Announcing MulletDB (Formerly NoNoSQL)
sheddingbikes.com
sheddingbikes.com
The KV/document stores that have been coming out are trying to address shortcomings, perceived or real, with RDBMSs. I think some explanation of the trade-offs which each part of the "mullet" make and how they would be obviated by this system would be helpful. Right now there are a ton of KV/Document stores as well as SQL databases, and I don't know how this is going to be different or why I should pay attention to this project.
* Use protocols to scale, not backends
* Single interface to talk to db/kv store and disk
* In addition, use an SQL-like interface, if that's what floats your boat.
* Moving data between db/kv/disk is a server side operation.
Additionally, it's mostly a project I started on a dare that I started liking to hack on. Not really expecting anyone to use it for a while, if ever.
More competition, more innovation, better products for us to use in the future?
You probably don't need to pay attention to this project. Did you pay attention to Google when it first launched, or Apple when it made the Newton? If you're not that interested in the database technologies, you can just not pay attention, and hope that it becomes unavoidable if it is truly remarkable in the future.
If Zed does indeed release a Python driver under the MIT license, and you use that, it's essentially no different than using the LGPL in an end-user distributed app. You'd have to make any of your changes to MulletDB available to your end users, but your proprietary code could remain closed.
MongoDB takes this approach, and has written more at http://www.mongodb.org/display/DOCS/Licensing
I'm not clear about what the line is that says "everything from this point must have its source released" versus "all this stuff is OK to license as you please"