Yeah, sorry will update the repo description to make that clear.
To attract more dev's, IMHO a major cleanup and a internationalization of the codebase is required;
* Translating comments and moving them "out of the way". Latter can be interpreted as changing them to trailing comments, moving/grokking them down into the header comments, improving naming to remove them altogether. I seriously volunteer for this one, since I'm already heads in to get my hands dirty on the code, and it would make my code-reading more useful :D.
* Dividing codebase into submodules and putting them into their own dirs... currently its one big mashup with undocumented prefixes (got my head around a few that were not so hard to figure out, but for instance "[no]Xxxx" still eludes me although I think to understand the category)
It would also be beneficial if the codebase could stand on broader footing, meaning;
* Separating engine from entity definitions.
* Making entity definitions some sort of non-programmer-readable/editable kind of thing.
* Building a set of tools that make the asset creation fanbase-sourceable and a little bit more streamlined.
* Convert those weird formats that nobody can do anything with unless they leverage libsiedler2, to something more contemporary and easily viewable/editable.
* Bottomline... lower the hurdle for people to build and contribute for it, assuming that for most C/C++ is already to low-level. If they could just push around pixels and visualize/simulate in-game behaviour of their editing in an intuitive way it may get some passive fans to get their hands dirty with asset creation.
On a separate note, Overall it does not feel like any of this project is really aimed at the public yet.
* Dev work is still sort of kept behind a fence. As far as the motivation for this goes (if I understood correctly) it can be handled by hosting a private/public gitlab (word of warning; its a memory hog)... keeping half of the stuff in the private and the other half in the open, eliminating a partial mirror in the open and all the hurdles to get via this mirror to upstream.
* Other than the forum there is no direct attempt to involve people in asset creation. Also no roadmap and/or project status to get potential contributors oriented.
* There is also no walk through to get BlueByte assets in place, which is death for the end-user. I understand that the ideal would be a complete deliverable package... but since it is not possible, maybe at least walk end-users through the portion which cannot be delivered by the package directly?
Maybe there is a need for something like a wiki?
Editiert von Arlias
am 28.07.2014 17:16