Suggestion on new feature requests.

[ciraga]

You should use this tracker for new feature requests/ideas and/or set a new forum/wiki for them.
The point being that some features influence design choices and it's good to catch them early on.
Not to speak about bugs, but those are handled better "live" on irc.
For my chip in, while Adom is not FOSS it's the *rogue that think got things right, so please check it out.
More ideas to income in near future, stand by an issue tracker near you.

[JohnAnthony]

I'm watching this issue -- this is a great place to formally discuss new feature ideas and the project's direction.

[notwa]

1. Where do you plan on taking this? At the moment TROG seems like the basis for a roguelike without anything too special or a clear direction, thus making it difficult to provide suggestions.
2. Are you going to halt development on TROG and start a similar project in C? I heard you considering it on /g/, but you don't seem serious about it.
3. Is requiring C++11 really necessary? I know everyone's being pretty quick on support for it, but when I attempt to compile without -std=c++0x, most of the warnings/errors that occur are related to the way data is entered and initialized. In other words, not much that you couldn't rewrite to work under the previous standard.
4. Do you have any plans to externalize the game data and load it at runtime? The pros and cons of this are pretty obvious, and I won't nag about it for a (currently) small game. I'd just like your thoughts on it.

[JohnAnthony]

1. I want to make TROG a complete but extremely vanilla roguelike and release a 1.0 version. I then want to play it and have tons of fun and add all the stuff I think is fun to have in a roguelike at my (and others') leisure.
2. I have no intention of halting progress with TROG. It's here to stay and I'm intending to have it as a fun little hackable project for many years to come. I WILL, however, be beginning work on another game once TROG is a happy and playable 1.0. That game will be using a similar combat system but will be more like Might & Magic than a roguelike, complete with fractal terrain, cities, overworld travel and a whole party of characters. That's a big project, though, and I'm intending to be playing and hacking on TROG the whole time I'm working on it.
3. C++0x brings in the only feature I really care about in recent years with C-like initializers. Yes, people are getting warnings and I'm not particularly attached to them. If somebody wants to re-write the initializer lists to simply be inline constructors I'd happily include that in the project and drop the C++0x requirement but I'm not prepared to do so myself at the expense of coding time doing things that actually matter (like finishing the magic system).
4. No plans whatsoever to externalise the game data. Including a parser creates bugs and will add, at a guess, 2k LOC to a currently 5k LOC project (with aspirations to be a complete roguelike in well under 10k LOC). I just don't think a project so tiny and succinctly small needs it.

[JohnAnthony]

As an update regarding #3 -- I just worked through the code for compatibility and gcc 4.4.0 now compiles without warnings. I haven't tried anything older, but I bet it's fine. Considering Debian Stable uses 4.4.3 I think compatibility with older versions is now a non-issue.

[alcioa]

Forgot to ask : what are the requirements for write access to the repo?