Now we are in my area of expertise I feel safe to throw my two cents in without coming across as simply bitching.
The tool as it is just now is very incomplete and I understand that, everything I say here is based on this and I may talk about it as it is but I will also talk about how it may develop.
The real test of this tool is what Artists will do with it, not in the sense of what is created using it, but in what assets will be made for it. I can see a lot of potential for it, it is far superior to segment building alone which I understand is quick and easy but lets face it very limited though I do still believe there is room for this method of building. Segments, or prefabs if you want to call it such now, seems to me are exactly as they were but now with the ability to snap them together whatever their scale, due to vertex snapping rather than grid. Of course this will also mean limiting assets to simpler shapes for use in this tool. There is no point in thinking of organic models being used with this, there is no advantage at all in placing a rock for example or a pile of rubble in this way versus simply placing it in editor in the usual way. I know this sounds stupid but if you think assets for this tool will expand beyond walls,roofs and simple shapes for building using it alone, then simply no if you want fuse boxes,pipes etc whats the advantage over entity placement to fixed position in the mesh?. But as I say prefabbing areas is where it's at with expanding on this.
The big advantage here is building prefabs from mesh's and storing them for use in other areas of the level. I wouldn't want to be trying to create an entire level with this using walls and floors, but the ability to import and build areas is really exciting and a real time saver for any developer. This should be expanded to selection of entity's in the top down editor and stored in the construction kit so outdoor areas, entire towns etc can be re-used and re-edited.
This tool is ideal for creating corridors and rooms, and your correct, I expect there will be screenshots of the massive outdoor structure created and posted with the comment 'my city I built' which won't show how bad it's lagging. Many users will create something using the broken down prefab parts at least once, then never use it this way again unless assets are well thought out and designed specifically for use. If a mesh building facade is imported then it won't work too well with floors, which don't match scale exactly and will lead to areas of Z-clashing,overhangs etc unless individual meshes can be scaled. Again if Artists create single models and don't have an actual theme and design which blends all assets together then it's going to be impossible to get much use out of it not to say frustrating.
This is where packs would be indispensable...something TGC don't want to deal with anymore preferring single assets in the store and something I don't want the extra work doing, particularly when it comes to component parts. If packs can be uploaded to the new store and administrated by the vendor themselves then I will be fine with that.
The problem with users will be that expectations will be too high that they can create their entire game using this alone, but it is as I say a huge leap being able to prefab small builds aside from rooms etc, this should be made a priority part of its development. As should the lighting system required to use it for indoor scenes.
The roller paint huds etc can go, they make this look like a toy and are completely unnecessary, probably using up resources better saved for actual developing, an icon will do fine.
To conclude I would be happy to create my models as component parts where a prefab can be un-grouped and parts re-textured,re-arranged removed or added to, which incidentally, will lead to more draw calls than a single texture model so should be used sparingly. Just expect these to cost a little more than the single mesh prop models.
There are advantages with single mesh models which can utilise the LOD system for instance but for areas you want added control over design this would work well. Anyone thinks I would create LOD for every single component part is deluded and obviously never did much modeling....joking of course
It would work well however if component removal from render was tied to distance/view via script (at least for outdoors where occlusion may not be effective) or uses the imposters system. Regardless, component parts should be lowest poly count possible anyway so LOD shouldn't even come into it in most cases. The ability to remove parts that are never seen at all would in fact be an advantage to the more savvy and thoughtful level designers who look to performance.
Here's something I was already working on that illustrates everything perfectly.
Aside from a single mesh model in the store it could be an editable prefab with all the parts being re-usable, more mileage for the buck right there. This is just one use I see for this tool and as it develops probably more will come, just drop the fluff so us grown ups feel like it's not aimed at 5 yr old's.
On my wish list for this would be import from modeler which has the parts pre-built or grouped and turns this into a prefab, which may be asking too much but would be an ideal time saver at my end.
NOTE: rolfy is not writing the Construction Kit (or the Rift Support, Object Importer or new LUA Engine system) and remains focused on models to the exclusion of all else (except forum posts of course). Rest assured all progress on these auxiliary modules are running in parallel to the main model making and are not distracting rolfy from what needs to be done
