Here's a repost from the Facebook page that talks about some improvements in workflow, etc. that we're doing for Baldur's Gate II: Shadows of Amn Reloaded (SOAR).
------------------------------------------------------------
Now that we've announced Baldur's Gate II: Reloaded (or SOAR as we like to call it), I wanted to start an initiative to do a better job sharing information with you all throughout development, which will hopefully result in more regular updates and "mini-blog" posts.
I wanted to share some of the ways we're looking to improve upon BGR in SOAR. A few months after BGR was completed, I created a "post-mortem", which is basically a reflection on what worked, what didn't work, and why. As much as we love BGR and are proud of the work we did, there are a lot of ways in which we could have done things differently, at a higher quality, and more efficiently. We learned a ton of things during the development of BGR, but, as a saying I just made up goes, it's hard to turn an oil rig around; it's sometimes better to just stay the course. With a new project, we get enact these changes from the start, which I'll detail more below.
1. Identifying a critical path. One of the most important aspects of pre-production (before you actually start developing a project) is to identify a "critical path" or what needs to get done first before other things can get done. In game development, this can be quite tricky because there are often a lot of cross-dependencies. In BGR, there was a lot of overhead in making any changes to our area scripts, creatures, or items because they were made in the wrong order; hundreds if not thousands of hours were wasted due to this. For SOAR then, we decided the first thing we would do is to compile a list of the custom content we will need and create placeholders. Following that, we could create all the item icons, then all the items. Afterwards, we could then make all the creatures (not NPCs)in the game and equip them with the newly created items. This is all done before any work on levels begin. The list goes on, of course, but you probably get the idea.
2. Using the campaign directory more. For SOAR, we've decided to move as many creatures, items, scripts, and dialogues to the campaign directory as possible and where it makes sense. In BGR, we ran into a lot of issues, especially early-on, due to each module having its own set of items. If we ever wanted to update an item, we had to update them all. Needless to say, this was tedious and error prone. Furthermore, because NWN2 stores items in stores and loot containers as separate instances of the item, updating an item in the module wouldn't do anything to the items already placed in a store or loot container. In BGR, we suffered a lot of items having an incorrect cost value, icon, appearance, or even stats. By moving everything to the campaign directory and following our critical path, issues such as the ones we experienced in BGR should be minimized greatly.
3. More testing throughout development. In BGR, we tested as we developed, but not always in the most organized way. Something almost every game developer does is to create test maps in which specific features can be tested quickly and easily. While we had a limited number of test maps in BGR, we intend to create more test maps to create stable, and ideally final, features as early as possible. For example, in SOAR, we created a test map for testing each item to ensure they behave properly, look correct, cost the right amount, etc. It's important that we get this right early so we can avoid having to make costly fixes later in development.
4. Encounter balance. Some of the encounters in BGR were not balanced in a way with which we felt happy. While a lot of this came down to the rather open nature of BG and our focus being diverted to actually completing the game, we want to make the encounters in SOAR as balanced, challenging, and fun as possible. Somewhat related to the above, we plan on having test maps to more thoroughly test creatures and major encounters in the game to avoid some of the difficulty (or lack of difficulty) spikes seen in BGR.
5. Better custom art assets. While most of the custom art we borrowed from the outstanding NWN2 community is extremely high-quality, the custom art assets we created were not always of the same quality level. In particular, the custom BCK textures we created stuck out a bit and didn't match the style of quality of the surrounding assets. For SOAR, we've already created a lot of custom assets that we believe are vastly improved and we can't wait to show some of those off soon.
6. General polish. While I think we ended up releasing a game that was pretty well-polished in the end, there are a lot of features we just didn't have time to tweak and perfect the way we wanted. In particular, resting felt a little underwhelming and not as cool as it could have been. For SOAR, we fixed this and we've rewritten or rest system which we'll hopefully demonstrate soon.
7. Improved Creature spawning. Many of the creatures in BGR happened to spawn in sight of the player and often in odd locations, such as behind the player. While some of this was related to the rather open design of the levels, there are some tweaks that could be made to the encounter spawning system that would have fixed many of these issues. Furthermore, creatures tended to spawn in weird groups (e.g. a zombie spawning with a wolf and a kobold--what?!?). For SOAR, we think we've implemented a fix to both of these issues.
8. Better level layout and gameplay. One of the problems we had in creating fun encounters in BGR was that the level design was not designed to support certain kinds of encounters. This was mostly due to the areas being designed and often polished with final art before any scripting was completed. Because of the large amount of wasted time it took to redesign a finished area, we often could only make minor changes to spawn placement or minor level details. In SOAR, however, we are developing levels in a multi-stage process which will allow for more rapid gameplay iteration. The first step is what's called "level blockout" in which no attention is given to the art of the level, only the "feel" of the playable area. During this stage, we ensure the walkmesh is exactly what we want (so the player can navigate everywhere we want them to go, etc.) and the scale is correct. During this stage, we'll also put in doors and other area transitions. The next stage is to script the area for gameplay. During this stage, we can still shift around the area to ensure our layout supports the best encounters possible. The final stage is polishing everything, including texturing, using final art, audio, etc.
9. Smaller module size. In BGR, we grouped a large number of areas into a single module, which was great for short load times between the areas within that module, but made for some very long load times when loading into the module the first time or restarting after a death or a saved game. This was most easily noticeable with Baldur's Gate city, in which the load times could easily take over 30 seconds to a minute, which is far too long. Therefore, we will be making every exterior area and its associated interior areas into their own modules. This will help normalize the load times and even provide some workflow improvements since Shallina and I can stay out of each other's way easier.
10. Delayed Beta Testing. While our talented beta testers provided excellent feedback and bug reports for BGR, we weren't able to utilize the beta as well as we could have as we were still polishing and making the game. Instead, we will be pushing the beta until we believe the game is done, so we can focus on only fixing bugs to ensure for the most stable release possible. We were proud of our work in BGR, but we would have liked to have the stability we have in Patch 3 available at release.
There's definitely a few other items that I could have mentioned here, but perhaps I'll save those for another time. We're incredibly excited to be on our journey of recreating one of the best RPGs of all-time in NWN2. Thanks for reading and let me know if you'd like to see more posts like this!
Back to top
.
