Samba-TNG fork

Everyone pretty much knows now that the Samba codebase has forked, with the TNG branch being split off into a new set of releases. Despite some hilarious reports this is actually a good thing.

For quite a while a number of Samba Team members have been working on a ambitious plan to provide complete NT functionality in Samba, using a architecture that differs considerably from the one that has been established in Samba over the last 10 years of development. It was decided by the team leaders that this proposed architecture was not appropriate for core Samba development which needs to take a more conservative approach to new developments. Instead the team members who have been advocating the TNG approach were encouraged to do their own releases, thus forming a code fork of the base Samba code base. After some consideration they decided that they would take this approach and have formed a new project that they are currently calling "Samba-TNG" (after the code branch they have been working on).

This new project will give the TNG developers the freedom to try some quite radical approaches to SMB server development, including such things as a DCE-RPC endpoint mapper, multi-transport support, a virtualised SMB layer and a daemonised architecture that may make modularisation of Samba much easier.

As the original author of Samba I am delighted that this split has occurred. Many of the design decisions in Samba are showing their age, but as Samba is so widely used it can be difficult to try radical new approaches while keeping the code as stable as users have come to expect. With a new project developers have a lot more freedom to try innovative solutions to problems without any concern about stability. While we don't yet know how the TNG project will work out, it will certainly teach us something about how their proposed approaches work when they are given the chance to be fully tested.

I look forward to seeing more development in TNG now that the developers are not constrained by the more conservative elements of the Samba Team (such as myself!) and I will be delighted to see the project flourish. There has been only one viable SMB server solution for the free software community for far too long, and a world with only one choice is a boring place indeed.

Andrew Tridgell
October 2000