Log 302 Warehousing

You and your rich friends are entrepreneurs in Silicon Valley, but you don't have the smarts to either build new devices or write new apps. (Sorry.) Instead, you're interested in tapping a more prosaic market; the geeks who build new devices, and write new apps.
These guys and gals, both employees and graduate students, never seem to sleep. And they get hungry at all hours of the day and night. This has been good news for the various pizzerias and burger joints that never close, and also offer 24/7 delivery. But you have a feeling that the market for prepared food is saturated. Further, it doesn't satisfy everyone's needs. What about the geek girl who feels the overwhelming urge to cook a tub of spaghetti sauce at 3:00 Sunday morning, but doesn't have any oregano? What about the farm boy, overcome with longing for his Mom's cooking, who wants some calf's liver smothered in onions? In other words, what about the weirdos who actually want groceries at all hours of the day and night?
You have a tentative name for the business: MyShoppingCart.com. Customers visit the site, and select items for delivery using one of two shopping modes: by store (specify a business, see what it sells, and pick items) or by product (specify a grocery item, see which stores stock it, and pick a store). As usual, the customers fill shopping carts online, enter their plastic, and await delivery. You plan to charge outrageous prices, but hey - this isn't a price-sensitive crowd.
Order fulfillment would take place in one of two ways; either directly from a store, or from your own small, very selectively stocked warehouse. Here's now it would work.
If the store is open, you send a shopping list to the store, and they fill a box for you to pickup - and add their own markup, for the extra work. If the store is closed, one of your own agents, bonded and preapproved by the store, opens the store up, gets the stuff, and leaves an invoice at Customer Service. If a store isn't open, and you can't reach an agreement with the owners to let one of your guys go rummaging through the shelves in the middle of the night, then that store wouldn't be on your website during the hours that it's closed.
If the customer orders by product, then you have two options; either go to the nearest store that has the requested items, or fill the order out of your own warehouse. The items in the warehouse either belong to you, having been purchased from a wholesaler, or are there on consignment - that is, they belong to local merchants, and they're letting you keep them and sell them on their behalf. Perishables are either frozen rock-hard, or (in the case of fresh vegetables) not available unless a market is open for pickup.
All in all, it would be an ambitious, enormously complex enterprise. It would be impossible without cutting-edge apps, which your geek coworkers are developing for you.
At the moment, you're in the fact-gathering, preliminary planning stage. The immediate problem is the warehouse. What information would you need to determine:
1. A satisfactory site?
2. What items would need to be kept in stock, and the optimum stocking level of each?
3. What size facility is needed - both floor space (sq ft) and volume (cubic ft)?
4. An optimum system for locating items, so your employees will know where to put them when they arrive, and find them when they're needed?
Pertaining to number 4 above: You're anticipating the need for some sort of scanning system that identifies items as they arrive, keeps track of where they're located in the warehouse, and issues instructions for retrieving them. Most, but not all, of the items will arrive with some sort of computer-friendly label already affixed; however, the label doesn't necessarily provide all the information you might need, particularly for the items that are there on consignment. For those items, it would be nice to know to whom they actually belong. You're also interested in the possibility of generating barcode labels for customers, to be affixed to the sacks and cartons containing their orders. The delivery people could read those labels with hand-held devices, and receive real-time driving instructions that take traffic congestion and road closures into account. The three labeling options are UPC barcoding, RDIF tags (either single-use or reusable) or 2-D barcodes using one of the standard protocols.
5. Which of the three options would be best? Why? Explain.
In conclusion, you should give some thought to safety.
6. What are the minimum procedures you should put into place, to ensure that your warehouse workers (probably no more than one or two people) aren't exposed to unnecessary risks?