Thought I could contribute my 5c here. I have been working in retail for around 4 years now, and dealt with PoS systems extensively: searched for the right software (protip: there is none), developed a master DB for synchronizing the product catalog across several locations, gathered reports and maintained the infrastructure. After all, I find the lack of good PoS software frustrating (at least here in Norway). The more or less affordable solutions are decades old, running VB macros on top of MS Access databases, they are often incompatible with versions of Windows above XP. Everything seems stitched together over the course of many years by various WYSIWYG Studio dilettantes. Extending such software with new functionality is often hard or even impossible, which their developers admitted on several occasions. For instance, I've once got an estimate of $5000 for integrating an old DIGI scale, with a condition that I'd manage to find a datasheet for that protocol. I ended up writing a small COM port emulator and used it as a proxy between PoS and the real port.

Fun thing about PoS is the huge assortment of various hardware, and all of it could be abstracted in business logic. Almost every country has it's own ways of collecting taxes, own currency, own central company registers and different ways of reporting -- all of it should be considered when designing, a dream of every architecture astronaut. In addition come these necessary things like stock control, handheld scanners, BI, printing, Unicode (never seen it supported by any PoS; imagine running an international foods store with ISO Latin-1) and last, but not the least, user-friendliness.

I guess the PoS software market is much larger in the States, but what about the average software quality? PoS systems are not going anywhere soon, and I feel that a startup dealing with these issues could definitely see success. If anyone else feels like doing it, I'd be more than glad to contribute.