There were options for migrating from PDP-11, not PDP-10.

PDP-11 was a 16bit byte-addressable minicomputer that was extended to 22 bit physical addresses through MMU. VAX took PDP-11 and extended it into 32bit flat addressable paged system.

PDP-10 was 36-bit, word-addressed with 18bit addressing (plus segment extensions in KL-B that were only really implemented in 3rd party follow-ons, a mainframe-in-all-but-name.

VAX would mean starting from zero unless nearly everything was written very high-level, which is doubtful given the optimizations needed to run so many simultaneous users.

▲

FrankWilhoit 5 hours ago | parent [-]

It had more to do with the ecosystems and the corporate relationship. The -10s came from a earlier perspective. It was open source before "open source". You bought the software and you got the source, because of course you did, you couldn't even install it otherwise. Compuserve tracked TOPS-10 up to version 6.01, with their own additions and patches: far-reaching but well encapsulated. They could not track 6.03, so they "forked" it, to use the modern word. The question of support was always fraught. Compuserve offered all of their changes to DEC, who mostly sniffed at them. But from DEC's perspective, one of the major purposes of the VAX platform was to lock all that kind of thing down and go to a support model more like that of IBM.

	▲	p_l 4 hours ago \| parent [-]
		Another part was pretty vicious infighting between PDP-10 (already sorta on thin ice from start due to fuckup that was PDP-6) and PDP-11/VAX teams, and VAX team won it pretty decisively - if in part through consolidation efforts. EDIT: Another major difference was how PDP-10 was from an era where everyone with a (under)graduate students to throw at the problem was making custom modifications to hardware. Like some early multiprocessor PDP-10 being done with hand-soldered MP bus done by students. It was a wild place