[HECnet] Re: DECnet equivalent of PSH?

I don't know if I would go so far as to say hardly relevant.  The Unix stream model is quite close to Tops-20's.  For TCP, there is a certain amount of shared conceptual ancestry.  Both Tops-20 and Unix have the idea of attempting to divorce the program from the data it is reading, which allows you to pick different devices with relative ease at the user level (think pipes) BSD broke this paradigm with sockets because now you can't pipe into them.  Plan 9 doesn't do that.  Tops-20 broke this in exactly the same way except that they're called JCN's, but you have an entirely separate set of JSYi to talk to the network.  This is what I mean by a shared conceptual framework as I believe these were two of the early TCP implementations. Tops-20 dropped the JCN stuff to switch back to a JFN interface, which makes programming this a lot simpler.  MRC and the rest of the ARPA community made a pretty big stick about doing it "the right way" When you are trying to figure out the way something works, it is very useful to see how different operating systems are implementing the protocol.  For Tops-20 FTP, I had source for Tops-20, ITS and BSD FTP and that helped me understand a conceptual and then fix a problem with passive mode (which the Tops-20 FTP client had not implemented). It is a very sad fact that the number of operating systems in active use has vastly decreased because there are less cross fertilization opportunities.  And if you want to bullet proof your code, have it run on different OS's... I really grind my teeth when I see some You-Boob'ers talk about different OS's when they're just really different versions of Windows.  Same API, different glitz.  Big deal.