I recently attended the Montreal International Game Summit event. Though I went as a student, and used the event to network, I had an ulterior motive: I spent some time picking the brains of professional developers about the feasibility of using Blender as a tool in their pipeline for creating games. I compiled some information that I received that will hopefully make an impact on the development of 2.50.
Keep in mind that these are NOT just my opinions, these come from industry professionals, artists, and executives. As such, they are very difficult to discount out of hand. That said, I do now tend to share these opinions, for the most part.
Firstly, I was amazed but the generally positive view of Open Source and Blender specifically, especially from Insomniac Games(they already use open source programs and have even open sourced parts of their engine). I think that there is an assumption in this community that users of other programs are closed minded to the prospect of using Blender, when in fact, this is not necessarily the case.
By far, the largest complaint I received was regarding the interface. We tend to complain that every time there is a new Blender release, the announcement threads on forums like CGSociety are essentially hacked by complaints about the UI. This cannot be handed off simply to a bias towards Maya or Max. The fact is, Blender’s interface, while foundationally strong, is quite mangled on the surface. I will not spend time re-iterating William Reynish’s UI paper, but there is some merit in this argument. Complaints were related specifically to the layout and location of buttons, and that there is poor implementation of any sort of heirarchy (i.e. the outliner). From the artists that had used Blender, I did recieve comments that once artists understood and worked with the interface, it was very positive, and they could model swiftly. However, getting to that point was a nightmare, and the time that training would take was not attractive or feasible in a production environment.
One of the attractive features of Blender for developers is it’s Open Source nature. They really like that it can be re-tooled to fit their pipeline. However, they have found that the code is firstly, poorly commented, and secondly, very poorly documented. Companies need to be able to dive in feet first and be able to make necessary changes quickly in order for this to be an attractive option. Not being a programmer, there is not much more I can add here.
Another stumbling block for artists trying to learn the software is the absence of reliable documentation. Personally, I learned from hundreds of different sources how to use Blender, often using tutorials dating from the 2.25 days that I had to figure out. The Blender Wiki is a great step in this direction, but, it is still somewhat unreliable. Companies and artists want to have every feature of Blender documented accurately to the most current version, and they want all of this information in an easy to find and navigate place.
One often cited reason for not using Blender is that, like I said, it is difficult to learn, and there is very little incentive to actually try to learn it. There is little that makes Blender unique from the standpoint of Maya or Max users. Most artists are trained to use those other programs, and they have no reason to switch to another, more difficult program to learn that will afford them no further functionality. Sure Blender can sculpt and composite video, but it does these things much more poorly than ZBrush or Final Cut Pro. So they will learn those other programs rather than Blender, and use them in tandem with their modeling software.
I have this entire next section devoted to the GE, because, of all parts of Blender, this recieved the most negative comments. The largest issue I learned was that the BGE tries to please everyone, and by doing so, it pleases no one. It’s not good enough at any one thing for a company to use it to produce anything. While it does support snazzy GLSL effects, there are somethings that it needs to have at the base level that are absent. First, there is no system for creating a GUI in game. Secondly, supporting bitmap text only is an enormous block. Realistically, the time spent managing the arbitrary formatting of bitmap text prohibits anything complex being built. An RPG for example would be incredibly difficult if it used any dialogue. From a design perspective, the BGE is impossible to work with since there’s no good way to store or organize huge types of information (lists of assets, lighting setups, etc.). The pipeline is very clunky as well. Lightmapping a scene is a good example of this. I’m not going to keep covering more specifics, but the point is, the engine is not feasible, and never will be feasible without some of these things. It will also not be feasible if it still tries to tackle all game types.
These are the main criticisms I received. My sincere hope is that the devs pay attention to this writing, as, like I mentioned, these are not my opinions, but the input of industry professionals on the viability of Blender. These should not be dismissed out of hand like I said. I’ve seen a tendency for people to dismiss complaints about Blender instead of trying to rectify the issues raised. This hurts everyone involved. It prohibits the growth of Blender, and turns people off of what is really, fundamentally a fantastic programs and open source achievement.
I hope this is wisely considered.
-Brandon
Edit: I also hope that the release 2.50 is very closely considered by the devs. I wouldn’t mind waiting for a year or more for things to be well implemented. 2.50 really has the potential to enable Blender to become a viable competitor, and I would hate to see it squandered.