In article 47b73fd7@qaanaaq,
Andy Hall writes:
On 2008-02-16 18:56:49 +0000, (Andrew
Gabriel) said:
How strange -- I designed X.25 switches at GEC in the 1980's,
and had to get those tested for PSS approvals in early days,
and the line modules had to get electrical approvals. One I
remember was our G.704 module which we designed to provide a
raw unstructured 2Mbit X.25 link over BT's Megastream service.
Went along to Baynard House, and the bloke took it and looked
at it, and said "that's fine". It was the first G.704 module
they'd seen, and they didn't have any tests defined for it!

Well the problem was that this was a product that had originated in the U.S.

For some reason, best known to themselves, BT had an addressing scheme
(group number IIRC) that was 4, while in every other country it was 0.
On this particular box the addressing wasn't configurable, simply
because there was no need anywhere other than in the UK.

Not too sure what you mean. PSS always required the DNIC to
be present (2342), whereas most other networks only required
the DNIC for calls to other networks. It was a bit like always
requiring the full international number to be dialled for a
phone call, even a local one. Actually, I thought this made a
lot of sense. Generally there's no one typing in X.121 addresses
for each connection (excluding dial-up PADs), so having short
forms of addresses was just overhead in software to recognise
(a bit of software I wrote a number of times in different
products;-). Since we sold in other countries too, we had to
cope with this though.

For the LCGN, PSS supported all 4 ranges if you subscribed to
them, although BT stopped selling new PVCs around 1985 because
their switches ran out of table space to record them. I can't
recall if other country implementations supported multiple LCGN
ranges (most didn't support PVCs, which would have removed one
of the LCGN ranges in any case).

Anyway, the BT guy agreed to sign it off on the basis that I agreed
that we would have a software fix to allow configurability before any
were connected to their network. I suspect that he figured that if
this wasn't there, it wouldn't work with their PSS environment anyway.
All of the other HDLC and other tests had passed and they seemed more
concerned about that for some reason. There was no interest in the
hardware at all.

Imagine paying by the packet (well segment)..... but one did.

Yes.

--
Andrew Gabriel
[email address is not usable -- followup in the newsgroup]