It's a good idea but fairly challenging in practice. The amount of information you need to reveal in a scientific publication may make many companies uncomfortable. My view is that academics should not just be focused on getting another paper on their CV -- there is value in having the industry experience even if no papers come out of it.
I'm not in academia per se, but (in Germany) just getting to do my (equivalent of a) Master's thesis at a company (still with a supervisor at university) was hard enough - because sometimes they just seem to search for people to do work based on their topic suggestions and if you want to do something that's a little off the tracks... Let's just say there were no real arguments, mostly just "we don't like companies that are not university spin-offs and oh shock that company might reuse that thesis to advertise their exploits instead of granting the kudos to the university.
But maybe that's not the general case, but as someone coming from the tech sector to get a degree (and not the other way round) this sounds all too similar :)
