Autoconfig, Hardware Manufacturer Number?

[olsen]

Is there any management of the Hardware Manufacturer Numbers used by Autoconfig that were assigned by the Commodore Amiga Technical Support in West Chester, Pennsylvania in the past?
(as mentioned here googlecode.com: AUTOCONF.TXT)

Please have a look at the hardware manufacturer ID page on the AmigaOS wiki. Although I can take care of assigning IDs if you need it, the procedure outlined on the page is the official one now.

The page has been recently updated and polished, and it also contains a list of IDs that are on record.

Please note that the list of IDs on the page is probably incomplete. Which means that you can't just grab one from the historic number ranges (say, 1999) and expect it to work. While the risk may be low that there will be collision, why take chances?

[olsen]

Incomplete, yes. For instance missing Petsoff, makers of Delfina and DblScan.

Which ID would that be?

[olsen]

Re amiga.resource.cx:
Delfina (Classic)  Petsoff, Finland   Date 1996     Autoconfig ID 14501 / 0
Delfina Lite  Petsoff Date 1997  Autoconfig ID 14501 / 1
Delfina Plus Petsoff Date 2001    Autoconfig ID 14501 / 2

My, that's a weird one. To the best of my knowledge, this ID was not assigned by Commodore, Amiga Technologies GmbH or anybody else who assumed responsibility for registering IDs since then. Commodore would have picked a number from the 2000 or 4000 range, and Amiga Technologies GmbH, etc. would have picked one from the 5000 range.

Anyway, I'll make a note to record the ID.

[olsen]

Is this any help?

http://aminet.net/package/util/libs/BoardsLib

No, unfortunately not. There is no list of manufacturer IDs included, it's "binary only".

[olsen]

Disassemble it?

Or make UAE trigger various IDs and grab them from the utility?

Since providing a complete list of assigned and appropriated manufacturer ID is probably not even possible, I guess this might be a bit more work than is worth it...

Besides, sending an e-mail to the author of the library and asking him for a copy of his ID table might be just as effective

[olsen]

I think this is not so much about where to get data, as to who makes the final call on ID assignment.

It's simple enough: if somebody asks for an ID and provides the information which is filed along with the ID, he will receive a new manufacturer ID. There is not much call for new IDs, and so far nobody has tried any funny business.

I guess there has not been any ID-collision yet?
(given that SKIstor 2000 (1989) has the autoconfig ID 32768 and netbsd sources hint that it's 16-bit unsigned, collision may not be that unlikely)

It's hard to say if there ever was a problem with ID collisions. From what I know, most of the Amiga hardware that ever was and needed a manufacturer ID was produced when Commodore was still in business. Fewer than 300 developers requested manufacturer IDs, and from the different fields they were working in, it would have been a real coincidence if any Amiga would contain even two different expansion boards by different developers sharing the same ID.

How many new devices is produced per year btw?

A year with more than one new manufacturer IDs to be assigned would be the rare exception. In the last 15 years only 11 new ID have been assigned.

[olsen]

Which ones are the latest additions?

Funny that you should ask  This is not a particularly exciting subject. The latest additions are the numbers 5014 through 5017 (assigned in the last two years). They were assigned to developers who are making smaller batches of hardware expansions, such as CPU cards, memory expansions, I/O expansions (USB/Ethernet) and even virtual hardware for new Amigas (if I understand the product description info correctly).

Seems at least like the ID space won't run out like IPv4 will

At this rate, we'll have run out in 50,000 years or so. Then it's somebody else's problem