Freelance. basicly, for the sake of this discussion, the art of scratching,or kitbashing a structure/bridge/locomotive/whole railroad/etc. from concept, as opposed to, from prototype. Walt Disney once said that things need not be accurate to be believed, just pausible. And what makes something plausible?....it has to look as if it really could have been. Freelance can be an excuse to build a diesel locomotive that looks like the starship enterprise (which has been done), and if that's what you want, so be it. Or, it can be the freedom to create something which might not actually exist, but looks as if it did. The art, then, is to look to the prototype, actually several different prototypes of the same type of structure. Look for, and note all the things these structures have in common. Construction, hardware, style, accessability, setting, etc. Why? industrial buildings will generally exist in similar locations, or have similar receiving/processing/shipping facilities, all of which contribute to the appearance of the building. Locomotives will generally have a similar appearance to other locos used in the same industrial application. They also have a common development history...some things work, and some things don't.Appurtanences will generall be mounted in "convenient to service" locations. That's why you will see these things in similar places on most railroads. What's easy for one to reach is usually easy for all. When you create a freelance model, the one thing you should try to do, is include the common things that the prototypes have, in a manner that is unique to your building. There are certain "conventions" that appear in construction methods. If your structure is to be a brick building, study brick buildings for the common items.(the material usually dictates the method of assembly, sometimes learned from previous mistakes) Wood structures, likewise, engineering is a trial and error process, and much of how things look today is the result of correcting the mistakes. You may not model the exact size, shape, or even material of the "prototype", but you should always try to model the "engineering". Observe, record,and use this information to design your model. If you do, you will have a distinct item, that looks like it very well could have been.