I sort of agree with Frosty but, i have just started a project and elected to do it in 1.4 because.
- I already have a lot of code blocks and whole modules i can reuse and time is of the essence
- S2 is not stable and what i am doing needs to be stable from day one
- Documentation for S2 while good is not at the same level as that for S1.x
- A lot of the plugins for S2 are works in progress whereas many of those for S1.x are solid and again well documented
- Speed performance is not an issue so S2's gains are no help to me
- The client does not care as long as they get a deliverable
- S1.x (well 1.4) will be supported until end of 2012, over 18 months which is often a lot longer than the average life of a pice of software, especially in the web arena
- There is not native admin backend and i really need one that just works
- In 6 to 12 months in can reverse engineer, by hand, this project fully working and tested in S2.x with oodles of documentation to hand
- Fabpot is a very clever guy and will probably have written a script to convert 1.4 to 2.x so i don't have to worry (ok perhaps he won't but all the other hold)
The above boils down to yes S2 is probably stable enough and the documentation is useable and if you have enough time to go up the learning curve but it will depend on individual cases as to which way the cookie crumbles.
