That probably depends on the result. The current heuristical scheduler
provides reasonably good results for most common cases. It's far from
perfect, but fast enough for practical use.
It's certainly possible to provide alternative schedulers. If you can
demonstrate good results with acceptable performance, I'd be certainly
interested. See the 'How To Contribute' section in the manual for some
guidelines.
I'm optimizing the scheduler performance right now. Many internal data
structures are in flux. I should be done in a few weeks. But you can
certainly start looking at the code.
Chris
The server is written in Ruby, just like the rest of the applications.
We make very little use of Java script.
Chris
On Thu, Jul 7, 2011 at 2:20 AM, Jeff Chimene <jchi...@gmail.com> wrote:
> Attached, please find a spreadsheet implementation of the proposed
> scheduling simulation.
you probably have to explain a bit more what this is, what it can be
used for and how.
Chris
I still haven't understood the relationship between TJ projects and
your spread sheet. What's the purpose? How should it be used with TJ?
Chris