<!-- Content Here -->

Daniel Jacobson on de-coupled publishing systems

Daniel Jacobson, NPR's Director of Application Development, has an excellent article on the philosophy of de-coupling the content management tier from the delivery tier. He calls this strategy COPE: Create Once Publish Everywhere. In particular, the diagram is particularly useful in showing how all the pieces fit together.

If you are in the content publishing business (like NPR is), this is absolutely how you need to think about your content technology stack. Your content repository, editorial systems, and distribution channels can get sophisticated and highly specialized so compromise and lock-in can be costly. The delivery tier that came with your content management system may not scale or may not allow you to push the cutting edge if an opportunity to innovate should arise. All of my big name publishing clients have adopted this strategy for their core publishing platform.

However, as I have warned in earlier posts, the flexibility may not be worth the cost for all publishers. Unless your business model depends on aggressively leveraging your content and you can afford to play on the cutting edge, a lighter weight "website in a box" style architecture may give you the flexibility you need without the additional complexity and cost of building and integrating these de-coupled systems. As an example, Drupal is rapidly becoming a popular platform for small to medium publishers and also for smaller initiatives in larger publishers. And you cannot get a tighter coupling of management and delivery tier than Drupal. One strategy that has been used by early Drupal adopters (that have grown out of their forked versions of the platforms) is to use Drupal to publish into a custom delivery tier.

As an architect, I love the COPE model and I think that most successful, large scale content businesses will eventually converge on that strategy. However, as a pragmatist who also serves publishers in the lower and middle tiers of the industry, I know that the resources and expertise may not be available to go straight to that architecture. Still, at the very least, every publisher needs to start thinking on this level: creating and storing content in presentation neutral way to keep options open.

Attending Django Master Class

I am looking forward to attending a Django Master Class, taught by Jacob Kaplan-Moss, on October 16th in Springfield, Virginia. I have been building a prototype application for a client in Django over the last couple of months and have been very impressed with the framework. It looks like the application will be going to Beta soon so it will be great to pick up some expert tricks that will take me to the next level.

Another flower war

Not quite the War of the Roses, there is a little skirmish happening between Magnolia International (as in the CMS company) and Ma.gnolia (the off-again, on-again social bookmarking site). The squabble began shortly after Ma.noglia's recent resurrection (it had shut down in February 2009) when Magnolia International issued what appears to be a Cease and Desist order. It is a little unclear because it was written by Magnolia's CEO, Pascal Mangold (not an attorney) and it seems to be more of a request than a demand.

I am no trademark attorney but, since that hasn't stopped anyone else, I figured I would weigh in from a common sense perspective. I understand how Magnolia International would want ma.gnolia.com to be out of the picture. They probably also want ExxonMobile to stop using magnolia.com. ExxonMobile probably wasn't paying attention when, then Obinary International, Magnolia registered magnolia.info to market their new open source CMS.

Does Ma.gnolia have the right to use their domain? I don't know. There are very specific legal distinctions to determine legal and illegal trademark infringement but I don't think it really matters. In the real world, it happens all the time. This is why you try to grab all the variations of your domain (.net, .org, and .biz as well as -sucks.com, -sucks.net, and -sucks.biz) when you are branding your company. You do your best to claim your little spot on the web and you work with what you've got. Getting lawyers involved is an expensive distraction.

I am sure that Pascal knows all this and that is why he wrote his own letter. Hey, it was worth a try. Maybe this will lead to a negotiation for Magnolia to purchase gnolia.com. Who knows how serious Ma.gnolia is about their business. Personally, I don't think I would use a bookmarking service that had a track record of closing down and opening up again.

This is not a particularly interesting story. What would be more interesting is if Magnolia International went after ExxonMobile. What the heck is ExxonMobile doing with that domain anyway?

Feature Request: Audience-Inspired Keywords

In order to maintain my vendor neutrality, I refrain from advising software companies about their products. But that doesn't stop me from having ideas and opinions so I am starting a series of features that I would like to see. Please feel free to use them and, if you do, tell me about it! If you don't like an idea, let me know why.

Today's feature request can probably be implemented (at least primitively) as a minor customization on most content management systems.

Audience-Inspired Keywords

When a visitor searches for content on the site (or arrives at the site from an external search result), the presentation tier records the keywords that he used. These keywords are then listed as options that content contributors can use to tag their content. This would allow contributors to tag content in the language of their audience. To keep things manageable, there should be a mechanism for filtering out typos and organizing the keywords. There may be some system that would allow you to group synonyms so if people search equally for "bike" and "bicycle," the contributor only has to use one of those terms and the asset is returned on both "bicycle" and "bike" searches.

The original seed for this idea occurred way back when Arjé Cahn, from Hippo, told me of an experimental feature that used the terms a contributor used to search for related images and links as keyword suggestions for the asset he was working on. This feature idea takes it one step further by using the language of the audience, rather than the contributor. Another part of the inspiration comes from the "Best Bets" strategy where you train a search engine to recommend a certain set of assets when a specific keyword is used.

Paul Graham on Post Medium Publishing

Paul Graham, has an interesting essay on Post Medium Publishing. The general gist of is that publishers have never really been in the "content business." They have been in the business of selling paper (the embodiment of the content). Digital publishing, with its ease of replication, devalues paper as the embodiment of content and exposes the weakness of a pure content business. Publishers can either give content away and look for ad revenues or they can come up with some new embodiment that is both desirable and defensible. Personally, I think that publishers are in the audience business. They command an audience that advertisers are willing to reach. But the audience business has become more difficult with the increased competition from all sorts of information sources. Many people, like Graham, are looking for that new embodiment or channel that will bring back the exclusive audience attention that was so profitable back in the golden years of publishing.

Graham also invalidates some of the successful content market examples publishers are looking to emulate. For example, Graham says that iTunes (held as the gold standard for a successful content business) is not really a content store. Rather, it is a "toll-booth" that charges people to access the most direct path to their listening devices. Apple controls the path, it gets to charge a fee. If you don't control the device or the channel, you are back in the unprofitable content business. It appears that Time, Inc. has a strategy here but who knows if they will be able to create that market.

New Alfresco Review

I am pleased to announce an updated version of my Open Source Web Content Management in Alfresco report. The report evaluates Alfresco Enterprise 3.1's WCM capabilities for both traditional web publishing and as a framework for building dynamic web applications. Like all Content Here reports, Open Source Web Content Management in Alfresco is highly technical and gets into details that a potential buyer should know. In writing the update, I interviewed systems integrators and technology managers from customer companies for their candid opinions of the product and the software vendor. I have also personally evaluated Alfresco, supporting documentation, and third party books. I can safely say that you are not going to get a more thorough and unbiased evaluation of Alfresco anywhere — not even if you pay several times the $200 price.

Long time readers know that Open Source Web Content Management Alfresco was originally published in February 2008 as part of a larger report called Open Source Web Content Management in Java. Because all of the products reviewed in that report have undergone significant upgrades, I have been selling it at a deep discount. The front matter that explains the marketplace and significant portions of the evaluations are still accurate and relevant so I have decided to offer a bundled product consisting of the original report plus the updated Alfresco review for $400 — that is still 50% off of the original list price. As I complete updates to the different reviews, I will add them to the bundle and incrementally raise the price to the original full price.

If you are evaluating Alfresco for web content management, save yourself time and reduce your risk buy purchasing Open Source Web Content Management in Alfresco. If you work for a Java shop and are starting to consider open source alternatives to commercially licensed web content management software, consider the Open Source Web Content Management in Java bundle.

Snow Leopard Issues

For some reason, I upgraded to Snow Leopard at my first opportunity. To tell the truth, the end-user improvements are pretty modest and it is difficult to perceive Snow Leopard's slight speed improvements. That is not to say that the upgrade did not have impact. Some software that I regularly use stopped working. I am all up and working now and here is now.

Looking back, it would have made more sense to wait on the upgrade and see what other people on the mailing lists said about their experiences. But, since I figure I would share what I learned.

Ingredients for a good software development project

While nearly all of my consulting work is in software selection and technical strategy, I try to have at least one implementation project in the works so that I can stay relevant as a technologist. For my current project, I am building the technology infrastructure for a new web based startup. The project is going really well and I thought I would share some reasons why.

• Great Client. My client is totally non technical and this is his first startup. He found me through a mutual friend that we both think very highly of. His lack of technology experience means he has not picked up any of the bad habits of technology owners. He didn't force a detailed Gantt chart with a bunch of imaginary dates and tasks. His trust in me helped us to collaboratively prioritize and plan work around his schedule. I have followed through on all of my commitments so the trust has only gotten stronger.

• Fluid Design. The concept has evolved substantially since its original inception. A critical part of the design process involves going through the application with prospective customers and then talking through how to incorporate their feedback. We have done a really good job of keeping things flexible without adding complexity. Some of the more speculative ideas are faked a bit until we get an indication of their viability. The underlying machinery is only reworked when we are really onto something. We have plans to do a holistic refactoring once the design is frozen and we are ready to start preparing the application for beta customers.

• Strong Frameworks. Building the application on top of the Django web application framework makes development incredibly fast. Python has a great unit test system where the tests are written in the comments of the code. This makes them easy to maintain. All of the data administration interfaces are provided for free and, because everything is data driven, the client has a lot of control over what he shows potential investors.

• Simple Tools. All of the work is planned using tickets and milestones in a service called Unfuddle (looks a lot like Trac). The information is very accessible and email notifications keep everyone up to date. I am probably going to bring another developer on board to help with the beta and I think that setting him or her up with a development environment is going to be really easy. Just pull everything down from Subversion and launch the text editor of choice.

I am not fooling myself into thinking all projects can be like this. You don't always get to choose your clients or your tools. Often other constraints are forced upon you. But you can make some adjustments to achieve incremental improvements. For example, on another project for a very large, established client, I used to say "we are going over a waterfall in an agile barrel" meaning that the upper management saw a waterfall model project plan but internally, we were being as agile as possible. We tried to be vague rather than guess and commit. We made decisions that didn't preclude other options. We made opportunities to refactor code. So, rather than throwing up your hands and accepting a recipe for failure, think of how you can get the ingredients for success.

Jeff Potts explains the difference between Alfresco's DM and AVM Repositories

Jeff Potts has written an excellent article explaining the functional differences between Alfresco's two different repositories (the original DM and the WCM). The enclosed chart is a useful cheat sheet that belongs on any Alfresco developers cubicle. Nice work Jeff!

For those of you who are new to Alfresco, the WCM repository (or AVM: Advanced Versioning Model) was introduced to the platform when Alfresco added web content management. The AVM is designed for managing structured XML content and does very cool things like snapshots and contributor "sandboxes." But the AVM is slower than the original DM repository and has less powerful metadata capabilities. Typically, web content will be managed as XML in the AVM and then pushed over to a cluster of DM repositories to power a dynamic website.

The Drupal Divide

Recently there has been a lot of chatter about friction between the designers and developers within the Drupal community (see posts by Earl Miles and Greg Harvey). The Drupal community is huge and dynamic and is bound to have various skirmishes as members with different backgrounds and needs try to get what they want out of the platform. But I didn't really expect that the divide would be along the lines of developers and designers. I thought it would be between high-end adopters (like media companies) and low-end adopters (like individuals, small companies, and non-profits that make up the bulk of the Drupal community). Or it could have been between people who see Drupal as a business application and those who see it as a development framework.

The Drupal community is keenly aware of the importance of usability and has done a lot of work to identify issues. With the upcoming version 7 release (code freeze any day now), Drupal upped their usability efforts by bringing in some high profile designers to participate. The transition from being developer-dominated to having a more even balance between design and development is important for the Drupal platform as a business application. Mark Boulton and Leisa Reichelt conducted a very open process to incorporate peoples design ideas and feedback. Now that I think of it, this rift between new ideas and the old Drupal guard was pretty inevitable. This kind of transition is not easy. Both sides need to adapt. I don't think there will ever be total harmony between pure developers and pure designers but I do think the project can achieve a healthy level of tension that improves the software.