I had such a good time at cmf2007 that I vowed to go back and was the first speaker to confirm for jboye08. A couple of days ago, the J.Boye team announced that Robert Caiilliau, co-inventor of the World Wide Web, will be a keynote speaker for the event. The anticipation builds...
Jeff Potts has an update to his cogent post on Alfresco's insularity. According to Jeff, John Newton and Kevin Cochrane promised to create a system that would allow non-Alfresco employees to contribute to the Alfresco code base. The changes are expected when Alfresco releases 3.0 of the Enterprise Edition. At this time, the Community Edition (which Alfresco treats as a "lab" for introducing new, unsupported features) may start to accept contributions from non-Alfrescans. The prospect of committer status is unsure.
What makes this a big deal is that Alfresco has always hobbled the Community Edition to drive sales to the Enterprise Edition. First, they stripped down the Community Edition. For a while they made it badgeware. Most recently, Alfresco's policy has been to forbid integration partners from helping Community Edition users. Because Alfresco doesn't fix the bugs in the Community Edition and Alfresco partners will not help you implement and maintain it, the Community Edition is too risky to use for most companies to consider.
Opening up Alfresco Community Edition would allow a community to form around the project that may make it more viable to use in the same way that companies use Fedora Core as a lower cost alternative to Red Hat Linux.
In the open source content management world, Magnolia is a good example of a commercial open source company that has non-employee committers on their Community Edition. The Magnolia Community Edition serves as the core on which the Enterprise Edition is built. In order to get support, you need to buy the Enterprise Edition whether or not you intend to use its value added features (Sitedesigner, clustering, etc.). Other commercial open source projects like Hippo and Daisy have one open source licensed product and make all their revenue on support and training contracts which are entirely optional for companies that adopt their software. However, Magnolia, Hippo and Daisy are small and not so heavily venture funded, so they have less pressure to convert adopters into paying customers.
While Microsoft Word is undeniably a poor content production tool for all the obvious reasons (poor separation of content and presentation, non standard formats, instability, etc.), it is amazing how much textual content starts out in Word. Even content management gurus and purists can't resist the automatic impulse to fire up Word before they start typing even for things like blogs (as an aside, I am loving MarsEdit as an offline blogging tool). They accept the effort that it takes to strip out all the MS Word cruft for the convenience of working in a familiar word processing tool. As much as I hate to admit it, Word is everywhere.
But there is another little text editor that everyone likes to use and that is an email client. Many of my colleagues just start typing in an email as the most convenient way to capture information. After they are done, they can decide whether to turn it into an email or paste it in another application to manage it as content. Email clients usually have basic word processing functionality like a spell checker and simple formatting.
Readers of this blog know that I have been very critical of email as a collaboration tool. My complaints of email center on email's failure as a repository (what you get is a lot of duplication but very little retention). But as a text editor, I actually like email because it is focused around a standard. Email editors, even the WYSIWYG HTML kind, try to limit the amount of non-standard formats that they produce because they have no idea what kind of reader will be accessing the message. Of course, many users undermine this benefit by writing Word documents and then emailing them.
Some CMS have tapped into email as an authoring tool and probably many more have been customized to extend the native functionality to incorporate email for content entry. The functionality is most commonly found in blogging tools. For example, WordPress has a commonly used blog-by-email configuration. One of the better out of the box implementations from a full blown WCM that I have seen is GOSS's iCM product which allows users to mail in their articles. The iCM processes the article based on who it is from and puts in a workflow to be prepared for publication. Some users can have the rights to publish directly from email (or an SMS). You don't see GOSS outside of the UK as they are a regional vendor specializing in government clients. The knowledge management product Intraspect (acquired by Vignette) had a nice feature where every folder in the repository had its own unique email address.
I would like to see this feature implemented more frequently. It would be especially useful for simple, article based content types where the subject field maps to the title, the "from" field is translated to a byline, body == body, and attachments could be treated like related images or files. Anyone else know of WCM products that do this particularly well? I know Drupal has a configuration to enable this.
Over the past year I have been wondering about Serena Software's plans for their Collage product. If you haven't seen Collage, it is a simple but useful WCM system that is relatively widely used in higher education. Serena has been steadily marginalizing the product for a while. It has gotten so bad that the only way I can find Collage on the Serena website is to do a search for it.
Well, the Duo Consulting blog has a scoop that Serena's plans to sell Collage. Content management historians know that Serena acquired Collage when it bought Merant software whose primarily line of business was configuration management (source control) software (PVCS). Serena may not have even wanted Collage but it came in the deal.
I don't know what the acquisition market is for a middle of the road WCM product (especially with today's credit crisis) but I have a feeling that some company will come along and pick it up. In the meantime, Serena will continue to do the bare minimum to keep the product alive.
Matt Raible has an interesting post about some framework evaluation work that he did for LinkedIn. During the project, Matt helped LinkedIn address the issues of framework proliferation and the maintenance of a proprietary framework. Spring MVC and Struts2 were considered but neither of them solved the problem of a long development/deployment cycle: it takes a lot of time to change and run code.
There is also nice mention of Rails vs. Grails. But the biggest news was buried in the middle of the post. Matt is joining LinkedIn as a full time employee.
A few days later, I was pulled into the CTO's office and he offered me the job. He offered me the challenge of building this team and told me I could do it remotely (from Denver) and hire my own people to help me with it. I gulped as I realized I'd just been offered the opportunity of a lifetime.
Matt is going to head up the UI Frameworks team and he is hiring.
I am extending my trip to Europe with a stop-over in Gent for a fireside conversation about open source web content management hosted by Outerthought. Save the date: May 8th, 2008.
The format is going to be informal, conversational, and (luckily for me) in English. I will go over the key findings of my report Open Source Web Content Management in Java and there will be plenty of time for questions and discussion about all sorts of content management topics.
The event is free but space is limited so you need to request an invitation. I am really looking forward to visiting Gent and meeting the content management community there. Hopefully I will see you.
I was just reading Jeff Pott's Alfresco article "Working with Custom Content Types" and it is one of the best developer guides on Alfresco that I have read. The article goes through the creation of content types in a step by step narrative with some pointers to other interesting facts along the way. Jeff is hosting a whole series of these articles on his website. The Web Scripts introduction is also very good.
For a vendor neutral analysis of Alfresco's general utility in web content management, I am offering my Alfresco evaluation from Open Source Web Content Management in Java as a stand-alone report. The report covers Enterprise Edition version 2.2, which is due out any day now.
[Update: here are the slides and the recorded presentation in case you couldn't be there]
I am back home in Florence after a couple days at Drupalcon and AIIM. It was great to see all my colleagues at both events. I didn't attend any of the AIIM sessions but I did catch a few of the Drupalcon talks. Here are some quick notes and observations.
Drupalcon was huge. The attendance doubled from 400 last year in Barcelona to over 800 this year and there were people turned away because the venue was at capacity. AIIM was also pretty well attended - at least on the Expo floor.
I arrived a little late and squeaked into the session "Report from formal Drupal usability testing at the University of Minnesota Libraries," which reported results of non-Drupal users trying to do 7 tasks in Drupal. The study looked very professional and well executed. Unfortunately I missed Task 1 which most of the users got hung up on. For more information, you can go to the Drupal Usability Group where there is already a nice discussion and the team plans to post the slides, video and other findings.
At a high level, I think it was great that the Drupal community was able to do this. No one was at all defensive and everyone seemed excited to fixed the problems uncovered. They welcomed external criticism and that is a good characteristic of an open source project.
Some of the specific points from my notes:
Newbies are confused by the stuff that the pros just ignore.
Help was not helpful.
Even after tweaking of the terminology (less references to the word "node"), users are still confused.
There needs to be a good glossary. There used to be one but it was removed
There are too many options and that buries the important stuff.
There should be tutorials.
A common complaint was that users lost the page that they just created.
Dries did a good keynote on the state of Drupal (that is strong, by the way). Some specific notes....
Drupal 6 came out with February 16th, like previous Drupal releases, it was big. The number of downloads over the first month doubled the first month (100,000 vs 50,000 for first month of Drupal 5). There were 741 contributors (Drupal 5 had 472). And several thousand more fix/enhancement patches than Drupal 5. There are now 20,000 live Drupal 6 instances pinging home.
Looking forward, the Drupal community wants to improve the project to make it ready for the next wave of adopters. This includes addressing pain points like module compatibility and expanding Drupal's relevance. One nice quote was "the Social Graph just connects people. we have the opp to build a graph that connects everything." Dries went into a discussion of the semantic web and RDF and how Drupal could be a major player. Not a lot of people can get away with promising about the semantic web. I guess Dries is one of those people. Here are some ideas:
Web 1.0 = WCM
Web 2.0 = Web 1.0 + user management + infinite extensibility
Web 3.0 = Web 2.0 + infinite interoperability
RDF is the way to achieve infinite interoperability and federated data sources. In my opinion, Drupal is very weak in this area and in things like web services and REST style interfaces. Drupal developers are very database oriented and tend to use MySQL replication to move data around rather than application to application APIs.
Getting into specific features, Drupal faces the same dilemma that Plone and every other content management system face: is it a framework for developers or is it an application for business users? Dries took a mathematical strategy to solving the problem. A survey of the Drupal community found that 32% wanted to focus on end users of small to medium sites, 44% wanted to focus on enterprise users, and 24% wanted to focus on developers. Out of that he concluded that Drupal 7 should have 70% user facing features and 30% developer oriented improvements. It is a simple solution but I don't know if I buy it. In particular, this approach makes it hard to do some of the major refactoring that projects sometimes have to do in order to prepare for a new set of features or reach a new market.
Dries also predicted that Drupal 7 would be out October 15th 2008. In the past the community had been very reluctant to put out target dates. It will be nice to see them make this date that they set. Having Acquia pushing things along and dedicating developers like they did with Drupal 6 will help.
More robust testing practices were also mentioned. There was talk about getting a testing framework into the core and establishing testing procedures. This will help improve the quality of the code and make it easier for Drupal to get into more demanding IT organizations. Better testing procedures will also make it easier for the contributor community to safely grow.
The Knight Foundation is giving away a lot of money through the News Challenge to users and developers that are using Drupal to transform community news. Lisa Williams was there to talk about PlaceBlogger.com and Benjamin Melançon talked about his "related" module. For more information on what newspapes are doing on Drupal, see Newspapers on Drupal.
Configuration management (as in managing multiple environments and promoting code forward from development to production and pushing content back from production to QA) in Drupal has always been a pet peeve of mine. To put it simply, there is no good or standard way to do it. The core issue is that Drupal stores a lot of configuration in the database and there is no good way to move a configuration from one instance to another. A second component of the problem is that IDs are overly important in Drupal because most developers use IDs in their application logic but these IDs are randomly assigned by database sequences. I think these configuration management issues will keep Drupal from being adopted by large development teams building huge sites. I sat through a session on this topic and a Birds of a Feather discussion only to learn that everyone is doing some hack or another.
The hacks break down into two categories: use a macro tool like the Devel module or a third party application to record and replay the clicking of options in the admin forms; use ranges of IDs for settings that are owned by different environments. The latter is the preferred
method by advanced developers. They do things like start the sequence of dev at 1 and the sequence of production at 5,000 so there are no collisions. Or they use odd numbers for development and even numbers for production. Neither approach is straight-foward or scalable. Developers talked about all these database scripts and diff'ing processes that they have created. I would really like to see some fundamental improvements in this area like using the file system to store these settings and best practices in coding to avoid using IDs.
[Update: Jean-Baptiste Ingold was kind enough to record the BoF session and post it on Blip]
I attended the Acquia panel at Drupalcon in Boston yesterday. I must admit that I thought this session was going to be packed with people interested in what Acquia is going to do and how it will relate to Drupal. Plus, Dries always draws a crowd. It turned out that the session on theming drew a much bigger crowd. In hindsight, I am not surprised. I was sitting in the Drupal IRC channel when the Acquia announcement first hit. There was a mention of the news but neither concern nor intrigue about this new addition to the Drupal family. People were more focused with the their immediate tasks working with Drupal. They trust Dries and the Drupal Association to do what is right for Drupal.
The session started out with an introduction to the Acquia team. The first thing of note was the early hire of Gábor Hojtsy whose primary responsibility was to get Drupal 6 released. I think that this demonstrates Acquia's commitment to the Drupal community and how they can help themselves by helping Drupal. Another significant hire was Jeff Whatcott from Adobe for marketing. Their community evangelist Kieran Lal was walking around in colonial period attire to signify being a founder although most mistook him for a pirate. He should re-use the outfit on September 19th.
Acquia's business model hasn't changed since the original announcement. They want to be the Red Hat or MySQL of Drupal and they want Drupal to be the web application framework for Web 2+. In the near term they are going to allow the community to shape their activity by taking a play from the Zimbra book. The Acquia Projects page will allow the community to suggest and vote for enhancements and services that they would like Acquia to provide. Paid technical support is another core service that Acquia will sell.
A little furtherdown the road (2nd half of 2008), Acquia's strategy focuses around two offerings: Carbon (a supported Drupal distribution with certified modules) and Spokes (a collection of network based services like safe automated updates and configuration control). There will be more value added software in the future. The product names were inspired by CEO and co-founder Jay Batson's (don't bother following the link. Jay has been too busy to blog. I know the feeling.) interest in cycling.
One of the big contributions that Acquia hopes to provide is rigor around testing. During Dries' keynote, he spoke about building in testing frameworks (unit, system, and integration testing) and it seems like Acquia has the experience and funding to make this contribution. This strikes me as similar to SpikeSource who also tried to create a commercially supported Drupal distribution. I was going to ask what made SpikeSource's SpikeIgnited Drupal fail but the answer is obvious. They didn't have Dries. In fact, they didn't have any relationship with the Drupal community and the resulting product was pretty lame: Drupal without the modules.
Another part of Acquia's strategy is establishing a network of certified partners that can work with Acquia to build and serve the market for Drupal. They are going to do this through partnership and certification programs. The latter will be run by trainer and Drupal book author Robert Douglass.
As for Dries, the deal is that he gets a minimum of two days per week to spend on anything Drupal related be it coding, community development, or being a spokesman for the product. He reiterated the point that he is able to make decisions that are better for Drupal than for Acquia. Hopefully these two interests will not be in too much conflict. I have not seen the agreement that was signed (nor do I expect to) so I don't know the wording of Dries's terms. My major concern is that VCs (which gave $7mm to Acquia) like to have control.
This brings me to a bigger question about VCs: what did they buy? It seems that much of the VC pitch centers around the growth and promise of the Drupal project. VC's would love the fact that every time a new release of Drupal comes out, the number of downloads doubles. But I hope that they understood that they did not buy Drupal itself.
As long as Dries is on board and Dries keeps his promise to the Drupal community, Acquia's success may mirror Drupal's. However, Acquia will probably not be able to monetize all of the Drupal phenomenon. Figuring out a way to capture the upside of Drupal's success without having a negative relationship with Drupal will be Jay's job. A commercial entity has the promise of helping the Drupal project by bringing in more interest and adoption for Drupal and also protecting the Drupal purists from the open source n00bs that are still running around looking for throats to choke. Acquia's services will probably appeal to medium to larger sized companies that have money to spend on software. Not the small, budget-less non-profits that bring a lot of the energy and passion into the community. If Acquia can target new markets with their products and serve existing markets with their Drupal contributions, I think they can achieve a nice balance.
CMSWire has a nice little article on Google Sites and MOSS. The general gist is that, while not as capable as MOSS, Google Sites may be just good enough for basic websites and collaboration. The absence of some advanced features may be compensated for by the simplicity of the platform. Sometimes just good enough is exactly what you want.
Over on the Sharepoint side, I just read Michael Sampson's excellent report The 7 Pillars of IT-Enabled Team Productivity: The Microsoft Shareport 2007 Analysis. The report uses Michael's 7 Pillars framework to evaluate Sharepoint. Not to be a spoiler but MOSS doesn't do so well.
I am not convinced that you want one platform to try do do everything but Michael assumes that this is Microsoft's goal (it certainly isn't building a robust WCM system). So, it seems that Sharepoint is caught in the middle. It is not the silver bullet, one-stop-shop for enterprise collaboration, and it is not the simple, cuddly tool that everyone loves to use. That is not a very fun place to be.
