Bruce Lane <kyrrin2@wizards.net> wrote to me instead of port-vax:
> ?!!? A -SCSI- RRD40? Pardon my skepticism, but I've only seen them in the
> proprietary DEC interface with the 15-pin connector. You're sure about
> that?
   Absolutely sure. It's an external box with 2 SCSI-1 connectors and SCSI
ID switches on the back, and I have seen several manuals for SCSI-only
machines detailing how to attach an RRD40 to the SCSI bus ("Take one end of
the cable and insert it here, take the other end and insert it there...").
Seriously, though, I'm quite confused by all these RRDs. This one is in an
external box. I haven't opened it up yet, but I have peeped into the
ventilation holes. What I saw there seemed like a half-height device with
some circuit board ON TOP of it (NOT part of it). It even had some chips
that look like EPROMs. Seems like the situation with TK50Z (the drive
itself is utterly non-SCSI but it is connected to a special adapter board
which is in turn connected to the SCSI bus). I have another disassembled
box (MV2000) on my desk right now, so I don't have enough space to tackle
the RRD40, but I'll probably put the MV2000 back together on Tuesday, and
then I will be able to tell you for sure what the RRD40 box really consists
of.
   In general the situation will all these RRDs is rather strange. The
strangest thing is the way in which DEC has assigned model numbers to them.
Normally DEC model numbers for drives consist of two letters and two
digits. The first letter should be R for disks and T for tapes. Sometimes
they use two letters to specify the device type, like using "TL" instead of
just "T" to indicate a particular kind of tapes. CD-ROMs are disks, so they
should be denoted by "R", but they are special in many ways (read-only, for
instance), so using "RR" instead of just "R" to indicate a read-only disk
is justifiable. But the second or last letter normally indicates the
interface. The ST506/412 interface is denoted by "D", the 26-pin TK50/TK70
interface by "K", and SCSI is almost universally denoted by "Z". If one
were to accept the theory that there was an utterly non-SCSI CD-ROM drive
and then the SCSI RRD40 was made out of it by adding an adapter board, the
name becomes more justified. In that case it's just like TK50Z. It's called
TK50Z, rather than TZ50, because it's not a native SCSI drive but a drive
with the 26-pin interface that has been artificially SCSIfied. In the case
of RRD40 they have probably forgotten to add the "Z" at the end.
   This theory can be confirmed by finding a drive that's just like RRD40
but utterly non-SCSI. Bruce describes some drive that connects to a Q-bus
system through a special MSCP controller. Actually, I remember reading
something like that in the mass storage options section of the MicroVAX II
630QE manual. However, it was called RRD50 rather than RRD40. This raises
another interesting question. It seems to me that the digit following the
interface designation in the model number denotes the form-factor.
Apparently, 8 denotes 19" rackmount, 5 denotes 5.25" full-height
horizontal, 7 denotes 5.25" full-height vertical, and 3 denotes 5.25" half-
height. But I have never seen a form-factor designation of "4" anywhere
except RRD40 and RRD42. RRD42 is an absolutely plain and generic 5.25"
half-height native SCSI CD-ROM drive, and the actual drive inside RRD40
also seems to be half-height, judging from what I could see through the
ventilation holes. Also the manual for VS3100 M40 says that RRD40 can be
internal and depicts its front panel as half-height. Of course, the M40
enclosure has plenty of vertical space, and if the RRD40 requires an
adapter board on top of the actual drive there's enough space for it. I
really see no explanation for the form-factor designation of "4". On the
other hand, if the original version that connected to a special Q-bus card
without involving any SCSI was really called RRD50, that's more
understandable. That was in the manual for MV II 630QE, and that's probably
what it has been designed for. 630QE consists of two rackmount BA23 boxes,
which are designed for full-height horizontally-mounted 5.25" drives only
(they are screwed on the bottom to plastic mounting plates which then slide
in and lock, and this method clearly makes no provision for vertical
stacking of half-height devices), and these happen to have a form-factor
designation of "5". Even if a drive is half-height it will occupy a full-
height bay, so it certainly made some sense to use a form-factor
designation of "5".
   Soon I'll take a closer look at my RRD40, and also Richard should find
the docs for all of these drives for me, so soon I should be able to tell
you more.
> >[...] I don't think that I would ever need to read CDs on a
> >UNIX box, but in case I do I will have this external RRD40 spare.
>
> <g> No need to cringe. Most of my data preservation and archiving
> activities take place on PC platforms.
   "Data preservation and archiving" is done on tapes. My main tape drives
will be external, so that I can attach them both to my IBM PC AT-compatible
running DOS at home and to my UNIX boxes at work. As for CD-ROMs, IMHO
there is no place for them at all in the world of free software and
individuals doing favors to each other. They are utterly undemocratic, as
they emphasize reading over writing and are designed specifically for mass
duplication of commercial packages. I certainly don't need to keep a CD-ROM
drive on Harhan, and even at home I don't use it very frequently. I use DOS
and small gadgets, written by hackers and mostly free, rather than Windows
and big commercial applications, so I don't need CD-ROMs for most of my
work. However, there are cases when I need to look at some DOS or Windows
CD-ROM, and therefore I have to keep a CD-ROM drive at home.
> >[...] The standard document sets for such machines normally
> >consist of a three-ring binder with several documents inside it, and
> >there is a part number for each document as well as for the whole
> >binder. I wonder whether the AZ-... numbers that you have mentioned are
> >for separate docs or for whole binders. If these docs have this
> >structure, could you tell me all the numbers?
>
> Separate docs, actually. Unfortunately, I don't have the overall
> structure numbers.
   Oh yes, right. That was my error. The structure with part numbers both
for each document and for the whole binder is used in MV2000 and VS2000
manuals, but not in the MicroVAX II ones. The latter don't have a number
for the whole binder. The ones that I have are for 630QB and 630QE, and I
don't have them in front of me (I'm at home and they are at work), but I
think the numbers for individual docs in that case do start with "AZ"... I
have screwed up this time.
   David Brownlee <abs@netbsd.org> wrote:
> Having a CDROM drive on a unix box can be a particular boon for
> users - both on a workstation, where a user can swap in disks as
> & when they want to, and on a server, where its a cheap way to put
> up large source code distributions & such like :)
   The workstation considerations don't apply here, as the machines will be
in a locked machine room and will be accessible to users only via the
network (that's why I prefer to call them "login machines" or "satellites",
rather than "workstations"). As for putting up source distributions and the
like, there is a problem with your approach too. I don't want to burn a new
CD-ROM each time I need to make a change. Even if it's something static
like frozen releases, one may still need to add something to the archive or
reorganize the directory structure. Plus I don't have a CD writer and don't
want to waste money on buying one.
   
   Sincerely,
   Michael Sokolov
   Phone: 440-449-0299
   ARPA Internet SMTP mail: sokolov@alpha.ces.cwru.edu