[Info-vax] RSTS/E and VMS connection
Johnny Billquist
bqt at softjar.se
Tue Apr 11 16:26:12 EDT 2023
On 2023-04-09 20:29, Arne Vajhøj wrote:
> On 4/9/2023 12:15 PM, Johnny Billquist wrote:
>> On 2023-04-08 00:26, Arne Vajhøj wrote:
>>> On 4/7/2023 5:22 PM, Johnny Billquist wrote:
>>>> On 2023-04-07 01:02, bill wrote:
>>>>> WE have often heard that VMS on the VAX was the logical procession
>>>>> from RS11 on the PDP. I was looking around at my bookshelf and saw
>>>>> my 1983-84 RSTS/E PDP-11 Operating System book and it kinda piqued
>>>>> my interest so I picked it up and did a little lite reading.
>>>>
>>>> Well. Internally, VMS is like a scaled up RSX in many (most?)
>>>> corners. Most data structures and designs were (are) more or less
>>>> the same as in RSX, but scaled up for 32-bits and whatever logically
>>>> came out from there. A lot of system calls is the same story. And
>>>> initial release of VMS mostly used all RSX cusps, running in
>>>> compatibility mode, and slowly getting replaced with native things.
>>>> Early versions of VMS I seem to remember even might still have used
>>>> ODS-1, or some hybrid between ODS-1 and -2. I clearly remember the
>>>> 9+3 character filename limitation in my earliest memories of VMS.
>>>
>>> Before my time (I started with 4.4) but I thought the 9+3 to
>>> 39+39 change happened with VMS 3.x -> 4.0 not with ODS-1 to ODS-2?
>>
>> My earliest VMS was 3.6 or 3.7. Which seems to match very well with a
>> memory that it changed with V4.
>>
>> But was it ODS-2 already at VMS V3? I guess it might have been, but I
>> thought the filename format, as well as directory structure hadn't
>> changed in ODS-2. But if you are right, then the ODS-2 spec must have
>> changed?
>
> My understanding (but totally based on hearsay as it
> is before my time) is that VMS had ODS-2 from day one
> (but also supported ODS-1 for PDP-11 compatibility).
VMS on VAXen still supported ODS-1 right up to the last release (V7.3).
I was sortof expecting that VMS used ODS-2 from day one. I read up a bit
on ODS-2 now, and it always allowed up to 255 characters in a filename,
in the directory structure, it seems. (And the file header actually only
keeps 40 characters, which is shared by name, extension and version number).
So I wonder if the 9+3 was just a parsing limitation, which was there
because of various tools still being RSX binaries running in
compatibility mode, and those would have the 9+3 limit.
Johnny
More information about the Info-vax
mailing list