Of course stratum was a serious leap forward and the collaberation with the operator of one of slush's pools major rivals at the time helped cerment it's future.
Just FYI, Stratum wasn't developed with any collaboration. It was just a coincidence that while Slush was making stratum, BTC Guild (with the help of ArtForz) was also working on ASIC-capable mining protocol changes. The solution to fixing getwork's inefficiencies was very obvious, and in the end the two different protocols were virtually identical in function, just slightly different in syntax. The BTC Guild protocol was marginally more efficient (10-30% depending on block size) due to lack of markup, but Slush's was more in-line with other Bitcoin API and protocols by using JSON. It was wasteful, but it means new implementations of the protocol could use a JSON parsing library instead of writing their own parsers.
BTC Guild ended up supporting Stratum the same night since it only took a few hours to rewrite our new pool code (using the new protocol we had designed) to use Stratum, and Slush had a working stratum<->getwork proxy available already. It would've taken me longer to write a [unnamed-protocol]<->getwork proxy than it did to just modify my input parser to use the stratum syntax. Giving up some efficiency wasn't a huge tradeoff, considering both protocols used so little bandwidth that a 14.4k dial-up connection could keep up with it.
EDIT: Actually, I do recall both of us were aware that the other was working on a solution to mining protocols. IIRC, slush had been in contact with the poclbm developer, while I had been in contact with conman/ckolivas, the cgminer developer.