Theoretical Structural Archaeology is a way of thinking about building foundations from the perspective of the architects and builders that caused them to be created, and while I cannot do full justice to the questions Tim raises, I will attempt set them in this context.
The Layout of Roman forts by Tim Holland
I’ve been frustrated at the lack of interpretation information at Roman sites regarding practical details on the construction; a feeling of “why did they do it like that.” I’ve had no luck with the guides and internet searches, so when I stumbled on your blog, you seemed to have a similar curiosity and distinctly practical bent!
The fort complex has been partially excavated, and work continues, but the geophysics has mapped it pretty completely. The northwest section has been conserved and revealed some of the standard barrack blocks, but the orientation has the verandas facing northeast. There doesn’t seem an obvious reason for this; a southwest orientation would have been more comfortable? I can’t see an obvious reason for the orientation, but I may have missed something.
Even more puzzling, there’s a large latrine in the northwest corner and the drain appears to have been channelled behind one of the blocks running parallel down the slight slope. Again, this seems a rather odd arrangement. The outflow presumably ran out into the river; was this the standard arrangement in forts? I’ve seen the same latrine construction at Housesteads located more practically by the downhill side of the rampart, but there’s no river there. Did it just settle in a big slurry heap by the fort, next door to where the locals seemed to be living? Am I being too C21st, but this doesn’t seem a very ‘military’ solution? Maybe it was recycled into the fields?
The same appears to have happened to a grain store. Originally one large building with internal columns supporting the roof, its been reworked to provide 2 separate buildings, separated by a narrow gap. But the new walls are substantial, 2 foot thick and presumably supporting 2 roofs. Again, this seems a strange design.
Laying out Roman forts, by Geoff Carter
"The Praefect of the camp, though inferior in rank ..., had a post of no small importance. The position of the camp, the direction of the entrenchments, the inspection of the tents or huts of the soldiers and the baggage were comprehended in his province. His authority extended over the sick, and the physicians who had the care of them; and he regulated the expenses relative thereto. He had the charge of providing carriages, bathorses and the proper tools for sawing and cutting wood, digging trenches, raising parapets, sinking wells and bringing water into the camp. He likewise had the care of furnishing the troops with wood and straw, as well as the rams, onagri, balistae and all the other engines of war under his direction. This post was always conferred on an officer of great skill, experience and long service, and who consequently was capable of instructing others in those branches of the profession in which he had distinguished himself."
BarracksThe layout of a Roman fort was, to a reasonable extent, standardised; this extends to the types and design of building, and almost certainly to their structural components.
The essence of building is creating roofed space; this is the critical aspect of the project and is mostly concerned with protecting people and materials from water. Wide roofs are more complex, so you will note that Roman military structures are narrow and long. As a general rule, forts have roofs of c.25’ to 35’wide. It is possible that timber forts had narrower buildings. The larger central structures are usually comprised of a range of narrow buildings arranged around a central court.
The narrow gap between the barracks allows for a shared drain [or water butt], and saves space. The ‘long axis’ of the fort, as dictated by the orientation of the barracks, should run downhill. However, the fort at Inchtutil in Scotland has barracks in both directions.
Roman forts, while strategic, are not always placed simply with defence in mind. Architecturally, the axis of the fort was the main through road, the Via Principalis, with the principal buildings being laid out in relation to this. However, its placement and orientation would have to have taken into account the slope in terms of drainage.
One by-product of creating large areas of roofed space and metalled surfaces is water runoff, in addition to the waste-water created by the occupants. This is distinct from the problem of getting clean water into the fort, though clearly the two problems are related. In terms of construction, after site clearance, laying the drains is probably one of the earliest phases of construction. A Roman fort would rapidly have become a fetid mire without a robust drainage infrastructure!
The practicalities of drainage probably overrides other considerations, such as aspect, in the placement of barrack blocks.
The positioning of the latrines at Housesteads in the SE corner of the fort probably relates to the Kings Burn, the nearest watercourse, 200m to the East.
NB. The latrine pits for the C1st legionary fortress in Colchester are legendary – huge and almost bottomless, I’m not sure we ever found the bottom of one! .
"An army unsupplied with grain and other necessary provisions will be vanquished without striking a blow "
These buildings are traditionally very robust and well constructed, dry, with raised floors and good ventilation. However, even if you create the correct built environment for grain storage, there is the danger of infestation, a principal threat. Timber granaries were often burnt down prior to rebuilding, which may indicate an accident, fumigation with extreme prejudice, or even malicious action. There is always something of a conflict between centralising resources and placing them ‘all in one basket’.
So why replace a single wall [dividing/party] wall with a double, which Tim observed at Housesteads? This is also found elsewhere, such as at Hardknott, Cumbria.
Even with the wider examples with a single wall, I am inclined to see these as having two pitched roofs with a valley in the centre, rather than a single large pitched roof. My reasons for this are that there is no advantage to be gained by having a larger and higher roof; a roof base of 50-60’ would be significantly more than c.30- 35’, which is usual for stone foundations.
Pushing two granaries together may save materials, but a double wall with a gap is a more effective firebreak, and a better barrier to vermin, than a shared roof with a valley. As with barracks, the shared drain was in this case preferred to a common valley.
Many thanks to Tim for raising these issues, and I hope others will feel free to comment and raise others issues to do with the archaeology of structures.
Sources and further reading http://adriennegoodenoughtravel.blogspot.com/2007/11/caerleon-roman-fort-baths-amphitheatre.html (Caerleon)
 After figs 63 & 70 in Wilson, Roger. 1980. Roman Forts. Bergstrom / Boyle Books. ISBN 0 903767 07 4
 See Crummy, Philip. 1988. Colchester Archaeological Reports 6, chapters 2/3 & fig. 3.17. ISBN. 0-9503727-7-3