I believe you are right.  That was a typical implementation method.
Deborah

On 7/12/19 3:45 PM, Clem Cole wrote:
> If I recall this was one of the implementations that wrote to a file 
> and then forked the next process after it got to eof.
>
> On Fri, Jul 12, 2019 at 2:44 PM Deborah Scherrer 
> <dscherrer@solar.stanford.edu <mailto:dscherrer@solar.stanford.edu>> 
> wrote:
>
>     I didn't do this port, so don't know the details.  But it was done in
>     the late 70s (I think) and had broad distribution.  When I collected
>     various Software Tools versions, I was not able to find the VMS
>     one. Sorry.
>     Deborah
>
>     On 7/12/19 1:45 PM, Paul Winalski wrote:
>     > On 7/12/19, Deborah Scherrer <dscherrer@solar.stanford.edu
>     <mailto:dscherrer@solar.stanford.edu>> wrote:
>     >> There was also an extensive port of the Software Tools to VMS,
>     done by
>     >> Joe Sventek at LBNL.   Included at the key tools, the shell, pipes,
>     >> everything.   Felt completely like Unix.
>     > How did the LBNL Software Tools for VMS implement pipes?  I'm
>     curious
>     > because DEC itself did a product in the mid-1980s called DEC Shell
>     > that was a VMS port of the Bourne shell and associated utilities.  I
>     > wrote a VMS device driver that implemented pipes as a true VMS
>     > pseudo-device, similar to VMS mailboxes but with true Unix pipe
>     > semantics.
>     >
>     > -Paul W.
>
> -- 
> Sent from a handheld expect more typos than usual