But if your goal is to have a POS application, stop writing code right now. There exist hundreds of off the shelf POS apps all ready. For Windows, for Linux, thick clients, thin clients, web, desktop, green screen, etc...
Your time would be vastly better spent finding an existing product and adapting your business process to it.
Not always. There's a decent open source POS platform called OpenBravo POS. I successfully put that into our business without much fanfare. The first year was rough as I left too much of the original intact, mostly the product management stuff. I disabled their implementation and built my own C# app to handle the "back office" stuff and things were better. I beat the living hell out of that app to mangle it into how we think of things but that was WAY easier than changing business practices that didn't actually need to be changed.
And when business needs change (selling organic cheese by weight from a local farmer) I just modify it and roll it out. It's way cheaper than actually buying a commercial product given that we've got 26 checkout lanes. I'd imagine the break point occurs somewhere around 12 unless you're really specialized.