Browsing through Steamindex having awoken in the early hours I happened on a mention (by LA Summers) of a GWR Dean era proposal for a water tube boiler on a 4-4-0. You'd think that came out of nowhere, but a couple of months ago I was given sight of part of the Swindon drawing office register of drawings for the time when the 3521  0-4-2Ts were being worked on. One thing that struck me was the number of drawings being produced at Swindon for the GWR's ships. They clearly didn't maintain a separate drawing office or outsource at least some marine work, even though I don't think Swindon designed their own ships.

It also seemed evident that at least some draughtsmen worked on both marine and rail drawings. I'm not going to double check now, but if memory serves me right a young G. J. Churchward worked on drawings for a marine boiler. One assumes, too, that at least the keener young draughtsmen would be readers of trade publications like 'The Engineer' which covered a very wide range of engineering topics.

Now I think of it I'm also reminded of Cook's tale of how big end lubrication for Kings, Castles and eventually LNER A4s was sorted out with inspiration from the design of a machine tool in Swindon Works. Collett did his apprenticeship with Maudsley's, a very high status marine engineering firm too. 

We're accustomed to think of a silo mentality in railway design in the 20thC, and there certainly was some of that, but equally the above suggests that design staff had a rather wider range of experience than we might expect.