that was not the reason that the broader OSS movement broke with the FSF (or vice versa)--the split is all about IP rights
-- There is nothing in the evolution of acceptance of the Linux operating system (characterized as the "success of the Linux operating system" in the article) that could in any way be tied to the idea that OSS=higher quality software (implicitly, higher quality than Microsoft since it is unlikely that anyone that makes this claim means higher quality than IBM's mainframe software); Linux-based operating software has been gradually replacing operating software based on earlier versions of UNIX over the last 15 years because Linux is a descendant of UNIX and that's the way the UNIX market has always evolved (acceptance of Windows continues to outpace acceptance of Linux--see my post of February 9 and other research particularly from IDC)
-- The word "source" in the term "open source software" also needs to be explained better and understood in terms of the OSS movement but as I said, Sunday is my lazy day

I told ebizQ when I started this blogging activity that I was agnostic about the Microsoft vs. OSS war. I am not going to enage in the Microhate that appears on most OSS blogs, and conversely, in my wider research, I am often critical of Microsoft where it is deserving. I say some things above that the OSS community will find pro-Microsoft but I contend that they are "just the facts, ma'am." If you want to debate about them or add some additional factoids to the discussion, please do not hesitate to comment below.

((By the way, I believe the FSF vs. OSS distinction is inconsequential because if Linus had not--almost aribrarily I think--chosen the FSF's General Public License (GPL) when he was figuring out how to distribute Linux (he could have chosen the Berkeley license or at least a half dozen other means), the FSF would be simply a footnote in IT history. To the extent the FSF is of any consequence today, it is because of its ownership of the GPL license and not because of technical contributions.))

Posted by dennisb in OSS Culture | Permalink | Comments (0) | TrackBacks (0)

One thing I noticed since I started researching open source software (OSS) for the purpose of my IT investment research practice last fall is that you can't tell the OSS players apart without a scorecard. By the very nature of the OSS movement, the structure of the OSS community and its intertwined business model looks like an old piece of pre-Y2K COBOL spaghetti code. This graphic summarizes and compares the basic differences between initial OSS code, and the organization responsible for it, and the sponsoring commercial organization (only a few OSS projects have none, at least only a few that you would care about). Some of the sponsorships are straightforward and some are byzantine (we need a spell checker here).

For example, Mozilla is OSS portal code worked on by the Mozilla Foundation backed by the Mozilla Corporation; Mozilla in turn was originally spun out of AOL/Netscape. And that is one of the simpler examples. The Linux operating system kernel comes initially from the Linux Foundation (see my January 24 post), which in turn is sponsored by H-P, IBM, Novell and others, and eventually ends up in almost every other type of packaged OSS infrastructure code. A spreadhseet listing specific codesets, communities and commercial backers is available on my web site (see link to the right under "My Work Elsewhere"). The spreadsheet uses these categories: OSS applications, OSS databases, OSS middleware, OSS operating software and OSS toolsets and includes accompanying commentary I wanted to keep track of (such as the factoid that the guy providing the commercial backing for Ubuntu has been to the International Space Station as a paying customer on Soyuz).

This is a work in process so I will make a note on this blog every time I make a major change to the spreadsheet. So far I have identified and categorized about 25 OSS codesets by this taxonomy (sorry, that's what old IDC analysts do--and young ones too). There are hundreds of other OSS codesets but I have--in this case--taken the perspective of my investment research customers rather than developers in that I am most interested in OSS codesets that have financial-market potential. But if you would like to add your favorite OSS code, let me know via the comments section below.

Posted by dennisb in OSS Development | Permalink | Comments (0) | TrackBacks (0)

Unfortunately I got “iced out” of traveling down to the LinuxWorld East event in NYC yesterday and today so I am trying to catch up virtually. Ice temporarily locked up my windshield wipers between Worcester and Hartford and driving into the metro NY area with my hand out the window to try to emulate the wipers just didn’t seem prudent.

From a distance, the most interesting news of the event looks to be the creation of an organization called the Open Solutions Alliance (OSA) and the first name-brand endorsement of that organization by Unisys. Such attention to real applications (as opposed to tools and infrastructure) is just what the open source software (OSS) movement needs and four of the members of the group meet that criteria:
• Adaptive Planning—4-year-old suppliers of a budgeting/analytics product
• Centric CRM—7-year-old CRM software provider
• Jaspersoft—7-year-old OSS report-writer community that morphed into a BI application in 2005
• OpenBravo—a 6-year-old ERP supplier
Of this group of four, all but JasperSoft have adopted the OSS business model either partially or totally after beginning as more traditional ISVs. As for the other charter members of OSA, they are tools/infrastructure software developers or software distributors.

I am not sure what “driving the adoption of comprehensive solutions” means but just putting the four applications together would provide a manufacturer or services company so inclined with a fairly complete enterprise applications “suite,” if deployed along with OpenOffice, under OSS license terms (i.e., so-called free). Note that the companies employ various differing OSS license structures. It is also unclear whether the user would have to go to each individual supplier if the user wanted service and support or if some combined entity that is part of OSA could help. Presumably, that’s where Unisys could come in as well.

Initially, OSA says in its press release, it will focus on defining and promoting tools, frameworks, and best practices for deployment and interoperability. In other words, the group wants to reverse engineer the equivalent of ABAP, PeopleTools, or their more modern equivalents such as SAP NetWeaver or Oracle Fusion. I am guessing that wanting their own framework rather than being tied to JBoss or an equivalent is what they mean by “vendor neutral.” But since some of the members are already hooked up with the Eclipse consortium, tying up with IBM Community WebSphere (nee Gluecode) would get the framework-building process moving faster.

OSA said it also supports motherhood and apple pie (and not all members are even American). Motherhood: It will “build “meta-communities” by partnering on projects that involve a variety of companies, communities and individuals to drive innovation and collaboration.” Apple pie: It will “coordinate joint marketing campaigns to raise the awareness of business-hardened, open source applications and solution suites.”

A key factor that Unisys brings is helping the OSS community understand the requirements of a mission-critical enterprise application suite.

Once I thaw out I’ll find out more.

Posted by dennisb in OSS Culture | Permalink | Comments (0) | TrackBacks (0)

Matthew Szulik, CEO of Red Hat, spoke this morning at a Merrill Lynch investment conference. He repeated the Red Hat executive observation from last week (see blog entry of February 7) that Oracle's November announcement about servicing Red Hat Enterprise Linux (RHEL) has legitimized the market. He also spoke about the Novell/Microsoft deal: he says "free doesn't cut it" for his users (those who want "free" can go to fedoraproject.org). Enterprise customers understand they need the service level Red Hat provides for a fee so he hopes everyone just uses up the Microsoft coupons (that get users SUSE Linux). When the coupons are gone, OSS market dynamics will get back to normal.

Szulik highlighted the server-side compute loads now being migrated over to Linux (graphics and telecomms now as opposed to simple things like mail boxes 10 years ago). Because of this, the discount offered by Oracle on a Red Hat Linux knock-off is a small part of TCO and is not affecting Red Hat's business. He tied the Fedoraproject.org activity to the later RHEL release schedule in the same way that Max Spevak had (see my blog entry of February 3).

The Red Hat CEO highlighted Red Hat's own virtualization "incorporated" in RHEL Version 5, scheduled for release the end of this month (as opposed to SUSE Linux with its Xen-based virtualization "on the side") . The EMC move last week to partially spin out VMware has made virtualization a hot topic but Szulik cautioned that virtualization is not trivial (I give him points for reality).

On JBoss, Szulik commented on the organizational process (e.g., departing executives) and the functional advances Red Hat has made to JBoss already (e.g., localization). He cited the great cross selling opportunities because only about 3% of Red Hat Linux systems had JBoss on them. Also he said JBoss was mostly a development phenomenon and top execs (even top IT execs) had very little exposure to it. It is Red Hat's challenge to take advantage of that opportunity. JBoss saved them efforts in middleware development that they had already started; that team was able move over and take over some of the holes left by JBoss management "retirements." SOA is going to take 3-7 years. Again, he gets high marks for reality from me.

He spoke of Red Hat's opportunities by industry and geography, all of which were cited as mainframe or UNIX migration opportunities. He gets more high marks for reality. He feels Linux on desktop success will follow because of efforts such as Mozilla and stateless Linux but that is more of an opportunity outside the U.S. (that jibes with Microsoft's results). Again, reflecting reality, he is not simply talking about knocking off Windows personal productivity users by the millions.

For the integration software user, Szulik's understanding of
-- your needs by industry and geography
-- IT history (he had a lot of references to the past successes and problems of Unix, object technology, development tools, and so forth; problems he says he can help you avoid),
- the fact that you want choice
-- the entire stack and to leading edge functionality such as virtualization (not just Linux)
are all good signs.

NOTE: Early in my blogging schedule, I feel Red Hat is getting too much attention. This is partially because they are one of the few OSS pureplays. But it is also a sign that there few other OSS "solutions," particularly true of real applications. In my opinion, Linux is still at the stage infrastructure software products such as VAX VMS and DG RDOS were in 1975. I will be digging for more solutions this Thursday at the InfoWorld Open Solutions Conference (formerly east coast LinuxWorld). If you are aware of OSS middleware or applications you want followed, let me know using the comment section below.

Posted by dennisb in OSS Business Issue | Permalink | Comments (0) | TrackBacks (0)

View image

Novell and Microsoft added some color and timing today to their plans announced on November 2, 2006 to bring the Windows and SUSE Linux worlds together. Subsequently these plans were blasted by the Free Software Foundation (FSF), Groklaw.net, and other parts of the OSS culture. Assuming anyone in the OSS world will work with them, the Novell WS-Management/Distributed Management Taskforce (DMTF) initiative will have impact on any Windows/Linux interoperability. The other activities are totally in Novell's and Microsoft's control and can move ahead with or without cultural blessings.

If you are in the open source community, using the comments section, let me know if you think it is worthwhile to work on this project with Novell.

Posted by dennisb in OSS Development | Permalink | Comments (0) | TrackBacks (0)

According to Gartner, Microsoft gained share as the operating system of choice in PDA devices in 2006. Suppliers such as Intermec, HP, Dell, Symbol Technologies, and TDS manufacturer the devices themselves. The statistics vis a vis Linux are especially revealing because such "appliances" would be one area where I would think that Linux could make up for lost time vs. Microsoft's operating system dominance (and it would be a much more reasonable market segment to attack than the desktop fantasy of the World Domination 201 crowd).

But there is no relief in these numbers for the OSS players:
-- Linux picked up a tenth of a percentage point but still had less than 1% of the market
-- Microsoft picked up eight percentage points and now has over half the market.
To be clear, there is likely lots of open source software in the RIM, Palm, and Symbian-based devices that hold market share position between Microsoft's 56% and Linux' 1% in Gartner's listing. In fact, there is likely lots of open source software in the Microsoft-based devices. Remember, at least as of late 2005 (and still probably true), the most popular middleware running on Linux was BEA's WebLogic, not JBoss. And conversely, Apache is as popular on Windows systems as it is on Linux. That's open choice.

From an infrastructure software user perspective, these types of statistics are important if market acceptance is important in convincing the boss to go with an OSS tool or application that you think is right for your enterprise. If you are in government, education, or IT services provision, this is probably not an issue. But as I noted in my Feb 7 blog on how Red Hat wants to and plans to grow the ecosystem, for other industries, OSS acceptance is still an issue. I'll keep my eye peeled for more favorable statistics; if you see any first, please comment below.

Posted by dennisb in OSS Business Issue | Permalink | Comments (0) | TrackBacks (0)

Other interesting stuff:

IBM is into building its OSS ecosystem just like Red Hat will be (see earlier blog)

Jonathan's explanation of why KKR "invested in Sun"

mySQL had a great year. Of course, it helps when you don't have to define "great."

Posted by dennisb in OSS Business Issue | Permalink | Comments (0) | TrackBacks (0)

Mike Evans, Red Hat VP of Corporate Development, and Dion Cornett of "Inside Sales" at Red Hat spoke today at the Credit Suisse Disruptive Technologies Conference, a traditional place to break new ground for the investment community. I listened in via the web (using Windows Media Player unfortunately which is against OSS culture; Oracle investor conferences are not even supported by Medial Player). Red Hat didn't directly break new ground but hinted at a couple of areas where a spade will be used shortly:
-- specific channel programs, including "inside sales"
-- new indirect sales model changes pumping the OSS ecosystem.

Red Hat had announced yesterday expansion of its channel activity into Japan and announced last week a new certified services provider program for its entire channel. Despite the fact that the channels VP that was the advertised speaker was not there, Evans and Cornett dove into the subject.

This is important to you as an individual OSS user or someone considering OSS because Red Hat has been concentrating for years on how to spur OSS adoption. Red Hat has helped OSS along as much as a couple of hundred million dollar company can by advising and investing in the ecosystem. The pair sort of preannounced a reseller/referral sales model (IBM almost instantly--one might say purposely--rolled out its own similar program).

Ironically, the Red Hat pair said that Oracle's and Microsoft/Novell's OSS-related announcements late in 2006 actually increased interest Red Hat's offerings and business model. This is the classic "the big guy legitimizes the business." IBM did for both the minicomputer and the PC.

When Credit Suisse asked about the competitive environment, Evans said the ISVs and companies that he talks to as propsective Red Hat customers are "standardizing" on Windows and Linux. They are seeing very little demand for Solaris for example despite the fact that that software is now also open source.

Cornett also discussed RHEL 5's status as a complete platform, not just an operating system. He highlighted virtualization just as Max Spevack did last week at FUDCon in Boston (see blog entry from Friday February 2, 2007). He said the product featured standardizing and security features, espeically for government customers. He also said virtualization will ramp up slowly among the customer base just as is true for all bleeding edge technologies--tell me what you think of that using the Comments button below.

Other factoids:

-- JBoss is now integrated into the Red Hat stack. Now that JBoss is part of Red Hat we can see how much business JBoss was really doing by looking at proformas on the last two Red Hat quarterly reports. The investors asked a few questions indicating that they are disappointed. Database will be a further area where Red Hat will add to its integration middleware stack.
-- From the 30,000' level Evans sang the praises of the open source software development model (leverage on development and distribution). He said OSS now has the critical mass (a global dynamic, real applications, and a maturing infrastructure in products such as Linux, JBoss, Mozilla) to form a real ecosystem.
-- Cornett addressed market analyst-firm press releases about how Linux server shipments have been doing vs. overall server shipment. He says that Red Hat's business does not coordiante well with analyst's view because analyst firms don't understand how to measure "free." I actually have not seen any analyst firms questioning Linux' popularity; the bone with the OSS people is that the analysts say Linux' success has come at the expense of Unix, not Windows (and Evans comment noted above about what new firms are telling VCs confirms that)

I am happy to see that the Red Hat pair punted on an investor question about the FSF/Novell GPL license brouhaha (see Monday 2/5 blog entry). Red Hat supports a couple of dozen licenses and is not going to get itself embroiled in the GPL controversey. Of course, their not taking sides is part of the controversey.

Posted by dennisb in OSS Business Issue | Permalink | Comments (0) | TrackBacks (0)

I can't figure the genesis of the Reuters story about the Free Software Foundation (FSF) and Novell that was released Friday Feb 2. I don't see a press release from either party nor an event where this might have been the topic of discussion. Slow news day I guess but there is nothing new here that I can see. I have asked both parties to comment.

It seems to be all the same stuff that FSF said in November when the Novell-Microsoft deal was announced. That is, Warning: FSF is going to structure the next version of its license so that such deals are not allowed under its General Public License (GPL) open-source-software legal framework.

A lot has to happen before this becomes a major issue for Novell:
- FSF has to put out its new license (a couple of months late now and counting)
- Open source software (OSS) organizations such as Apache or the Linux Foundation or ??? have to adopt the license; many have said they are reserving judgement (not because of the Microsoft-Novell deal but because of other restrictions); others such as mySQL have already rejiggered their use of the existing FSF license so that their products don't automatically fall under the new license
- Someone would have to sue someone I guess to say Novell was violating a license (I believe this iswhy the FSF lawyer is so excited about the prospect; this stuff has never made it into court before; the highly publicized SCO-IBM case is really about UNIX and run of the mill licensing/patent trivia)

I opine: All this is about is what the FSF has always been about. FSF is against you having open choice.

Posted by dennisb in OSS Culture | Permalink | Comments (0) | TrackBacks (0)

FUDCon, officially the fedoraproject.org User Development Conference, came to Boston Feb 2-4. Unofficially FUDCon says it’s all about “con”tradicting the Fear, Uncertainty and Doubt that they say is spread against the adoption of open source software (OSS). Fedora is the Red-Hat-sponsored open-source community with—naturally—an operating-system focus. (Note: next stop is Brussels at the end of February.) The chief FUD spreader was unnamed during the morning sessions that I attended (which I found refreshing given some of the anti-Microsoft diatribes I find on the web). On the other hand, I wasn’t on the pub crawl Friday nite.

The planners’ choice of Boston in February speaks to the seriousness of their efforts. The venue was Boston University (for the third year I believe) and it was a good host. The timing of the event during the first week that I am officially blogging on OSS gave me a chance to do what I plan to do most of in this blog: Talk to and about the people involved in the OSS culture, development effort and business and talk less about the technology itself.

So I was able to hook up for a half hour with Max Spevack, the Red Hat Fedora project leader. In the OSS triad (culture, development effort, and business model), his role is among the dozen most important on the development side of the triangle. The events this month mark the end of the first year of his time in this role and he was proud of the release of Fedora Core 6 in October and the planning in progress for “7.” Given the way the OSS culture works, this FUDCon meeting was a key part of that planning process of course. Max presented a full update of Fedora in the afternoon, which I will blog on sometime in March.

Max made a couple of key points during our conversation that lead me to opinions ebizQ readers might want to keep in mind:
• Spevack Point: The non-Red-Hat attendees represented over 100 people that took a day off from work (many of whom paid their own travel expense if not New Englanders) to be part of the movement.
• I opine: How much of your software is written by people with such passion?
• Spevack Point: They’ve measured over a million installs of Fedora Core 6 using a strong methodology
• I opine: I like this measurement much better than the nebulous “downloads metric” I used to rail about at IDC when someone told me how JBoss had surpassed WebSphere.
• Spevack Point: Latest version has some interesting virtualization features but more to come in “7.”
• I opine: OK. Latest, greatest. These guys are not just rewriting UNIX here.

We’ll keep track of Fedora as it progresses. From a business model perspective, which users should never ignore (anyone who promised the boss their first born if the enterprise would bet on the great Pick operating environment can explain why to you younger guys), the key question is how this helps Red Hat and to what extent it permeates beyond Red Hat into the wider IT supplier community. Outside of government, education, and the tier-one information technology community (e.g., guys who build business or IT services on top of OSS because of its flexibility and cost effectiveness), users are most likely not going to deal with Fedora but with suppliers that deal with Fedora.

Speaking of such services suppliers, you may be surprised to hear that Google was one of the sponsors of the event (which at an OSS conference means they provided the T-shirts). Google helped out along with BU, Dell, Red Hat of course, the Wiley publishing organization, and KDE e.V., an EU-based open-source GUI community.

I opine: rattle that Google connection around in your head for a day or two.

Posted by dennisb in OSS Culture | Permalink | Comments (0) | TrackBacks (0)

Look at the leading information technology (IT) suppliers’ activity today and you will find what’s happening to the Open Source/Linux movement now happened a decade ago in the Open Software/Unix movement. This week’s merger of Open Source Development Labs and the Free Standards Group into the Linux Foundation is Exhibit A. The powers that be could have even used the same press release they used 11 years ago (see press release February 11, 1996) in merging X/Open and the Open Software Foundation (OSF) into The Open Group. X/Open like FSG was a specification and certification “group” that tried to help overcome the by-then multiple forking of UNIX. OSF like OSDL was a consortium founded by DEC, HP, IBM and other corporate players to promote a single UNIX.

Although the 1996 merger turned out to be too little too late, I think this week’s announcement might be meaningful. It is a sign of maturity for open source software (OSS), along with events such as the Microsoft/Novell announcement of November 2006, giving users a choice in the enterprise or in your personal life: OSS where it makes sense, Microsoft where it makes sense, legacy systems and software where they make sense. The open choice idea is reflected on the Linux Foundation’s web site (especially in a Jan 22 New York Times story on the merger) and perhaps signals a change from some of the OSS vs. Microsoft (and Microsoft vs. OSS) tirades that have permeated OSS discussions for the last few years (and as recently as December 2006 with the absurd badvista.org promotion). After all, open source software has been around since the first meeting of Share in 1955.

As a sidelight, I asked the Linux Foundation: “Why highlight “Linux” in the name and not some variation of “open source?”, which would imply a broader mandate. It could be as simple as that this is the place where Linus Torvalds draws his pay. But I suspect it has more to do with not competing with that other group IBM and others created 11 years ago. I’ll let you know if they get back to us with a more strategic answer.

Posted by dennisb in OSS Culture | Permalink | Comments (0) | TrackBacks (0)

By way of introduction, I have been an information-technology (IT) market and product research analyst for a long time. Bull SA, Data General, the Datapro division of McGraw-Hill, IDC, with a few “sidebars” (less than a year each) such as at Kendall Square Research, Raytheon Data Systems, and Noblenet.

You’re correct: short attention span, low marks for commitment, no gold watch for me.

The good news is that I have consistently researched and analyzed the same topic at all of those places—What’s hot in IT now? What’s not and why? Translate that list of companies above into market and product research topics and I’ve covered a good bit of the history of the IT industry: custom-built inventory control in the warehouse and accounting on the desktop 40 years ago; Multics, PARS, virtual memory management, minicomputing, microcomputing, and the Soul of a New Machine (the market for the machine not the book) during the 70s; material requirements planning (MRP), the first integrated office automation software and the industry’s first laptop in the 80s; the open-software movement, the Common Object Request Broker Architecture (CORBA), the debut of R/3 in the U.S., shared-memory high performance computing (not shared processor), and more in the 90s; application service provision, business process management (BPM), the enterprise service bus (ESB), and Web 2.0 so far this decade.

All were the “latest greatest” at the time; there were simply different reasons for me to do the research at different stops (better tech documentation, more effective advertising, cost-effective recommendations to users, vendor product-development planning, actionable investor advice and support in due diligence). I liked doing research for the last reason the best so since the fall of 2006 that’s what I have concentrated on (see www.itinvestmentresearch.com).

And what I found in this latter concentration was that the “latest greatest” in IT today revolves around “open source:”
• Open source as IT culture.
• Open source as IT development model.
• Open source as IT business model.
“Open Source,” I found, was more than just an updated version of the Digital Equipment Corporation User Society (DECUS). I find open source so interesting because it covers the whole array of what was the “latest greatest” IT before it.

So I will try to balance this blog’s entries among
• culture subjects such as the acceptance of the GPL license or the philosophy behind the World Domination 201 treatise
• development subjects such as Eclipse, Apache, Fedora, and the Linux Foundation
• Business subjects such as what IBM, Red Hat, Novell, Sun and so forth—including Microsoft—are doing with the results.
Because this is ebizQ, the emphasis will be on open-source integration software but I will also stray up and down the stack to operating software and applications as the spirit moves me. Because ebizQ readers are mostly users, I will look at things from your point of view.

Finally “open source” means something different to users, vendors, and investors. And it means something different again to me, a market and product researcher. That’s why I am identifying myself as agnostic about the "religious" open source movement; I take no sides in the “movement’s” self-proclaimed almost spiritual battle with Microsoft. I am willing to be convinced of either side's position on a case by case basis. In the end, I don’t think there should be “sides.”

Comment on that or anything else that hits you early and often. I look forward to it.

Posted by dennisb in OSS Culture | Permalink | Comments (0) | TrackBacks (0)