Use the fork, Luke

Open source means (amongst other things) that you have access to the source code. But what if you have no intention of even looking at the code, is access to the source code important to you?

It should be.

To illustrate why, let me paint a picture:

You’re IT department has installed a new piece of software across all desktops in the organisation. Your users like it, in fact, it becomes a key part of their daily life. Everyone is happy.

Over time the software improves and bugs are fixed. It’s time to upgrade to a newer version.

It’s going to cost money, so you enter a new procurement phase, the decision is upgrade to the latest and greatest version of the current solution or switch to another solution. The upgrade is cheaper, in terms of technical, licence and training costs and the competitors are not significantly better, so upgrade it is.

[Notice it makes no difference up to this point as to whether you have access to the source or not.]

More time passes, technology has moved on and your users are demanding another upgrade so they can do all the latest cool stuff. Time for a new procurement phase.

Now, lets imaging that the company behind your chosen product has decided to stop supporting it. It is at this point that access to the source becomes important.

A closed source product would have been left behind. There would have been no development from the point at which support was withdrawn. You have limited choices, stick with the old version or switch to another product.

On the other hand, an open source solution may have been picked up by any of its users, or more likely a community of its users and development could have continued. If nobody has picked it up you now have that option, instead of spending money on a new product, data conversion and retraining you could take the code and start a new project, either with internal resources or by paying a third party. Of course, you could still choose to switch to another product, but remember that your data conversion costs will almost certainly be lower since you have access to the source code of your existing solution.

This is exactly what happened recently when Linspire chose to discontinue development of their web authoring system Nvu. Users created a new community and have just made a bug fix release.

We don’t like to fork open source, but sometimes, if the original community is broken we need to. It is the potential to fork that makes open source sustainable.

4 thoughts on “Use the fork, Luke

  1. Tony Linde

    I’m not sure things are quite that cut and dried. I’ve seen plenty of commercial products where the original developers have given up and another commercial company has picked it up: I don’t remember any that have picked up again – best (like WordPerfect) is if they keep returning some revenue.

    With Nvu, I think it is more the case that there was no real community to start with, just the development team and the users. It is still early days with Kompozer: be interesting to see if it takes off. I’d be interested to hear of any other products where the original development died and was picked up and taken onto success with a real community around it.

    I don’t think one company or IT team has much chance picking up the source of a product and doing anything with it. I’ve done product conversions in the past and I don’t think the source code would have ever helped: generally get the data structure from the storage dbms. So source code for conversion to another product isn’t significant. As for picking up the code and trying to restart the development effort, that’d be beyond most IT teams, in terms of effort and desire.

    AFAICS most forks seem to happen when one group of developers falls out with another group in the same project 🙂

  2. Ross Gardler

    Thanks for your comments Tony. I agree that the options are not simple options. However,they are options that are not available under closed source licences. My point is that options are good.

    Consider this, if it costs less to take the software in-house, or to fund continued development via a third party then that’s a good option isn’t it? Sure, it won’t be easy, but neither is being forced to switch.

    Of course, one has to be careful when choosing a software solution, closed or open. One should always choose software one expects to be supported, but we can’t be right all the time.

    As for the code not helping in data migration, it really depends on the code in question. Well written software will be highly modularised, thus the data access layers can be pulled out and wrapped in interface code. This is a technique I have often used when I specialised in merging IT systems as a result of company acquisitions. Of course we couldn’t do that when the data format was undocumented in both code and text, but having access to good source code made the job far easier. Is all open source code good? No, of course not, but surely you inspected it before rolling out?
    Finally, you claim “most forks seem to happen when one group of developers falls out with another group in the same project”. I would phrase it very differently, but only to remove the emotion. I would agree that most forks happen when a set of active developers find that the existing project management has a different set of objectives to their own, and there is no other way to align those views. That includes a complete halt on development.
    Whilst I agree with your assertion this is a different point to the one I’m making in this post. Nevertheless it is an important point and leads me to…

    You say “I’d be interested to hear of any other products where the original development died and was picked up and taken onto success with a real community around it.” That’s the point of this post. The KompoZer community has made 5 releases since July 2006, still early days, but 5 more releases than NVU in that time.

    If you were dependant on NVU on your desktop wouldn’t you be happy it was opensource?

  3. Tony Linde

    The KompoZer community has made 5 releases since July 2006, still early days, but 5 more releases than NVU in that time. If you were dependant on NVU on your desktop wouldn’t you be happy it was opensource?

    I didn’t know that – Kompozer only showed up on my radar a few days ago – I’d given up on Nvu a while back. I agree it would be good to know it has been picked up though I’d be wary about relying on it a second time – but that is the same with resurrected commercial s/w.

  4. Pingback: Is MariaDB replacing MySQL? | OSS Watch team blog

Comments are closed.