The Pros and Cons of Distributions

Posted on by | 15 Replies

It’s been far too long since I blogged.

I’ve got a few good excuses such as a baby, a business, a book, the OSM board and almost weekly trips around the U.S. However, I’ve also got a few good topics stored up, so expect more posts in the coming weeks.

First up, some thoughts on Molajo which is the shaping up to be the first widely-marketed Joomla distribution. Joomla’s had few distros before such as Non-Profit Soapbox and the pre-packaged installs put out by template clubs and extensions like Virtuemart. However, those were never really presented as distributions and that’s what the Molajo folk are aiming for.  They’re expressly aiming to replicate Drupal’s success with distros. I’ve spend about 50% of my time this year in the Drupal world, so I’ve a few thoughts on their pros and cons:

The Pros of Distributions

Innovation and Audience

The huge benefit of distributions is that you can have innovation and an audience. Sadly, those two often are mutually exclusive.

Many innovative projects get lost in the crowd. Check out a site like http://www.cmsmatrix.org and you’ll see hundreds of projects with thousands of interesting ideas that never get any traction. CMSs are like restaurants – 90% of them fail without leaving a mark.

On the flip-side, the core of CMSs often have limitations on how much they can innovate because so many business, add-ons and side-projects depend on them being relatively stable.

Distributions, particularly done under the umbrella of the big 3 CMSs, seem to hit the sweet of spot of allowing innovation while still being able to reach a mass market.

Projects like Open Atrium in Drupal and Molajo in Joomla have the best  of both worlds because they immediately have a large audience share plus the freedom to innovate.

Personally, I’d love to see Molajo capitalize on those advantages and carry through on it’s plan to innovate in terms of Joomla’s usability.

Less Confusion

Because Drupal is further down this track, I’ve really had the chance to discover distributions through the eyes of students.

The basic concepts of Drupal such as nodes, user levels, themes, blocks are pretty easy to grasp for most people.

The two places they really get hung-up with in Drupal 6 are the admin area (the overwhelming number of configuration options) and above all: modules. Drupal is a set of building blocks, but it’s pretty intimidating to realise that even something as simple as a photo gallery needs to be set up using several modules, including Views, which is probably the most complex, important add-on in all the big 3 CMS. Not only do beginners need to research these modules and learn how to put them together, they need to do so without feedback or reviews on http://drupal.org/project/modules. It’s tough.

So having a pre-packaged distribution where experts have already chosen and combined the best modules makes a lot of sense.

Makes More Sense to Decision Makers

From the people I’ve spoken with, having distributions directly targeted to performance (Pressflow), Intranets (Open Atrium) and universities (Open Scholar) is a big selling point. Rather than needing to take a general concept such as Drupal and sell them on the potential, they say it’s much easier to present a product which is already configured for their needs.

They can sell the bacon, not just the sizzle.

The Cons of Distributions

More Confusion

I did just mention “less confusion” above, right? Well, it depends on the situation. I’ve come across people who have been through a long evaluation process of choosing Drupal and then they find themselves immediately confronted with another question … which Drupal? Several students mentioned that finding and deciding on the right distribution added another level of complexity to their project.

Less Documentation

We’ve already talked about the pros and cons of admin themes several times on this blog. The major downside is that most of the documentation, videos and books out there immediately become useless. Add in a whole new set of modules and you really are in a situation where you’re left to rely on the documentation of that distribution.

Documentation is hit-and-miss at the best of times for Open Source projects but for the main versions of Drupal, Joomla and WordPress, there’s certainly no shortage. People are so keen to get Drupal 7 books at the door that one publisher had planned to release there’s in April 2010 and several more out already, even though the project just hit beta.

Read More

Over to You

What do you think? What are the advantages and disadvantages of distributions? Would you use them for your clients?

 

  • Oliver Ratzesberger

    Great article. I have seen many distributions over the years. The vast majority fails very early but there are very successful ones like SuSe (for Linux) and others have certainly contributed much to the success of Linux and other OpenSource initiatives.

    So I lookup the site and here is something weird about it:
    It is strange to see that this distributions website is not built on Joomla nor the group that stands behind it. How can a distribution advertise itself to foster the development of Joomla but run that community on a ning powered site?
    I find discussions of GPL/Open Source related issues on a website powered by a non GPL commercial platform.

    Certainly takes away most of its credibility from the get go. I would have loved to find an advanced Joomla site that showcases all the things that are being talked about.

    But then again, as you said 90% of project like this fail.

  • Scott Cooper

    Great post! A few of my thoughts…

    I don’t know much a/b NP soapbox but it seems to me it’s more of a “Joomla As A Service” offering than a distro.

    I agree 100% re: innovation. I think it’s one of the things holding Joomla back in a lot of market spaces, particularly when it comes to innovation that enhances the UX.

    It’d d be nice to see the innovation that comes from Molajo and others trickle into the core. I realize that much of what may come from distros will not be suitable for the core but some of it certainly will. I give Kudos to the Molajo crew as it appears they are expressly interested making this happen. If it is done well and under the “Joomla! community” umbrella, I think the confusion you discuss will be minimized. All too often, distros end up doing too much of their own thing and becoming a community in and of themselves…not necessarily a bad thing but it does have cons.

    I’d love to see a distro geared towards addressing the needs of the US Gov’t. I think it could contribute a lot to access control, session management, and other things they’re particular a/b. Many of these things could be worked into the core and a huge market exists for such a solution.

  • JoeJoomla

    Correct me if I am wrong but isn’t a distribution simply a package put together using specialized extensions and components but doesn’t change the core code?

    When the core code is changed can it still be simply a distribution? If Molajo becomes successful and continues to make changes to the Joomla! core at what point is it no longer Joomla?

  • Steve Burge

    Hi Joe, Scott and Oliver

    Interesting points … it’s difficult of course to keep a unified set of rules and beliefs as a community grows. I know even official some Joomla communities such as Italy have a very different set of ideas on the GPL compared to those on Joomla.org.

    Same for being Joomla or non-Joomla. Visit http://openatrium.com and http://openscholar.harvard.edu/home and you’ll see almost zero mention of Drupal.

    Both issues could either fit under Pros or Cons depending on your beliefs.

  • Amy Stephen

    Steve – Thanks for framing this discussion so perfectly. One of the allures of distributions is shortening that time to market for site builders. Those in the Drupal community who hope to build an intranet, for example, are fortunate to find Open Atrium. I believe we need those types of solutions for Joomla! users, as well.

    Scott – we agree that sharing innovation back to core is essential. You’ll find in our [url=http://groups.google.com/group/all-together-as-a-whole]email list[/url] that we share code back and intend to continue doing so. The problem is, not everyone’s code can be accepted in core and the project has the right (and the responsibility, in fact) to decide which code is accepted or not.

    Which leads to Joe’s very good question: “What is a fork?” It’s a question I have spent a great deal of time researching before deciding a distribution would be a good idea. I’ve got five years in the Joomla! project of good service and I have no interest in harming the brand I helped build.

    When we come to a situation where a patch is not accepted, and we determine that we need it for the distribution, we go through a series of steps.

    If we can create an Extension, we do so, problem solved. If not, we try to override a core class in a helper file. If that won’t work, then we could include a library, much like the JXtended library, to augment what core does not have available.

    We would be extremely hesitant to hack core code. Having said that, if we must do it there are ways to hack the code that are perfectly acceptable. Joomla! has a forum for developers to share core hacks and instructions for how to share Extensions that hack core. One must manage their hacks with upgrades to ensure we all use the same common code base and can install and uninstall hacks without harming the core base. We will obviously use that approach if the need comes.

    To this point, we have *not* had to hack core and have found one of the other three methods sufficient. We share Joe’s concerns that we not deviate from core, but rather augment what’s available.

    This is a great discussion – much appreciated.

  • Mark Keller

    Well, I’m not a fan of distributions. There is the risk that the distribution begins to “grow” to an fork. And that would harm the Joomla! project.

    But I’m also not happy with the Joomla! project in the moment. To many problems with the JSST (not responding on security warnings) and other Joomla! teams not accepting good patches from perfect 3rd party developers.

    So I can understand the way you go to have the possibility to bring your patches in the project. But this could lead to a fork.

  • Paulo Pereira

    Actually for language purposes, distributions of Joomla are already implemented. They don’t require (my language at least doesn’t use it) any code hack but some install files are already changed (has there is hard coded words). Same goes for sample sql and some plugins in the main package. But if we go for distributions for specific markets, then language issues would create more work for some languages.
    I see distributions interesting only for new sites, not for keep them updated. Also some developers would find an easy way to pack all of their own extensions (even if no one really went this path (except Virtuemart)

    Questions on my mind:
    1- It would be important to know if the amount of work to have the specific distributions up to date (core joomla and 3rd party extensions) would be less than actual plan
    2- How will they be updated (as many extensions doesn’t allow update without uninstall)
    3- Do you expect more then 2 or 3 distributions (with more than one 3rd party developer extensions packed inside) will survive after 1 year to be deployed? Or they will move to forks?
    4- Why the (many times requested) distribution model was never a solution implemented for Joomla?
    Note: Joomla its a fork and will be forked also. Smaller communities will always be a temptation to experts, the problem will be how it will manage big communities with newbies.

  • Amy Stephen

    Mark Keller –

    I used to also feel that distributions can lead to increased risk of forking. But, I personally believe it will actually do the opposite because we will learn to deal with some of the weaknesses you have identified and figure out better ways to collaborate as a community of developers.

    Taking on the core 1.6 code would be no small task. I helped start the bug squad. I am fully aware of the workload involved. We want to leverage and augment those community contributions. The pressure is on collaborating. No one has time to double their work.

    Good comments and concerns.

  • Amy Stephen

    Paulo –

    I answered a set of your questions at ATAAW, too, but these are a little different so, I’ll answer those here, as well.

    Since we are using Joomla! 1.6 core, we are also using those language files.

    1. Our plans for keeping Molajo updated is to take advantage of the cool update ability Sam Moffatt built into 1.6. The distro will be updated via a single link – whether it’s the core code, or the add ons.

    2. As I said in the previous response, Molajo updates will be one click, whether it’s core code, or the new template framework. No uninstall required. (Joomla! Extensions should use the ‘upgrade’ method to avoid this.)

    3. Why would we fork and take on that work? There is no reason to fork.

    4. On why we have never had vertical market distributions (I think that’s what you are asking) – we actually have had some. However, I’ve always felt like as a collaborating community, we are a couple of years behind Drupal. Two years ago, Drupal started down the distro road. It was met with some skepticism, too, but now, it makes sense to folks. I do not believe any of those distros resulted in a fork either. But, site builders sure do find they get to market a whole lot quicker. That’s the goal – strengthen our ecosystem.

    Thanks!

  • Bruce Letterle

    I always thought that joomlapackages.com was a great concept. I just never used them because they didn’t seem up to date on the best extensions.

    I don’t see anything in Molajo that would interest me. It seems likely to be a fork. I like industry application specific distributions, not just someone’s idea of what a Joomla core distribution should be.

    I’m a big fan of Acquia Drupal, Open Publish, and Drupal Commons distributions for getting those types of sites started. I just wish you could easily combine them, as we sometimes need aspects of each on a project. I have never seen anything like these in Joomla.

  • Klas

    Bruce, not everything is for everyone. It is not just someones idea, it is a group of people behind this. And idea is to make a distribution that we see as a better fit to our needs as sitebuilders, webdesigners and coders. And yes in that sense it is our idea of what core should be. We are starting with unified distribution but later on it will split into multiple, more narrowly targeted and more application specific.

    Fear of the fork is like fear of the dark, it is gone when you turn on the light. I hope we will light some in the near future. It is only the begging – just watch us move :)

  • John Coonen

    Steve, Excellent article.

    As you know, I’ve been a big fan for an enterprise distribution of Joomla! in addition to a community version.They support each other. My hope is, J/OSM can muster up the resources to create one under the Joomla umbrella.

    J/OSM has great potential to re-gain momentum this year. I’m very glad you’re on the team, and we’ve seen a great deal of renewed excitement.

    Forks are inevitable for any great project. The key for the original project is to stay focused on their own mission, vision, values and productivity along that path.

    An Enterprise version, along with a Community Version would be stellar for Joomla!

  • ssnobben

    First congrats to the baby Steve! Maybe first then it will be changing times! lol Father of two kids myself I know the problems and fun.

    Great article as always and hope hear more from you and also the ideas that comes up from Joomla Molajo!

  • Brian Peat

    My worry is that I tend to use theme maker distros (such as the Rocket Theme Rocket Launchers) to start my site, then customize the heck out of the template and add a few needed components. If RT doesn’t offer BOTH a regular and a Molajo Rocket Launcher, it means I have to choose which one to start with…and I’d likely choose the RT installer.

    I suggest finding a way to make an optional installer (in addition to the main distro) that would install all the Molajo parts without screwing up the front end or the installed template. I have no idea how hard that would be or if it would even be possible…but I suspect at the least, guys like the RT team could put a switch in their templates that says Molajo Compatibility that would insert any needed code in the template.

  • ssnobben

    Hi there Steve!

    Sorry I hijack this comments and just want to know if you know or have done some comparison testing of Joomla 1.5 template frameworks? ie Gantry, T3, Zoo etc

    Would be nice to have a professional judgment of those from a guy like you ;-)