A legion on the march typically built a castra every evening, or as close to that as they could manage. That goes for both the legion of the Middle and Late Republic and also the Empire. Our best sources for their layout and construction are Polybios and to a lesser extent Caesar, as far as literary, and then the archaeological evidence of the numerous castra of the Roman period found all over their former territories, but especially in Britain, where the unique soil composition preserves their construction very well.
The process was begun by specially-trained engineers. The gromaticus used a groma, a sort of ancient survey device with sights and plumb lines, to mark the center of the camp and establish the initial geometry depending on size of the army and the expediency of the situation. From there, guys with decempedae "10-footers" marked out the internal structure, the course of the fortifications, etc. The construction and digging was done by the soldiers themselves, and they were as good at building as they were at fighting. There was no dedicated, non-fighting force of laborers, though we assume that slaves in some quantity were present in most camps. Each soldier had a specific job in a specific location of the plan, knew exactly what he was to do, and the whole process was accomplished with frightening speed.
A basic castra has a trench or ditch (fossa) around all four sides of the square/rectangle area, and the earth excavated to make that ditch was piled up on the inner side to make a rampart (agger). This served to double the "height" of the obstacle from the bottom of the fossa to the top of the agger (diagram). Atop the agger the soldiers created a rampart of wooden stakes (sudes), at the very least, which the soldiers carried around with them. An internal wall of timber was optional depending on available time and resources (the vallum). There was some variation depending on terrain and whatnot. Internal "streets" were laid out, always in the same pattern, upon which the centuries laid out their tents. There was also a central area (the "forum") for the command tent, for religious rituals, and for addressing and review of the troops. If there was timber readily available, further fortification could be elaborated from there, given resources and time. This would often include elevated watchtowers at intervals along the vallum. The just-for-the-night camps were usually not terribly elaborate, since they would be dismantled the next day. If the legion(s) would be spending an extended time there, like when Caesar's legions went into winter quarters, then more time was devoted to sprucing up the place. Timber and stone ramparts, stout towers, more comfortable barracks with stone foundations, sanitation works, granaries, slaughterhouses, smithies, and even little bath houses are all known from examples. These long-term forts also attracted new villages staffed with all the civilians who can profit off the presence of thousands of soldiers with wages to spend, especially those performing the world's oldest profession. The forts along the borders of the empire, at Hadrian's wall or along the Danube limes, for instance, were even more elaborate and got more complicated over the decades.
Was there a common theory about building camps?
Yes, and there was remarkable consistency across time and space. We have no idea how engineering know-how was disseminated in the Roman world, or how it was learned. I wrote a lot about this topic over here.
