Content Log

Cloud vs. Open Source? I don't think so!

2012-02-10T17:21:46+00:00

Photo Credit: aufziehvogel2006 (Kai) at Flickr

There are some who are asking if cloud will have an impact the growth of Open Source. For instance, my fellow AIIM board member Lubor Ptacek from OpenText suggests in his predictions for 2012 that "the Great Open Source Movement will hit the trough of disillusionment in 2012 - courtesy of the cloud." Presumably Lubor means Content Management in the cloud, because the cloud runs on Open Source. Operating System? Open Source (Linux as in LAMP)! Database? Open Source! Web Server? Open Source! Big Data? Open Source! App Development? Open Source! The cloud runs on Open Source and not Open Text -- or Oracle or IBM or Microsoft. I am assuming that he means the Open Source that has been a threat to his business of ECM (OpenText) or WCM (Vignette) and therefore he means Alfresco for ECM and Drupal and Joomla for WCM.

At Alfresco, Open Source offers no hype cycle "trough of disillusion"; we only seem to be seeing the "Slope of Enlightenment". Our Open Source Community platform continues to grow apace driven by new releases, such as the new Alfresco 4. There are now over 3.5 million downloads of Alfresco and the number of Community installations is in six figures. As a business, bookings grew 60% year over year in our last fiscal quarter ending in November. This growth is based upon significant revenue figures, not small numbers, and we are now the second largest Open Source company by revenue after Red Hat. We have a real business with over 2000 significant customers using Alfresco for mission critical information in many important industries. If that is disillusionment Lubor, I would hate to think on which slope OpenText is right now!

There is nothing to suggest that cloud and Open Source are mutually exclusive. The cloud is built from open source and if open source is architected to run in the cloud, it will. At Alfresco we have spent the last several years building multi-tenancy and scalability features into the Alfresco system to ensure that it can run in the cloud. This January we launched our cloud offering into beta and expect it to go into production later this Spring. The cloud will become yet another means of using Alfresco as well as trying the product before you buy. I have already articulated the use cases that we think will be important for running in the cloud. In addition, since Alfresco is capable of running an entire system on a single virtual machine or Amazon AWS instance, Alfresco is already used in many private and on-demand deployments. Rather than a threat, the cloud becomes a really big opportunity!

So rather than the cloud being a threat to Open Source, I presume that Lubor means cloud-based file sharing services such as Dropbox, Box, Huddle and others are a threat. Some of these are on the sharp edge of the Hype cycle and are gaining a lot of marketing oxygen as a result. There is no doubt of the popularity of Dropbox compared to anything in ECM, there are reportedly in excess of 50 million users. However, Dropbox is very complementary with Alfresco, which is why we will be providing an integration to/from Dropbox in the near future. Box is a company with more enterprise aspirations, comparing themselves to SharePoint, and although we rarely see them in the market today, we may in the future. However, Box is nothing like a SharePoint, or Alfresco or any other real content management system, as it lacks important features for content and document management such as metadata, workflow, rules and more than rudimentary policies. We shall see how cloud-based file management services and content management in the cloud evolve over the coming couple of years based upon the use cases I mentioned earlier.

As with on-premise proprietary systems, Open Source offerings in the cloud can provide advantages over closed cloud-based offerings. Rather than being the domain of tinkerers, Open Source is mandated as preferred first choice by many governments and organizations. Open Source is transparent and you know exactly how it works. Unlike running software on-premise, you have no idea how software works in the cloud, how it is encrypted or how your data is being stored. In addition, Open Source means that you are not locked into the software that you are running, whether it is running in the cloud or not. You have the choice to deploy it elsewhere and there is not cost to doing so. Open Source as a development model means that the Community is creating new solutions and news extensions, many of which will be very useful for cloud-based solutions. I look forward to sharing some of these with you in the coming months. Far from being at a disadvantage, I see Open Source as being a distinct advantage in the Cloud.

I think a more interesting question is how both cloud-based file management services and open source content management affect the existing ECM sector, which is already stuck in neutral. Having the flexibility to move into the cloud, even a private one, offers new opportunities to explore new applications, deployment scenarios, and business models. As far as I am aware, OpenText, Documentum and FileNet (correct me if I'm wrong) are incapable of running in the cloud. They all require multiple servers running different complex configurations just to get a single instance up. Last I heard, OpenText has a fragile url configuration that couldn't possibly work in a virtualized, cloud environment. And last I heard EMC Documentum was not even certified to run on VMware. I doubt we will see a public service of these systems in my lifetime. What use cases and business opportunities could this preclude?

Seems to me that ECM's "Plateau of Productivity" is tailing off and that represents threats for the large guys and opportunities for the new. I am actually very excited about the cloud and what it can bring and I have never had more faith in Open Source as a way to address those opportunities.

Symbiosis of Cloud and On-Premise Enterprises

2012-02-07T16:06:02+00:00

Symbiosis: One is free to move, the other is not. But they both need each other.

Eventually we will see computing as a utility like electricity that will be consumed on demand. I also believe that cloud infrastructure providers will ultimately be able to undercut the cost advantages and carbon footprint of the data centers of even the largest companies. Just not yet.

As Gartner’s 2012 Predicts, Cloud Computing dated December 8, 2011:

"The perceived (and real) risks to the technical reliability, integrity and security of applications and data that are entrusted to cloud providers hold back organizations' initiatives."

For many CIOs, we as an industry still need to overcome security issues, build trust from the IT organization, ensure open exchange of information, and even negotiate treaties on the rights of information domicile and ownership.

However, there are benefits to using the cloud and even the public cloud starting today that can transcend these issues. There are several use cases that require a cloud platform, particularly in the context of content and collaboration, that would not be possible or at least very difficult today. The first is when sharing information between companies, such as those that are currently described as an extranet, only easier to set up. The second is when information is public anyway and the efficiencies of cloud make distributing this information more efficient as well. The third is where the risk or cost of hacking or intrusion is low compared to a relatively high cost of operating computing resources internally. The fourth is when the need to provide information externally, particularly for the purposes of accessing in mobile environments, outweighs the relative risk of exposure of that information.

Business to Business Collaboration

Credit: Confederation of Indian Industry, Jan 2011

Business-to-Business Collaboration through IT systems has always been difficult when any one side of the collaboration is responsible for the infrastructure. Trying to get the IT organization in a large corporation, even a technology vendor, to set up an external system that is accessible from outside the firewall is nearly impossible—not from a technology perspective, but from an administrative perspective. And there is no way that the IT organization is providing a VPN.

This often leads to the default means for collaboration, which is email. I have been involved in enough standards and board meetings to recognize this is a real problem for even the big IT vendors. Perhaps this can be done with Private Cloud instances outside the firewall, but that again would require thinly stretched IT resources. Yet this needs to be done when two or more companies cooperate on development of products, joint bid offerings to customers or coordinated marketing campaigns. At the end of the collaboration, the results need to be owned by everyone, not just the one who set up the collaboration.

Public Information

Information that is public by default provides another opportunity to share in the cloud. When I first started broaching the subject of Cloud Computing with CIOs and IT departments, I was met with a bit of incredulity about why anyone would want to do this. However, the idea of putting information that was already or should be public in the Cloud didn't really meet much resistance. After all, this is what everyone is doing with web content, either in a DMZ or CDN and Cloud provides an excellent vehicle to manage that content. Marketing departments want to provide marketing material to potential customers and see that material as wasted as long as it is behind the firewall. Also some government agencies must provide certain information to the public by law. If it is going to be public anyway, putting it behind the firewall is only hiding that information.

Small, Medium and Underserved Organizations

Medium and small size organizations are likely to have an IT department that is swamped, if they have one at all. This is the situation we have for internal systems here at Alfresco, like HR and expenses. However, this situation is not limited to smaller companies as even large companies have their underserved departments with strained resources likely focused on core business processes. Many collaboration and content needs may not get met and it may be impossible to know what the impact will be.

Smaller companies are more likely take the risk of putting content in the cloud or with a SaaS vendor, but increasingly, it is becoming more acceptable for non-mission critical departments to also store information in the cloud or with a service. That is if that service can be shown to guarantee security.

Sometimes speed of execution outweighs the risk of exposure. Many on-demand projects or customer interactions require IT resources "Right Now!" or business will be lost. Limiting the time of exposure in the wild or public sphere may help, but if business demands immediate access and IT cannot provide that, then organizations will seek ways to make it happen.

The Mobile Worker

Finally, there is that nether world outside of the office, but still in the domain of the enterprise. Increasingly, we are doing more and more work outside of an office. Right at this moment, I am writing this in my home office, away from any distractions. I am often on the road working - in the air, in a hotel, in a convention center - on presentations, plans or reports, much of which I would just keep in the enterprise. A similar situation exists where workers are out in the field or at a customer site, where working outside the office is essential. Increasingly, this work is being done on tablets and mobile devices.

These are situations where you would want to have the mobile device or laptop be seen as an extension of the firewall. However, using VPN connections is not always practical and can create their own security concerns. If you could put and store the content and materials you need to get the job done in the cloud, you would not only make yourself more productive, you could raise the productivity of the entire organization. This of course assumes that the content is secure and perhaps available only for the duration of the task at hand. The value of the work should exceed the risk of exposing the content.

Content Symbiosis

There are many more situations where you wouldn't want to put content into the cloud. Any corporation that is regulated should consider documents and content under regulation and be stored on premise. Highly sensitive information such as board documents or legal documents should be maintained behind firewalls. I know of two very large corporations with very popular cloud services that are using Alfresco to store their documents behind the firewall for this reason. Records whose final disposition is managed and destruction guaranteed should be stored in a controlled environment. Mission-critical documents that must be absolutely accurate and must not be tampered with should also be considered behind the firewall.

However, these new use cases are so much easier in the cloud. They become the reason to start to move to the cloud and actually add value to the enterprise systems to which they are connected. As Gartner puts it:

"Users should plan for a gradual shift from on-premises IT architectures toward a hybrid model in which these architectures coexist and interoperate with public-cloud-based architectures. For most midsize to large user organizations, a shift to a totally public-cloud-based strategy will not happen during the next five years."

I have heard the term "Hybrid" used for this combination of In-the-Cloud and On-Premise enterprise structure. I prefer to use the term "Symbiosis" since these environments are very complementary and help maintain the health of each. The more symbiotic these systems are, the more effectively they will work. Just like symbiosis, the cloud and on-premise may be very different, but have evolved to derive mutual benefit. If you remember your high school biology, this is called Mutualism. From a content perspective, we can see how these use cases of on-demand content services, enabling mobile working and facilitating B2B collaboration can help the enterprise thrive. Likewise, the enterprise is the protector of information and the producer of content from internal processes.

Like any symbiotic relationship, these systems will reach a balance. As I said previously, I think it will be a long time before the cloud becomes a dominant component. However, equilibrium of 80/20 where 80 percent is in the enterprise and 20 percent is in the cloud could be something we see in the next few years. We are far from that equilibrium at the moment, but the signs are there. If we look at this as a symbiotic integration of systems, it's easier to see how we get there.

Our next lesson on Content Biology will be about “Content Osmosis”.

Content Monkey Predicts 2012

2012-01-31T15:31:19+00:00

It's the last day of January, so I am not too late for predictions for the new year. Besides, it's my birthday and what a great time to predict the future. This year I am influenced by a book I received for Christmas, "Thinking Fast and Slow" by Nobel prize laureate Daniel Kahneman, from our one of investors, Accel Partners. Every year they send out a profound book to their portfolio companies and boy is this one potentially life changing.

I won't go into the book in detail, but one of the things that Kahneman, who is a psychologist and decision theorist, discusses is why experts are so bad at predicting the future. There are a few reasons including some things called the Theory-Induced Blindness, Optimistic Bias, Hindsight Bias and the Planning Fallacy among others. But put simply, experts just try to be too clever.

Kahneman referenced the 20-year study by Philip Tetlock at the University of Pennsylvania that studied 284 experts in economics and politics and 80,000 predictions. The experts were were asked to rate the probabilities of three alternative outcomes: the persistence of the status quo, more of something such as political freedom or economic growth, or less of that thing. As Kahneman states:

"The results were devastating. The experts performed worse than they would have if they had simply assigned equal probabilities to each of the three potential outcomes. In other words, people who spend their time, and earn their living, studying a particular topic produce poorer predictions than dart-throwing monkeys who would have distributed their choices evenly over the options. Even in the region they knew best, experts were not significantly better than nonspecialists."

Tania, the dart-throwing Content Monkey (attribution: Quixado on Flickr)

So there you have it, if we want to know what will happen in 2012, we need a dart-throwing Content Monkey to decide the future. All you analysts please take note, she (I asked her what sex she was) will _probably_ (and the operative word is probably) be more successful than you will be. Her name is Tania Majerus, see http://www.kleimo.com/random/name.cfm and I asked her a series of questions related to "either/or" answers or "less than, same as, more than" answers. I will add my "expert" analysis of the result, because Kahneman suggests that we humans need to think causally and will always look for a cause even when events are random. I swear these are the answers that "Tania" had generated.

Will the ECM market still exist by the end of the year?

Content Monkey says "Yes!"

I guess a market that is some $4 billion or so wouldn't disappear overnight, with a year being classified as overnight. There is some talk about it disappearing or perhaps being absorbed into the infrastructure or being merged with the social business sector. It may not be exciting, but it's still there and still very useful.

Will Mobile be the next big thing in Content Management?

Tania says "Yes!"

It's certainly early days, but with the spread of "Bring Your Own Device", users are certainly challenging IT and the organization to accommodate these mobil devices. Smart phones first and then iPads are starting to create their own demands on the usage of content. I see that OpenText has some sort of offering that bridges mobile devices to internal content management systems. It's not exactly a Cloud offering, but I can assume that other ECM vendors will follow suit.

Will Cloud be the biggest influence in Content Management?

CM says "Yes!"

I guess it is only natural to assume this given the high degree of interest in various surveys of CIOs. We have CEO saying that they "must get into the Cloud" whether they know what that means or not. Flexibility is the clear driver more than cost advantage. The largest corporations say that they can better deals on their own systems than they can from Amazon, but they can add machines on demand. The big question is whether it will be private or public clouds that will have the influence on Content Management. However, I'm not really sure what I would have said if the result was no.

Will companies start to put content into public cloud offerings?

CM says "No!"

If the answer was yes to this one, I would have rationalized this as they are doing this whether they know it or not. People are putting stuff into DropBox and Box whether the IT department and management knows it or not. Still I think we are still a bit away from resolution of a few security issues before companies consciously put content into the Cloud. An AIIM survey last year said that only 6% of corporations are even using a corporate cloud, but that 30% would consider using a public cloud in the future. This may be ignoring the informal usage on the file sharing services.

Will "Big Data" really have an impact on Content Management?

Tania says "No!"

This was one of the big predictions of Real Story Group and Lubor Ptacek at OpenText. To be fair, Lubor said that "Big Data" in general would be the hype of the year, but that no one knows what that really means. RSG has aggregated their analyst predictions, presumably to avoid the failing pundit syndrome, and make Big Data sound like business intelligence. When I think Big Data, I think Map-Reduce and Key/Value pairs in extremely parallel processing a la Google and Hadoop. IBM sees the former as the future and I would think the latter. Still Tania says "Who care?"

Will Open Source market share in Content Management grow, remain the same or shrink?

Tania says "Grow!"

Whew! I wonder what I would have said if Tania had said shrink? I know Alfresco is growing and quite rapidly. I know that a lot of Drupal and Joomla is eating the commercial vendors lunch. But I can't comment on how other commercial vendors are doing. Still, good call Tania! Sorry, Lubor - did you consider that Cloud and Open Source go hand-in-hand and that most of the Cloud runs on Open Source?

Will EMC sell off Documentum?

CM says "Yes!"

There have been interesting rumors over the last year about Documentum. At this point, I can probably count on one hand the number of people I know who are still there. Interesting to see a number of the engineers going off to Salesforce. I had heard a rumor that SAP was considering buying without a lot of substance. I had also heard a very interesting rumor that a strategic acquisition by Autonomy that they had referenced in their earnings call where related to Documentum. That made sense prior to the HP acquisition. Now, you have to be wondering if EMC is thinking more about the potential drives they could be selling more than the prospects of their dwindling ECM brand.

Will SAP enter the ECM market?

CM says "No!"

I'm not sure if SAP would consider whether Tania is right or not. This is an adjacent market that SAP has been reselling OpenText for a while. Could they consider, once again, whether to purchase OpenText? SAP is also building out a cloud solution and CMIS is part of the cloud solution. Even so, they have said it before and will probably say it again that they are not interested in entering this market. And of course, they are using Alfresco in their enterprise.

Will HP grow, shrink or maintain the size of Autonomy?

CM says "Shrink!"

I would consider encouraging Tania to throw this dart again. When you go to the autonomy.com website, it hardly seems like anything is different. That would call for at least a "Maintain". I can see Meg Whitman investing in the service, but she could consider undoing yet another decision of the hapless Leo Apotheke. Autonomy's cloud offering has been somewhat murky from the financial analysts' perspective in terms of what it actually constitutes. However, it may add a bit of Cloud pizazz to HP's line up and it's not likely given the sunk costs of acquisition. Still Kahneman in the book suggests that sunk costs should not be considered in any rational economic decision and that companies are usually very good at removing CEOs who use "mental accounting" to avoid losing sunk costs. Meg's job is to avoid that mental accounting.

Will we see more consolidation in Content Management?

Tania says "No!"

There has been a lot over the last 10 years and perhaps we are seeing an equilibrium of the old guard of enterprise software. This prediction may fly in the face of a Documentum sell-off. I could also see OpenText folding into a larger software portfolio. But monkey says...

Will SharePoint market share in Content Management grow, shrink or remain the same?

Tania says "Shrink!"

Sorry, I can't see this and I can't explain it away. Sure, growth will slow and the Microsoft brand being tarnished by some bad handling of consumer products. But Microsoft Office is still the main game in office software for a long time to come and it's going to be natural for Microsoft to hook in SharePoint as part of the deal as well as SharePoint driving Office sales.

Will OpenText's market share grow, shrink or remain the same?

CM says "Shrink!"

I am going to be accused of being biased of even acknowledging this answer. I personally say "Remain", but we'll see.

Will those anti-Documentum ads that Oracle are producing actually work?

CM says "Yes!"

These ads really piss me off. Sure we compete against Documentum as well, but I can tell you that 8-Track tapes were well and truly dead when we started Documentum. Kids these days probably think we were all living in black and white and called each other "Man!" And can you tell me that Stellent's architecture is really that more advanced than Documentum when they were started within 5 years of each other 20 years ago. Anybody could get great results if they use an Exadata. Just give us a machine and we'll show you.

Updated Bonus Question: Will 2012 be the Year of CMIS?

Monkey says "Yes!"

I guess I have been taking CMIS for granted, otherwise why wouldn't I have asked this question in the first place. It has come a long way over the last several years. Apache Chemistry has made it possible for so many to implement it and Chemistry also means that the lighter weight JSON bindings (think brower plug-ins or mobile apps) can connect so much easier. As the Cloud becomes more important, what better standard to have to connect to content than one that was designed for that purpose and has so many systems like SharePoint, FileNet and OpenText supporting it.

Finally

My final prediction, is my personal one, which is these predictions will beat the others that are out there and that they will have around a 50/50 chance of being accurate at the end. Unfortunately, I won't be able to take credit.

HP's Board: What were they thinking?!

2011-08-31T16:57:17+01:00

What are they thinking? HP buying Autonomy! It took everybody by surprise, probably because it isn't the most logical of moves in today's industry. The last rumor I heard connected to Autonomy was nearly a year ago that they were considering buying the Documentum (or should I say "Intelligent Information") division of EMC. But HP giving up on mobile and the PC business to go into search, content management, customer engagement and a much bigger, fuzzier information management realm?

As a product strategy it could have been kind of interesting. Get out of the commodity PC and device business, concentrate on servers and build competitive advantage on the software of the most valuable assets on those servers. It's really too bad they didn't think through the whole Palm acquisition before starting this, but it could have been disruptive. However, this acquisition makes no sense on a lot of different levels. There are three things that will make this impossible - personalities, dis-economies of scale, and trying to do this in a post-technology market.

First, the personalities. I could just about imagine Mark Hurd making this work, maybe. However, Leo Apotheker? Did the HP board not realize why Leo was "asked to leave" SAP after a little under a year and a half in the sole CEO job. Perhaps they attributed it to a fatal flaw in the Hasso Plattner-dominated SAP organization. In reality, he was demotivating the whole organization with unclear direction. He was a competent technocrat elevated beyond his abilities in one of software's most demanding leadership positions. Hardly the leadership needed to integrate and realign a major Silicon Valley legend. In addition, his actions in relation to lawsuits from Oracle. Is run and hide really a strategy? Compare this with the decisive way that Larry Page has handled lawsuits against Google. On top of this swing from we are in mobile and now we are in only in servers, can we really discern vision. What are you going to think if you own HP hardware, HP software or Autonomy software? Is this guy going to stick with it?

Best defense is to run and hide? An enthusiastic receptionist at Oracle shared this with me last year.

The other personality is Mike Lynch, or Doctor Lynch as he apparently insists on being called by staff. Mike is an extremely intelligent man and has steered Autonomy well through the economic downturn. Creating the largest British tech company, he should be an inspiration to all of us British software companies. He brilliantly positioned the mathematical musings of an 18th Century British theologian, Dr. Thomas Bayes, and turned it into the branding of a must have piece of software. Software that was originally designed to detect submarines. Simply brilliant marketing! However, the man can be very difficult. My colleague John Powell and I met with Mike Lynch ostensibly to OEM some software from Autonomy. What we got was an hour lecture on information management and then finally toward the end he explained that IDOL was everything we needed without listening to our requirements. Autonomy's sales guy saw us out at the end of the meeting and apologized. We didn't really see much reason to proceed.

"Call me the Doctor." Michael Lynch: photo CIO Magazine.

Combine these two personalities and how will the two work? Will Lynch subordinate himself to Apotheker? Will Lynch take over? And how can you swing a giant supertanker of a company like HP in an entirely new direction quick enough? Not only are you jettisoning whole product lines, but you are putting existing and future integrated products into an uncertain future. HP as a hardware company has not exactly distinguished itself in software. EMC actually acquitted itself in software after the Documentum acquisition by the sheer charisma and software sales/marketing expertise of Dave DeWalt. Although both Apotheker and Lynch have tons of software experience, their leadership styles and personalities seem misaligned with the realities of try to change the direction of HP.

HP, with its large size, organization and bureaucracy, will not be able change quickly for all the reasons of "Innovator's Dilemma", "Competitive Strategy", "Innovation and Entrepreneurship" and all those other business books that say it just isn't going to happen. This isn't Larry Ellison or Bill Gates saying "Just do it!" These are both interlopers interfering with a Silicon Valley institution. At this scale, you aren't going to get new solutions that combine the best of both HP hardware and Autonomy software to create magical new software for the Cloud. Despite Autonomy's claims that it does 30% of its business in the Cloud, that business is actually.... It is not clear that anyone is using or even capable of using the IDOL suite as part of SaaS solution. Is the vision to just take same old hardware and same old software and host it in a farm somewhere. Now is the time for innovation for new solutions that deal with the coming privacy, security and scale issues. However, that is likely to be done by new startups. Perhaps the HP of old could have done it. Not this one.

Is this the Exxon Valdez or HP changing course?

Finally, IDOL (Information Data Operating Layer) is a complex concept and it is the basis for the potential new range of server-based offerings. It is also the basis for Autonomy's cloud offering, supposedly the reason why HP is buying Autonomy. As Nick Patience notes in the 451 report, it's not super clear what Autonomy's cloud offering is, how much of the business it really represents and how big a growth factor it is. Is it "IDOL Cloud", is it Zantaz SaaS, is it some other offering that allows HP to play in the new Cloud world?

Complex concepts don't sell well now. We are in a post-technology world where people are going to ask what are you doing for me. HP/Autonomy can start going on about Bayesian distributions et al. However, what HP/Autonomy will get from customers is: does it work on my iPad and does it work with Google. Playing a full stack game against IBM and Oracle is a hard defensive or offensive move with a confusing message.

I love Cringely's thoughts written back in February on who is going to take over this mess. But maybe he's wrong and maybe Mike Lynch will get the job. If you go by the Wall Street Journal's Al Lewis, he's just going to kill the company.

A Month in the Bay Area

2011-08-31T15:49:21+01:00

Sixteen years ago, my wife, my 9 month-old son and I left the San Francisco Bay area to live in the UK. I was still at Documentum, but I had always wanted to return after starting a company, which ended up being Documentum. Since then I have constantly flown back and forth between the UK and SFO.

Friday before last, I returned from one month in the bay area with my family. This is not something I tweeted about or put onto Facebook. There were enough warnings about this on the Internet to make sure I didn't say "Hey burglars! You still have 31 days left to rob my house!" In fact, I unintentionally made it a one month holiday off of Twitter, LinkedIn, Google+ and most things SocialNetwork.

That's me on the left - that little dot next to the Golden Gate Bridge

The purpose of the trip was to re-establish connections in the Bay Area (I'm a Cal grad!), to network in Silicon Valley and do a reality or "sur-reality" check on the state of technology. As William Gibson noted "The future is here, just not evenly distributed." Living in Europe has its own interesting sur-realities and in most cases stuff just moves faster (too fast?) in the valley. I have tried to take an extended period of time every 5 years or so and the last time was much longer than that. Besides, I have a lot of family in California and a day trip into SFO or San Jose just doesn't allow any sort of good catch up time.

I guess the way to think about it was that I was doing good old fashioned social networking, not internet social networking. I also had a few objectives that I wanted to accomplish.

First, I wanted to understand how consumer and enterprise were intersecting. As Geoffrey Moore noted in working session with AIIM, end users of software are asking "Why am I so powerful as a Consumer and so lame as an Employee?" The conclusion that most enterprise software companies have come to as the antidote is to make their software more like consumer software. "We are the Facebook for the enterprise!" they will exclaim. I have never been convinced of this and wanted to get other opinions, particularly of consumer internet companies. I also wanted to see if the traditional enterprise software guys are staring to get it.

Second, as we are approaching the launch of our Software as a Service later this year, I wanted to explore what the best practices are in operating SaaS in the Cloud. What are the adoption rates? What are the hardware requirements? What are the best practices in pricing and billing? How do you value a Freemium service? Etc., etc., etc.! Interesting there is more experience of this in San Francisco than on the peninsula. A real reversal of fortunes from when I left the bay area in 1995, not that either is doing too badly!

Third, I wanted a first hand observation of the Clockspeed effect first introduced to me by my friend and Documentum co-founder, Howard Shao. (It's too bad he was skiing in South America, which he always does during the North American summer.) The book Clockspeed (http://www.amazon.com/Clockspeed-Winning-Industry-Temporary-Advantage/dp/0738201537) by Charles Fine says that markets consolidate to a point of unsustainable Oligopoly or Monopoly until their whole ecosystem is exploded by a new wave of either vertical or horizontal integration. One example given was the horizontal integration of the Wintel platform that is now being challenged by new vertical integration platforms. These new vertical integration platforms have everything to do with cloud, mobile, social and consumer. This is going to lead to radically new ecosystems and I wanted to meet and work with the participants in these new ecosystems.

Finally, it's hard to overemphasize the role that personal life can have on professional life and the need to balance the two. I am of a generation that didn't do this very well and I can only try my best. Normally, I fly into SFO, try to see some of my siblings, niece and nephew, and the then head out to non-stop meetings as jet lag takes it toll. This is no way to do either. As my sister said last year, "Of course John, we love to meet with you and watch you fall asleep over dinner. :-)" I not only want to get some quality time accomplishing my professional goals, I wanted to see old friends and spend time with my family. I wanted my kids to see where we used to live and what it's like living in the bay area. In the end, I could have done better on this and everything really, but the trip was very fulfilling.

I will be blogging (I hope) about the first three points over the coming days and weeks, but not the last part. If we are connected on Facebook, you can learn all about that side of the trip. However, in the meantime, I thought I would include this photo of some of the old-timers from Documentum. We were able to meet up at the Hopyard Brewery in Pleasanton. It was great to see everyone so well and so healthy given how old we all are. ;-) I hope we can get more people next time I do this.

Jim Donahue (Dr. Object), Kendall Smith, Nicki Crosswhite, Dana Ford, Roger Kilday, Lalith Subramanian, John Klassen, me, Cathy Herbert, Jeff Miller.

Alfresco launches new Activiti Business Process Management Initiative

2010-05-17T14:52:56+01:00

Today Alfresco launched a new open source project, Activiti Business Process Management Suite along with the Spring Source division of VMware, Signavio and Camunda. We are also very pleased that Tom Baeyens, project founder of JBPM and BPM expert, has joined Alfresco along with his fellow architect Joram Barrez. They bring a wealth of business process experience to a clean slate to build a next generation BPM system that will be licensed under the Apache 2.0 license. This combination can have profound implications for both the business process and content management spaces.

Tom Baeyens

Activiti emerged from our desire to have an Apache-licensed BPM engine. Although we were quite happy with the jBPM engine, it's LGPL license was preventing us from OEM's Alfresco to larger software companies that were concerned about any open source license with the letter G in it. It's irrelevant that they shouldn't be concerned about it, we intended to take care of it. It's understandable that RedHat did not want to change its license, but our business needs dictated that we needed to find an alternative.

At the same time, Tom also felt that jBPM should be Apache licensed, but for different reasons. The OMG was making strong progress toward BPMN 2.0 as not just a modeling language, but adding execution semantics to create a language that could be used both for design and for execution. He felt that with the primary design vendors using this tool, the time was right for an Apache-licensed BPM engine. An Apache-license could take the BPM system to where BPM was needed, which was everywhere!

After a very short discussion, Tom and Joram joined Alfresco just a couple of months ago and they are already ready to put out their first alpha version of the Activiti BPM engine. They have both been working on this so long, they can create an engine from scratch in their sleep. However, the blank slate has allowed them to think about some of the issues that weren't even around when jBPM was created. For instance, how can you use NOSQL and eventual consistency to create an engine that can scale into the Cloud? What kind of role can scripting through JSR 223 play to enable the use of languages other than Java, such as Groovy, JRuby or JavaScript? How can you create open touch points to content management through CMIS and the OpenCMIS API in Apache Chemistry?

By answering these questions, Activiti is addressing the requirements of business process management for new applications. The Activiti engine as small as a few classes that are embedded in your application or as big as an internet and consumer scale engagement server. Applications that wouldn't have even considered a large scale, stand alone workflow server because of cost and complexity will now be able to freely embed a business process engine. However, new Cloud applications

In Activiti, we don't only have a BPM engine. There is a complete suite with the engine including a designer from Signavio, user tools and control consoles. Signavio had already been working with BPMN 2.0 and had an MIT license for their designer. Although they have working relationships with other ECM vendors, they are quite happy to work with the wider open source community. Their browser-based and AJAX approach to BPM design will make process design more accessible to non-developer business analysts. This approach and the future proofing and lock-in removal of the BPMN 2.0 standard may be one of the most revolutionary aspects of the Activiti project.

Activiti Process Modeler

Our intention is submit the Activiti engine to the Apache Foundation. We have another new employee, Nick Burch, who is the lead for Apache POI http://poi.apache.org/ and will help guide us through the Apache process. Nick has already been instrumental in moving the OpenCMIS project into Apache Chemistry. We shall provide support for Activiti when it is used in conjunction with the Alfresco engine, but do so in a style that is consistent with an Apache project. We also intend to continue support for jBPM for our Alfresco Enterprise customers as well as connections to other process engines, but Activiti will become the default BPM engine for Alfresco.

Activiti is available as Alpha now at the Activiti web site. Screenshots are also available. We are looking at General Availability at the end of the year. We are looking to incorporate Activiti into Alfresco in a release at the end of 2010.

We hope to see more people involved in Activiti. This is as much a good thing for our open source competitors as it is for us. With a common set of interfaces around CMIS and OpenCMIS, this will allow all of us, including those built upon PHP, to share common workflow capabilities and tools. Who couldn't be happy with that? :-)

There have already been a couple of interesting blogs from:

Tom Baeyens, Alfresco Creates Activiti
Matt Aslett, Alfresco’s new Activiti en route to Apache
Sandy Kemsley, Open Source BPM with Alfresco’s Activiti

A shift in Alfresco Community license to LGPL

2010-01-27T19:14:22+00:00

First of all, it's my pleasure to recognize Alfresco's 5th birthday. In January 2005, a group of eleven of us started on a fantastic journey from a familiar world, Enterprise Content Management, into an unfamiliar one, Open Source. During this time, we have re-kindled our excitement for ECM, learned a lot about open source and become the largest private open source company. Not bad for a little company in suburban England.

One of the things that we have had to learn and be flexible on is the whole area of open source licenses. It's been nearly three years since we went to the GPL license with Alfresco Community. The GPL was and is the most widely used open source license and provided a fairness of use that meant that we could feel comfortable to grow the project, company and brand. Alfresco is now one of the strongest brands in both open source and ECM.

Now we have grown the Alfresco brand considerably since its beginning, we believe that we can now move the Alfresco repository to the LGPL license. This moves us back to the model that is similar to the JBoss license model (see JBoss's "Why We Use the LGPL"), which we experimented with in the very early days of Alfresco. Compared to 2005, we see more of an opportunity to be a platform beyond individual applications, particularly with the emergence of CMIS. What the LGPL license provides over GPL is the ability to link in the Alfresco repository without affecting proprietary software that links it. As stated in JBoss's document:

We use the LGPL for JBoss because it promotes software freedom without affecting the proprietary software that sits alongside and on top of JBoss.

This reflects our feeling too. The LGPL code is share and share alike, but you can link it with any proprietary code without affecting the license of that code. We have considered more liberal licenses as well, but we currently have two main LGPL components - Hibernate for database access and JBPM for workflow - which prevent us from going to something like Apache or BSD licenses. However, this is something we may consider changing in the future.

We do this in the spirit of making Alfresco available as a CMIS platform and a general ECM platform to build content applications without inhibiting your business opportunities. What we hope is that your applications will build demand for Alfresco services from Alfresco Software, particularly in larger enterprise environments.

What this means practically, is that we are changing the license of all the software in the alfresco.war file. This is not an overnight operation, since every single files header needs to change to reflect the new license. Thus the license officially will change with the Alfresco 3.3 Community release in March. If you wish to consider alfresco.war today as LGPL, you may do so.

If you are an enterprise customer, then this won't affect you. You still have a full commercial license from Alfresco and have the full freedom of those terms.

8 ECM Predictions for 2010

2009-12-31T16:44:42+00:00

At the recent AIIM Board Meeting that I attended this month, AIIM President John Mancini asked us to blog using the number 8. Why 8? Because no one else is using that number and AIIM can brand around it. So what better way to start than with my 8 predictions for ECM and Open Source - emphasis on the former. I found that I had a lot more that I could have written, so I cheated and added two more on section 7+1 and section 8+1 for a total of 10. Still, I didn't get around to my thoughts on Apple, mobile, Google, Google Wave, New Applications, Mergers and Acquisitions, etc. Since I have to go home and I am the last one in the office, this will have to do! I know it's a lot, but I have been doing a lot of thinking lately.

In no particular order...

1. Economy's affect on ECM.

Nothing affects the ECM industry more than the economy for good or bad. Any sudden downturn has an immediate effect on practically all enterprise software spending regardless of the payback. Until the consequences of the downturn are understood, financial controllers will hold off on purchases above a certain amount. Generally bad for large enterprise purchases. On the flip side, a recovery can and has been an excellent opportunity for ECM as automation is often a better investment than bringing on the old hires that you let go. Middling growth leads to middling results.

The prognosis for the economy is not entirely clear right now. Most recessions endure a double dip, although this one was so sharp it may have washed out most of the bad stuff that a double dip gets rid of. There is a great deal of uncertainty for both downside and upside opportunity in the economy and know one knows for sure what will happen. Most leading indicators are good, but a similar situation arose a couple of years after the 1929 crash and the economy dropped again in the mid-30s. If it gets bad, it can very, very bad as over-stretched governments have very little room for maneuvre. In addition, Tim Geithner has said that prospects for increased employment will not be great in 2010.

My guess is that economic recovery will continue and this could be a good year for ECM and enterprise software in general. Regulation is up, corporate purchasing is up, optimism is up. Content management will be an easier sell than most as investment for the future with less risk than hiring. I'd put the probability at 60-70% for good conditions for ECM. Downside is that is still a pretty risky environment. As the Boy Scouts say, "Be Prepared!"

Having said all that, government will be a growth area for ECM in 2010. Stimulus packages will kick in and new regulations will emerge to put yet more content pressure on organizations. This is one of the reasons we invested so heavily in Compliance, Governance and Records Management in 2009.

2. ECM in the developing world.

Growth in Asia, particularly India and China, is obviously outpacing Europe and North America. While the developed world spent, the developing world saved with more reserves for investment in infrastructure and IT. Even eastern Europe is doing better than expected because of lower debt. Could the developing world be a major growth area for ECM. Pressure on space and skyrocketing real estate prices in rapidly developing urban areas means that digitization of information can be a valuable tool for freeing up space. ECM is ready for the developing world.

If I were a traditional software vendor, I wouldn't count on this for revenue though. These guys got to where they are by being cheap and bargaining ruthlessly. This is no place for a traditional enterprise software sales person. Also, the size and scope of the developing world means that traditional distribution channels won't work and relationship selling can take forever to develop. And it is not just software that is feeling this - so do hardware manufacturers, manufacturing equipment makers, even medical device manufacturers.

I think this is an area where open source will absolutely shine. Internet distribution and download is a much more effective way of getting software to people, but will probably develop new models of monetization. We're seeing a lot of activity in India as an example, but demand is going to come in the form of developer support rather than traditional licenses, particularly in various outsourcing houses. Microsoft too will benefit from the extensive reach of its indirect distribution channels, although the software pirates may benefit as much, if not more.

3. SharePoint in 2010.

You have to hand it to Microsoft; they seem to have scared the bejesus out of the traditional ECM vendors. They have all seemed to rolled over and played dead in the wake of SharePoint 2007. Either that or they have just buried their heads in the sand in complete denial. To claim that SharePoint is not really ECM is ignoring the messages that are coming from the podium of Microsoft conferences. Most seem to say - "Oh we support SharePoint, we offer archiving, records management, etc." Yeah, but who owns the data? SharePoint is not just a front end to your system. It is becoming the platform for knowledge worker applications.

This is probably a make or break year for the ECM industry. If you are an ECM vendor, beware the ever tighter integration with office, better web support, new records management, and claims to better scalability and administration. Still, it is not invulnerable. Microsoft has chosen not to address its fundamental architectural flaw of storing content in the database and it is still an exclusively Microsoft-centric platform. Forget whatever database, operating system, language, browser you have - you better get used to SQL-Server, .NET, Windows, IE and don't forget Silverlight. If the traditional vendors can't battle the crap out of that, then they deserve to lose. The next 12 months will be critical during the transition to SharePoint 2010.

4. The E in ECM.

The whole idea of Enterprise software in the 21st century seems anachronistic. The term Enterprise really only took hold in the 90s in order to describe systems that were able to scale beyond the department. It meant big, powerful, flexible, but it also meant big, clunky and expensive. As Web 2.0 sites with their cheap (read free), simple, but scalable platforms scaled to millions of users in a matter of months, the whole idea of only being able to support thousands of users and take years to implement became ludicrous. Being Enterprise meaning you can support your heavyweight infrastructure of other Enterprise parts also seems less interesting when you consider that the largest databases on the planet run on MySQL using a concept called Sharding.

Support the Enterprise concept if you want, but Brand Enterprise has lost a lot of its value. And while we are at it, so is Management. In a world of wikis, blogs, Facebook and Google, a strictly controlled environment is destined to be locked up in a closet with just a couple of people using it. It's not that control is not valuable, but you need to really consider what would really happen if information were freed up. Transparency and tracking are more valuable tools today than locks, keys and authorization slips. Leave those to the information that really, really needs it. However, the Content in ECM will be more important than ever as more and more of it gets created at an exponential rate. Best to guide it rather than restrain it.

So what replaces Enterprise Content Management? I don't know, but my guess is that we start to find out in 2010. it might even have a new name by December. Hint: I wouldn't jump to the conclusion it is Social Software. Also see Application vs. Platform below.

For the sake of readability of the rest of my predictions though, I am going to stick with ECM.

5. WCM and ECM.

What a funny relationship these two have. ECM was born in the early 90s out of document management. WCM was born in the late 90s because ECM couldn't get out of document management. The in 1999 and the early 2000's (Noughties here in the UK - what do you call them?), both ECM and WCM decided that they were in the same business and started buying up each others capabilities and then jury rigging the architectures together. The logic went that both sides were managing content, they had repositories and their way of handing content was better. From a pure dollar perspective, that one didn't play out as well for the WCM vendors. The largest players from that time, Vignette and Interwoven, are now subsumed by OpenText and Autonomy.

Still WCM didn't do too badly in 2009. The web site is the store front of the 21st century. And most apps are web apps and most web apps are web sites, at least as far as the user is concerned. Companies still invested in branding and applications for customers as a way to weather the recession. Because of the plethora of web sites, styles and characteristics, there seems to be a plethora of WCM vendors as well. It must be the most fragmented industry that I have ever seen in the enterprise software space. Due to these characteristics and the fractious nature of WCM, if you see a WCM vendor and an ECM vendor in the same account, often it means that somebody shouldn't be there.

However, I think WCM should play a very important role for and in ECM and hopefully this will become clearer in 2010. ECM as repositories of trusted information predominantly deliver their content via a web interface. If its up to the end user, it would be the web site that they are using. If web applications are presented as web sites, are content rich and want to dynamically deliver content to end users, then the marriage of WCM and ECM makes a lot of sense. This is why Alfresco is investing in its WCM to work well with the Spring Java Framework - to help build web applications that are web sites and provide dynamic content and repository services for content-rich applications.

I hope the mapping of the WCM industry becomes clearer as well for 2010. We believe our positioning of focusing on WCM for Java web applications, particularly Spring applications, means that we can work with other parts of the WCM industry, such as Drupal/Acquia, Joomla and others. In 2010, some other ECM vendors should just go ahead and get out of the WCM business and stick to the knitting. They should cooperate with, rather than compete with the WCM vendors. I think CMIS will make this a lot easier.

6. The Cloud and ECM.

As my colleague Ian Howells has pointed out, the buzz around Cloud is many times more intense in Silicon Valley as it is outside. And it is a level of buzz that feels like the Web in mid 1990s that was not felt outside the valley. Could we be seeing the start of a huge land rush that we did in 1995? There is a lot of commonality with the experimentation, trials and shear enthusiasm.

The economics of the Cloud seem almost no brainers. Two years ago, the cost of running systems in the cloud were 1/3 the cost of on premise. Now according to "Above the Clouds", a report from the University of California on Cloud Computing, the cost is about 1/5 to 1/7. I know it is a stretch to say that if this trend continues we can see computing at 10% and ultimately 1% the cost of on premise, but this is what I believe. The reason is that we are only just beginning to see the industrial scale of Cloud facilities come on line that have enormous buying power, very cheap energy, and nearly free cooling being in cold climates or near cold rivers. The ever increasing bandwidth available means that location of computing is less and less relevant. There are still lots of obstacle to be overcome, particularly in perception of security and reliability as well as real legal issues of data domicile. But the implications of 100X compute power for the same cost as in house is enormous and unknowable.

A lot of people have predicted that 2010 is the year of the cloud. I think that 2010 will be the start of ECM in the Cloud. Steve Ballmer at the SharePoint conference in October made a big point of talking about SharePoint in the Microsoft Cloud. We have been and will be doing a lot of work in using the elastic power of the Cloud with the Alfresco platform. I expect that you will see other ECM platforms working in this area. I would also expect to see lots of open source in the Cloud. This can't be good for

7. CMIS.

I guess I have been talking about this longer and louder than anyone else out there, so you wouldn't be surprised to see me say I think CMIS will have a significant impact in 2010. In the past two years, I have learned a lot about the standards process, particularly OASIS, and I have learned not to be so optimistic when lots of players are involved, no matter how motivated. But I can see the end is near to getting CMIS 1.0 to an official standard in the Spring 2010. It will not be long before the ECM vendors have their implementations out. But I have also seen portal vendors also building portlets as well to integrate with the ECM systems. I would really be surprised if these aren't out by the end of 2010 and already having an impact.

CMIS is both and an opportunity and a threat for traditional ECM vendors. It is a threat because it provides SharePoint in particular an opportunity to ease users out of those traditional platforms and into SharePoint. SharePoint being a value player at the low end will naturally try to gobble up ECM low end implementations. It is an opportunity for a number of different reasons. This is a chance for the larger vendors to consolidate the small holdings of the lesser or defunct competitors. It means that ISVs and system integrators will be able to reuse solutions from other vendors and apply them to other ECM vendors solutions. More solutions will mean more money spent on ECM as it solves real business problems, thus making a bigger ECM pie. This in turn will create more solutions. All this played out with the standardization of the DBMS market and there is no reason to expect that it won't in the ECM market.

Whenever I talk to anyone about integrating with Alfresco, I suggest that they integrate using CMIS. CMIS is very rich and usually good enough to do what they need to do store, access and search content. When the integration is done, it works not just with Alfresco, but any CMIS-enabled repository. We have done this with Drupal, Joomla and Confluence in the last year. Expect lots of early integrations in 2010, particularly with portals. We see lots of Liferay used in conjunction with Alfresco and I would encourage portlets developed by customers to be done with CMIS as an example. Expect CMIS to be on tenders, RFIs and RFPs. Finally, you may see the first companies being formed around CMIS and content applications.

7+1. Content Platform vs. Content Application.

I once had a bizarre conversation with Gartner about what constituted "Visionary" in ECM. The answer is that you buy into their vision of CEVAs (Content Enabled Vertical Applications) and have lots of them. Hmmm - visionary means you buy into someone else's vision? But then I asked why SharePoint was considered visionary when they didn't invest in CEVAs. The answer was because of all the interest from their clients in SharePoint as a platform. Hmmm - so you are visionary buying into their vision unless you are Microsoft. Now I understand that CEVAs are gone and replaced or "rebranded" as Composite Content Applications. Does this align with being a provider of CEVAs or being a platform that allows others to create applications?

My belief is that we are seeing a natural swing that has pervaded content management from the very beginning. Is content management an application or a platform? Is it middleware or a fundamental subsystem? A lot of this has to do with how you create, capture and use content. Some of the process of content management is so generic that it makes sense to have an application to perform specific tasks. However, many so called out-of-the-box solutions are so heavily customized that it really is more a platform than an application. Delivery and consumption of content via the web can either be an applet, gadget, portlet, web part or an entirely new piece of HTML/JavaScript/Java/.NET piece of coding.

Content management was never like an ERP or CRM system - plop it in, conform your business processes to that system and configure the rest. Content management was in many ways something closer to a database. Sure it has a rich domain model, but ultimately customers want to use it for all sorts of different applications. They want to write queries, build relationships, define business processes and create rich and elaborate user interfaces, particularly for the web and web sites. The better way to look at it is, that some "CEVAs" may be considered applets that are part of a larger application or business process. ECM systems should make those applets work independently as portlets where they can be provided in context. Capture and consumption can really come from anywhere, so provide the APIs that can provide the content access and manipulation required in the language and development environment of the application that needs the content. The reality is that applications need content as much as they need data in a database, so let's give them the same sort of tools that they have with databases. If portions of the application can be packaged as applets, that is part of the job done, but not all of it.

Three trends will get us to think about ECM as platform more than application in 2010. First, CMIS as a platform-style API will get us to think of ECM and repositories as a platform. CMIS will highlight the areas where we need to build and extend applications by initially system integrators creating solutions with CMIS that will be portable to other ECM systems. Second, the desire to build mash-up applications as a result of experience with Web 2.0 will encourage developers to be pragmatic about pulling in the right tools, including ECM into mashed up or Composite applications. The ECM system is not necessarily to going to be the locus of the Composite application, the web site may. From this perspective the ECM system will be perceived as a platform in support of the web site or web application. Finally, the 5 parts service to 1 part product cost ratio (plus or minus) that has been around since the early days of ECM can not endure in our cost conscious times. If we stop pretending it is a whole app and provide the tools that developers need to pull together the whole app in a fraction of the time, particularly with Web 2.0-style scripting over hard core coding, everyone will benefit.

This is not a massive swing of the pendulum, but it is heading in the clear direction of platform. This is where we are thinking. This is also where Microsoft is thinking based upon comments from Steve Ballmer at the SharePoint 2009 conference.

8. Open Source makes strange bedfellows.

We go into 2010 as what a lot of open source experts have described us as the largest private open source company and cash generative to boot. Not bad going as we celebrate our 5th anniversary in January. In those 5 years, we have learned a lot about the politics, business and religion of open source and it has been absolutely fascinating and fun. We have had to think about licensing, pricing, subscriptions, alliances, packaging, downloads, new marketing models, confrontations and who are friends and enemies are. Throw a recession in there as well and it makes for interesting times.

As time has passed and we think about how we can apply the open source model, we have focused more on our core competence - Content Services and Content Repository. The more CMIS has become real, the more we can cooperate with others with out locking either party in. Thus we can move away from competition with many players and become best of breed in content services. This has allowed us to more work with a Drupal or a Liferay when in the past there may have been overlap in our product sets. CMIS opens up even more opportunities to work with others as well. Working with others creates a stronger and multiplicative network effect.

Because open source stands parallel to other stacks and encourages others to integrate, particularly through contributions. This opens up alliances that may not require integration work because it has been done by others. We didn't have to build the integration to Kofax or SAP and you will see more integrations in early 2010. Portals - open source or closed - have always worked hand in hand with content management, while not necessarily encroaching on each other's space - SharePoint excepted. These types of integrations in turn allow us to work with other vendors that may need to bolster part of their product set and not feel threatened. Expect to see more integrations and alliances between open source vendors and closed source vendors in 2010.

This is not new for open source. MySQL and JBoss have worked with many non-open source vendors and bolstered their capabilities. In 2010, this will take on a new significance though. The Cloud is requiring many vendors to bolster their product sets for new models of demand and elasticity. The growth and threat of SharePoint and Microsoft product sets can make some products seem incomplete. As we see in the next section, Social software and content management can be very complementary. The ease with which it is now possible to pull together an entire stack of software means that a lot of that stack may be open source and present some really interesting offerings in 2010.

8+1. Social Software and ECM.

Should you buy social software from your ECM vendor asked CMS Watch. Conclusion was maybe not. I personally think that social software will be a category in itself with vendors whose core competency will be social software. Just like any web site could have been Facebook, they weren't because they didn't concentrate on that. A lot of social networking centers on content-centric networking - discussing, arguing and collaborating on content. However, it doesn't necessarily mean that the social platform will own that content. Look at all the YouTube video that show up in social networking sites.

I do think that ECM vendors should concentrate on Content Collaboration - the act of working together to create and manage content, but that in and of itself is not Social Computing. It's about sharing ideas and content in the context of a common problem and it should be presented as such. This collaboration could benefit from more focused social computing platforms, such as instant messaging, on-line conferencing, activity feeds a la Facebook, walls, wikis, blogs and forums. By concentrating on providing content services for those platforms, we are creating a win-win situation. More than that, we are focusing on what we are good at - Content Services. These Social Computing or Networking platform usually have only the most rudimentary content services, which make for very complementary products.

SharePoint has a lot to do with why this question would even come up. For lack of anything else to use in the enterprise, people are trying to build these types of solutions. However, what was once labeled as collaboration in SharePoint is now Social Computing, but the reality of it is that it is generally still content collaboration. At least that is what most customers use it for. This may or may not become a stretch too far for SharePoint, but it is one that we are not going to try. I didn't always feel like this and felt that we had a big opportunity in Social Computing. But as we progress the company, we have more to offer integrating with others and sticking to our core. We can get more from others that do the same.

CMIS one year on - soon in public review

2009-09-29T22:24:54+01:00

Autumn marks the new school year, business year and CMIS year

It's that time again. Autumn and the new school year and we are about to take another major step with CMIS, the Content Management Interoperability Services proposed standard at OASIS, and we will be that much closer to having CMIS become an official standard. As of today, a majority of the technical committee has voted to put the CMIS specification to public review with zero no votes.

I was recently reminded that I wrote about the public introduction of CMIS around the same time last year. This was the time that IBM, Microsoft and EMC decided it was time to announce the Content Management Interoperability Services that they started October 2006 and release it to OASIS to become an official standard. At that time Alfresco was fortunate enough to have been involved in the process for the previous year before joining the original three, along with Open Text, BEA and SAP. For the year before that, IBM, Microsoft and EMC broke away from AIIM iECM committee to speed up a process that itself had been going since Autumn 2005. Much of the groundwork was laid with iECM and the parallel, but Java-focused JSR-283 committee. All in all, this will not be a fly-by-night standard, but a standard that has been carefully deliberated, considered and grounded in real world use cases.

This last year has seen the number of vendors participating increase substantially with plenty of beneficial results. We have had a few face to face meetings that have triggered some great collaboration in a very competitive industry. The number of eyes looking at the spec has increased, but so have the minds contributing new ideas and expertise. SAP presented real world requirements that needed security in the form of ACLs, so something that we had been willing to punt on before, we added in. Microsoft, IBM, SAP and other owners of search engines participated in a working group to define APIs to get changes to a repository to keep search indexing up to date. Also, Day and Nuxeo set up an Apache project to share Java API and SPI components along with Gabriele Columbro from Alfresco and our friends at SourceSense. David Caruana, Alfresco's chief architect, contributed interfaces to handle renditions and specifically thumbnail previews.

More importantly, committee members were building real life implementations of the specification at each stage to really test its implementability and completeness. As Ethan Gur-esh from Microsoft once said, "I relish the fact that CMIS is retrospective." It can work with existing repositories, not be overly prescriptive, but still describe a powerful content application platform. Regular plug-fests ensured that the systems interoperated with each other. This alone distinguishes it from all other content management standardization initiatives - real life integration matching the current state of the specification. Dave Caruana's Java test harness, now in Chemistry, provided a tool for other vendors to test the completeness of their server implementations.

Developers started creating new applications based upon the early specs, such as CMIS Spaces and CMIS Explorer, which tested its usability as a platform interface for content-centric applications. CMIS's initial focus on real-life use cases has so far paid off as the level of functionality seems to match new applications being built and existing applications integrate content management capability, such as Drupal, Joomla and Confluence have with Alfresco through CMIS. I certainly encouraged anyone who wanted to integrate with Alfresco to do so using CMIS.

Today, unless some catastrophic event occurs, CMIS will go into public review. This is the last step in OASIS before becoming an official standard. As the process is laid out, it could take a maximum of 6 months for completion of the step and CMIS becoming a standard. It also means that only typos and minor corrections will change in the spec, currently at version 0.7. The determination of the spec leads, David Choy from EMC as Chairman and Al Brown from IBM and Ethan Gur-esh from Microsoft as Editors, has ensured that we got to this point and they have done an excellent job marshalling the process along. Everyone on the committee seems determined to see this to as speedy a completion as possible. It seems likely that CMIS will become a standard in the first quarter of 2010. I can't see anything getting in the way of this other than the US government disapproving or Google, whichever is more powerful. All the rest of the major players backing the standard. (Craig Randall, formerly of EMC Documentum, just popped up at the last call representing Adobe.)

Being the eternal optimist, I have very high hopes for CMIS, but I believe I have reason to be optimistic. I have seen it before first hand with SQL. Without vendor lock-in, it will now be possible to build content-rich applications that can run against many different systems. Customers will be able to move content from repository to another. It will be much easier to find stuff in random repositories without having to build multiple connectors. As the breadth of applications and utilities becomes available against these repositories, usage and deployment of content repositories will become more widespread and grow a much bigger pie for all vendors. As Alan Pelz-Sharpe reports in CMS Watch, CMIS is already gaining traction in 15% of AIIM members before it has reached standard status.

I'm really looking forward to this new year!

Professional Open Source Software

2009-07-29T15:15:34+01:00

I am constantly asked about why we put our software out as open source. Is it some sort of altruism? Communism? People who have worked with proprietary platforms may not be familiar with open source and its implications or the advantages or disadvantages for the products they are using or for the users’ organizations.

Image under Creative Commons Attribution-Noncommercial-Share-Alike License

Much of the early development of what is now known as open source started with the GNU project by Richard Stallman in the early 1980s to create development tools that emulated Unix, but did not require a license from AT&T or any payment to anyone for that matter. Stallman labeled this “Free Software” and felt that the freedom to use this software was paramount. This early development culminated in the mid-1990s in the creation of a completely open and free Unix-based operating system called Linux by Linus Torvalds under the same license terms as GNU. Torvalds designed Linux to be open and to allow other hackers to contribute to it.

In 1999 at the inspiration of Eric Raymond and his book “The Cathedral and the Bazaar” and through the organization of Tim O’Reilly, the community of developers and hackers who had created this free and open software decided to call the movement “Open Source”. Raymond’s and others’ reasoned that “Open Source” would be clearer to the potential users of the software and less threatening to businesses. This change was also timed to coincide with the release of the Netscape browser and the creation of the Mozilla Foundation.

Not long after this, open source got a huge boost when IBM changed strategy of building proprietary Unix systems and decided to adopt open source and Linux as a means of competing against Sun and others in the development tools market. IBM started actively contributing to the Apache project, a collection of open source tools and the most wide used web server. By contributing to Apache, IBM did not have to reinvent tools for integrating with various web technologies and actually benefited from the development as others, just as others had benefited from the IBM’s contributions. Before long, other large vendors started actively contributing to open source.

This also set the stage for start-ups in the early part of the 2000s to create businesses based upon open source and to employ a professional open source business model. MySQL was a database system originally developed in the early 1990s and had used the GNU GPL license to distribute their software. When Martin Mickos became CEO in the 2001, he helped develop the dual license business model of distributing a GPL-version of the software and selling ordinary software licenses to those who wanted support or wanted to distribute their software without the GPL license. Shortly after this, Marc Fleury took his JBoss application server and licensed it under the more liberal LGPL or Lesser/Library GPL license. JBoss raised venture capital and started to acquire other Java-based projects. Both companies thrived in a down economy proving the value of distribution via the open source model and developing business through selling support and services.

Open source became a change or really a revolution in business model and means of production rather than anything technological. It is underpinned by three main principles:
Opening intellectual property and transparency spreads faster than concealing intellectual property and builds a bigger pie
Share and share alike of intellectual property will grow that intellectual property faster
We should all reuse other people's intellectual property and stop wasting time re-inventing the wheel

This change in business model led to a flurry of start-ups to use open source as their preferred distribution model, including Alfresco in 2005. The advantages of this approach became clear with the success of Linux, MySQL, JBoss and others. These include:

Broad distribution of product through internet download
The ability of users to try the product before buying, eliminating much of the sales and demand generation activities of ordinary enterprise software
Much broader testing and code review of products in the beta and product phase of development
Word of mouth marketing and open source meet-ups reducing the cost of marketing
Lower cost of development through use of other open source components and contributions from the many extensions that users create
Bug fixes from users end up fixing problems themselves

The benefits to the user and business community also became clear:

Users were not locked into a proprietary platform and they were free to use the open source version of the product
The vendors providing these products were the support source upon which they could rely if they needed fixes, advice, warranty or indemnity
The open source vendors were held more accountable than the proprietary vendors since there was always the possibility to switch back to the free version
Users could see the source code and either make the minor changes they wanted without waiting for the next release or even change wholesale parts of the product
Because of the lower cost economics and economies of scale of open source distribution, the lower cost of product was passed on to the consumers of services
The transparency of open source means that there is greater visibility of changes coming and what bugs already exist
Through the community and transparency aspects of open source, the users played a more intimate role in the development of the products with a greater influence on product direction

There are lots of open source programs and projects out there and many of them are terrible. However, open source works well when software is a commodity and you cannot tell one piece of software from another - e.g. word processors, web browsers, operating systems. When software is a commodity, you are paying only for a brand where the company is getting money for no changes in the software. Professional open source works well when enterprise software is commoditized. Alfresco was created because of the increasing difficulty in differentiating one enterprise content management system from another.

Open source development is painted as chaotic and unreliable by proprietary, but in reality it is generally just as controlled as any other project. There are project leads that make decisions and no more than a handful of people developing it. Sometimes these developers can come from more than one company. This is made easier because there are no secrets or intellectual property to hide. However, the project leads have the final decision on what goes in or out. There can, however, be many times more testing involved than ordinary software. In most successful open source projects, you can have hundreds of people downloading the software and testing it out and sometimes even providing fixes because they could see and fix the source code.

In order to make this process work, somebody must pay for it. Usually an open source company is in markets directly adjacent to the software, such as system integration. By supporting Linux, IBM makes its money by selling hardware that runs Linux and selling consulting services to build solutions for customers. IBM has a huge patent portfolio, but they make twice as much money selling hardware and services around open source than they make from licenses from their patents.

More recently, projects that had been purely community-driven are seeing the opportunity to create a professional open source company that is closely associated with the project. A couple of years ago, Dries Buytaert got funding for Acquia to create professional services around Drupal, the open source web content management project that he started in 2001. Recently, some of the core committers of the Lucene and Solr projects have created Lucid Imagination to provide support and enterprise features on top of those two open source projects. The demand was there in much the same way that previous professional open source project have seen for support, bug fixes and features. They also have fit into categories that are consolidating and commoditizing.

Alfresco Software makes money by selling technical support on the Alfresco system. Not everyone wants or needs support, but it isn’t necessary for everyone to want it. With over a million downloads, a small percentage of users who need support in a product environment, especially in the Global 2000, are sufficient to support the community, the project and the company. By participating in the $4 billion enterprise content management market, Alfresco has sufficient scale to continue growing and service the open source community and product.

Follow me on Twitter @johnnewton