I would call the "roadmap" of 2006 more of a "wish list", and certainly somewhat optimistic at the time, but we did manage to accomplish much of it which is great. That being said, we went through enough challenges and delays trying to deliver on our public goals (due to any number of very normal, common issues that arise in software development) that it became clear we needed be a bit less open and optimistic about things. People just get frustrated by it, and understandably so. Our approach at this point is to promise only what we know we can deliver in a relatively short time-frame, hence the Terragen 3 announcement more or less out of the blue just a couple months before final release. We are likely to continue that approach, with perhaps the occasional teaser or early look at a feature in development. A fully laid out development roadmap is not likely to be something we will publish in the future. If you look around the commercial software publishing world you'll actually find fairly few public roadmaps or other majorly forward-looking feature discussions from the developers; open source is naturally quite different, though even there you seldom see published lists of features planned for implementation.
Of course we do have an internal list of features and functionality we want to add. In fact we have a pretty detailed plan out to Terragen 4 and beyond at this point. And that's important to know, too - even though there is no publicly available plan, we're certainly keeping track of requests, feature ideas, improvements, etc. Equally important then is the reality that we have our dev plans pretty well solidified, with dev time estimates to get us where we want to be for future releases within the release windows we want to hit. This means that there's less room for random feature suggestions than would be if we were kind of taking things as they come, throwing features at the wall and seeing what sticks, but a more deliberate product/feature strategy has so far been a notable improvement for us and for users, and we want to continue with that approach. We still certainly welcome feature suggestions and implementation feedback, but there are few really new ideas at this point, almost anything one can think of has been discussed before in various threads sometimes going back years. Sometimes a feature doesn't get implemented because we decide it's not valuable enough to the product, sometimes it's just a matter of implementation time or development priority, i.e. we'll eventually get to it. Everything that has been mentioned in this thread has been discussed at one time or another, some are intended for future implementation, others we may reconsider when the time is right.
You mention competitive concerns as well, and yes this is another factor. We've seen competing products implement very similar - if not identical - features in remarkable succession to one of our releases with that feature (good examples would be full planet model and cloud altitude/depth modulation functions that allows "cloud follows terrain" easily). That is inevitable, of course. But it's not unreasonable to expect that if we had a clear roadmap laid out, some of those feature implementations might come even sooner, i.e. competitors could work to leapfrog us in features (especially as most of our competitors have larger teams than we do). Our edge remains in the quality and stability of features that we implement, but if we can pull out a few surprises, that helps us too.
We certainly hope you'll all be excited about where we take Terragen next. It's really our job to make sure that you are excited about it because that's how we make our living. If we fail to excite a majority of our user base and the wider market, it threatens our very livelihood! So for us the stakes couldn't possibly be higher, and we take the process of analyzing and predicting the market and its future needs very seriously. Part of that is looking at your feature suggestions and ideas, both explicit and implied, as well as deriving good feature ideas from problems or workflow needs we observe. We may make mistakes here and there in the process of prioritizing and implementing, but I think our biggest challenge continues to be development and overall staff "manpower", not lack of ideas or ambition. If we had a larger team, I have no doubt we'd be blowing your minds on a regular basis. As it is, we have to settle for hopefully blowing your mind once every year or two.
To more specifically address some of your questions (keeping mind that I can't lay out our larger, longer-term plans): Certainly more stability and minor updates are coming. As well as not-so-minor, I'd like to think. We have a 3.1 release planned for early in 2014 which should have one or two notable new (but not Earth-shattering) features in it, along with many bug fixes, etc. We are likely to release one or more other "point release" updates through 2014 as well. These will be free for all users of Terragen 3, as has been our previous policy.
We are absolutely not moving in the direction of a general-purpose 3D program. While other more general-purpose applications do seem to be adding some competitive features, and we watch industry developments in that area very closely, we continue to feel that by focusing in the area of modeling and rendering natural phenomena, we can remain a relevant and effective part of many artist's tool set. We are continuously evaluating various options for diversifying and strengthening our position, but we will almost certainly always remain pretty tightly focused on this area.
We do have some plans for completing the Wiki documentation faster and better, along with other potential documentation-related improvements. We haven't lost sight of the importance of it, if anything it's become even more clear over the past few months. But we're also working on partnering with other content producers and supporting the development of learning resources like Geekatplay's recent releases.
Regarding GI, Matt will probably kill me for saying so, but we *have* experimented with "brute force" GI. It's quite slow, as you would expect. So it's really not likely to show up as a feature any time soon. Instead we're likely to continue innovating for higher-performance solutions like the improved GISD in TG3. Brute force is useful mostly as a reference, but it is of course the most simplistic and generally slow way to achieve the end result; more clever and optimized solutions can achieve similar quality levels in far less time and so are a better use of dev time.
In answer to Badger's comments:
We do plan to add Populate on Object some time in 2014.
Mudbox/Zbrush integration, not entirely sure how to handle this to be honest. It appears possible as-is (as the workflow threads have shown), so it may be more of a documentation/verified workflow issue, which frankly I can't help with as I(we?) don't have either program. Autodesk isn't exactly highly responsive, but we may have some luck talking directly with Poxologic to see if they have any input on an ideal workflow.
RTO displacement is *not* likely to be implemented any time soon, unfortunately. What we will do (again, some time in 2014) is add an option to affect RTO per-object/population.
Alessandro:
Don't hold your breath on a particle system. It would be nice but is far too ambitious an undertaking at this point, especially considering how good many existing particle simulation systems there are. Our general approach is more to improve interoperability with other applications so you can get the best of multiple tools working closely together. There are whole software companies focusing on particle simulation tools so if we can help users integrate them in some way with Terragen that is likely the most effective use of our time (e.g. geometry export for particle simulation interaction, environment maps for rendering, etc.).
An improved populator with some of the features you likely have in mind is planned for 2014.
Painted shader improvements are certainly something we're thinking about.
An AO/Dirt Map shader would be nice. I'm not sure of the feasibility, but it's something I've mentioned to Matt before...
Spline terrain generation is also a somewhat "big" feature, one we're definitely interested in, but not likely to come real soon practically speaking.
For everyone else: lots of misc. feature requests here and there, we'll take note, but may not be able to respond specifically to each one.
I will say though, that the requests for "stability" are a bit surprising to me. We do aim for very stable releases and in general that seems to be something we've done well with over the years, in fact we'v had a lot of feedback, particularly from new users, that Terragen is comparatively stable. If you have specific circumstances where there are stability issues that are *not* simply related to basic system resource limitations, then we'd love to hear about them. We are aware of a few such issues recently, for example the "abort while populating" crash reported in another thread which we're looking into, and we fixed a few stability issues in the 3.0.10 release after the TG3 launch, but if you do have other stability-related feedback and experiences, please let us know with more details and specifics.
- Oshyan