An Intermediary Format
by Ben Swanson · in Artist Corner · 03/21/2004 (10:44 am) · 1 replies
(Pre-coffee so feel free to ask me to clarify any incoherencies)
I was sitting on the can today thinking about allowing for more diverse art paths, specifically giving artists the power to choose what modelling app they want to use without having to worry about what file formats are supported by that app. The idea is pretty simple: 1.Design a format that stores any data that would be relevant to games -- mesh data, animation(s), textures, 'markers' (mount points and similar), and so on... File size, speed of loading, etc aren't of concern because this won't be used in game. Straighforwardness, readability, etc are essential. 2.Build exporters for it for a few of the more popular modeling apps out there, revising the format along the way as you encounter issues that may come up when exporting from different systems, keeping in mind that the goal is to have your data end up exactly the same no matter what you export from, despite the fact that you're holding a lot of data here. 3.Document the hell out of the format, document the hell out of the exporters for it that you've already written -- a small child or well trained chimp should have no trouble understanding the format. 4.Encourage authors of modelling apps not already supported to pick up support for the format, get community members to do it, or do it yourself. It will be relatively easy to write exporters to the format if you're familiar with the API the modelling app uses because the model format is both straightforward and well documented, and people will want to support it because they will gain more potential customers... 5.Game developers, engine developers, etc don't spend time worrying about writing exporters to their game format for specific modelling apps, instead they just write a single convertor that grabs the necessary data from the intermediary format (once again realtively easy thanks to plenty of documentation and a straightforward format) and writes it into whatever the format is for their game. They've had to do considerably less work to support a wide number of modelling apps, which will grow as more pick up support, which will in turn cause more developers to write convertors to their formats, in turn causing more modeller authors to support it, and well you get the idea -- it grows as it becomes more beneficial for people on each side of the table to support it / use it.
(I obviously do not own Visio, but while a tad messy, this is just as effective)
Does this sound like a reasonable idea? Does this even make sense? Any major, obvious problems? Other thoughts?
I was sitting on the can today thinking about allowing for more diverse art paths, specifically giving artists the power to choose what modelling app they want to use without having to worry about what file formats are supported by that app. The idea is pretty simple: 1.Design a format that stores any data that would be relevant to games -- mesh data, animation(s), textures, 'markers' (mount points and similar), and so on... File size, speed of loading, etc aren't of concern because this won't be used in game. Straighforwardness, readability, etc are essential. 2.Build exporters for it for a few of the more popular modeling apps out there, revising the format along the way as you encounter issues that may come up when exporting from different systems, keeping in mind that the goal is to have your data end up exactly the same no matter what you export from, despite the fact that you're holding a lot of data here. 3.Document the hell out of the format, document the hell out of the exporters for it that you've already written -- a small child or well trained chimp should have no trouble understanding the format. 4.Encourage authors of modelling apps not already supported to pick up support for the format, get community members to do it, or do it yourself. It will be relatively easy to write exporters to the format if you're familiar with the API the modelling app uses because the model format is both straightforward and well documented, and people will want to support it because they will gain more potential customers... 5.Game developers, engine developers, etc don't spend time worrying about writing exporters to their game format for specific modelling apps, instead they just write a single convertor that grabs the necessary data from the intermediary format (once again realtively easy thanks to plenty of documentation and a straightforward format) and writes it into whatever the format is for their game. They've had to do considerably less work to support a wide number of modelling apps, which will grow as more pick up support, which will in turn cause more developers to write convertors to their formats, in turn causing more modeller authors to support it, and well you get the idea -- it grows as it becomes more beneficial for people on each side of the table to support it / use it.
(I obviously do not own Visio, but while a tad messy, this is just as effective)Does this sound like a reasonable idea? Does this even make sense? Any major, obvious problems? Other thoughts?
Alex "IkoTikashi" Zappe