[James Ross]

Genma Saotome, on 04 May 2016 - 02:44 PM, said:

My reference to "small ball" in the previous post can and should be rephrased into a more constructive POV, specifically, there isn't anything architectural there. In my own experience of running projects over the years I learned that if you don't get the fundamentals done up front they'll never get done because over time most of your resources will disappear and those that are left are busy with bugs and modest feature expansion within the structure that's present. IOW, small ball stuff.

I agree with this and have done a lot of work in the last few releases towards this goal - sacrificing almost all work on graphics in the process. There was at least one item on the board in this veign, "Separate 3D viewer from other code", but cjakeman has removed it - Chris, accident or got a plan? :)

Genma Saotome, on 04 May 2016 - 02:44 PM, said:

Perhaps there is a need to two lists: architectural and run of the mill features where the former is trans-formative and the later can go in where and whenever. Consider timetable mode -- it wasn't a simple extension of something KUJU had already done nor was it filling small feature gap they had left open. Instead it fundamentally altered how one thinks about scheduling. The core still need to be addressed.

We definitely need the core, foundational work - and I think it does belong on the board. It doesn't have to be me that separates the 3D viewer, for example, but I think people need to know we want to do these things, even if we're likely to do them in-house as the only way to get them done.

Some of the lack of such items is that I've done a bunch for recent releases and haven't come up with more. :) I have been mulling over some architectural stuff recently though, so I'll share that and see what changes are needed.

jovet, on 04 May 2016 - 04:58 PM, said:

I, for one, wouldn't consider offering suggestions to improve the Track Viewer since I see that as a separate product by a separate author. Am I mistaken in that notion?

Hmm, it's complicated. I would say that suggestions for contributed programs are fine if they are going to be worthwhile to OR more generally, but currently only make sense within that contributed area. They're definitely not at the same priority though.

cjakeman, on 05 May 2016 - 05:41 AM, said:

I noticed the dispatcher mode, which James has placed in 2.x.

What are the benefits we might get by packaging the multi-player mode as a dedicated server? Would they help to support tasks which are not driving, such as dispatcher, signalman, brakeman, fireman etc.?

A dedicated server would allow people to run servers without consuming lots of GPU/CPU resources unnecessarily, and maybe even run it on servers that don't have any displays, but I don't think it directly helps non-driving tasks. Any additional separation of components makes things better long-term though.