Ian Tree has written an open letter to IBM asking them to release Domino under an open source license. While I agree with him that open sourcing Domino could have lots of positive effects, he's ignoring the cost and time involved. All those good things might not happen given how hard it is to open source a large existing code base. Just ask Sun if you want to know how much work it is to release an existing product under and open source license.
- First off, it's going to take years.
- IBM will have to go through all the code, line by line (hopefully automated) and make sure that they own all of the intellectual property. Was it written by them? Were those developers under non-disclosures with other companies at the time? Did they leverage some of their patent cross-licensing? And so on. Any good lawyer could write pages of all the things they should check.
- Then when they open it, in the short term it's going to take more resources than they currently have, assuming they want to create an external community around it. Now not only do the developers need to work on the code, but they also need to interface with the community and bring other developers up to speed.
- And that community, while willing to help, will not be ready to jump in immediately. It will take a few years to grow that community – it may take a year or two for each individual developer to come up to speed on the project!
Picking the license and model (the parts Ian Tree suggests are hard) might be the easiest part. Making it happen will be the hard part.
So I'm not saying that opening Domino is a bad idea. To the contrary, it might be really interesting. I'm just saying that his reasons for why it's a good idea need to take into account that it might take 4-5 years to actually happen.
You know my University uses Notes and it is a total piece of junk. I would have a hard time believing that anyone would want to pick it up and run with it.
Notes and Domino are excellent products but expensive and often ignored by many IT departments because of the lack of M$ integration. I have always enjoyed using it but I agree, it would take years to create a viable open source version.
If Hula project is anything to judge by, the open sourcing of the code could be quite easy. What is hard is keeping traction on the project internally and externally.
With hula project, Novell dumped the code for Netmail out there and some development kicked off, the project then died off slowly as interest dwindled.
It could have been just what we needed in the open source world however it died a death…
Karl, thanks for the example.
Most companies dump code because they open sourced the product because it was old and they wanted out of the business. To create a vibrant community, it will actually cost them more in the short run.
Sir Marky, I wonder if they made it free for a while to see what the adoption rate was, if that would help them make a more educated decision on whether or not to open source it?
I am a big fan of IBM products and would love to seem them open up their portfolio but I don’t see it happening. IBM leverages open source where it makes sense. There isn’t much money in the OS market anymore, so offering Linux services is a given… the same reason it makes sense to sell Windows when you have your own AIX, i and z OS which are niche OSes designed to sell hardware and for existing shops.
I can’t think of anything that open sourcing Solaris really did well for Sun. It’s neat that they did it, and it probably got lots of testing and contribution to the x86 port. This was hardly a good business decision, especially the last part.
For the most part, the FOSS world likes home grown solutions. Take a look at the Firebird database, Solaris again, and the loads of Java apps/containers/etc.
I agree wholeheartedly with you and Carl: simply dumping existing code on the world and expecting magic and free maintenance is not realistic. The amount of vendor kick-starting required grows as a function of code size and complexity.
Kevin,
Your point that the FOSS world likes home grown solutions is right on. I think the preference for home grown solutions grows from a need to feel involved in the project from the beginning. It’s one of the reasons it’s hard to open source an existing project and create a vibrant community. It’s also the reason that I always encourage companies to open source their projects very early in the process, so that other people can weigh in early and help with the design.