From: Geoffrey P Waigh
To: Multiple Recipients of
Subject: RE: Euro - re-statement
Date: Friday, October 17, 1997 12:10AM
On Thu, 16 Oct 1997, Carrasco Benitez Manuel wrote:
> > As a non-European, you can take this with a grain of salt. However:
> > is any self-respecting European IT application really using seven
> > bits?
> > [Carrasco Benitez Manuel]
> > You will be surprised.
> >
There must be some kind of ancient computer preservative in Europe.
How are all the printers and terminals and programs going to be altered
to
support the Euro on 7-bit systems? Is it actually cost-effective to
maintain 20 year old equipment and have these beasts refitted with new
ROMs (good luck finding compatible chips) rather than replace them with
something more modern? Cast new chains for the line printers? All the
financial software is going to be hacked to support the new currency if
not dual currencies during the transition period, but it cannot be made
8-bit clean? People are going to convert all their 7-bit datasets to
another 7-bit encoding rather than go to an 8 or 16 bit encoding that
will
reduce the "we cannot represent that on our system" headaches they have
had for eons? IT staffs are not saying "We have a year 2000 problem, we
need to add Euro support and we cannot store the names and addresses of
our customers correctly, (+ who knows what else,) maybe its time to
phase
out our legacy system?"
A few years ago a client site had to upgrade their computer to
accomodate
the interface to our software and it cost them on the order of 40000 USD
for maybe 16K of memory (I don't know if that was bytes or words.)
After
a fair bit of downtime for the installation, they now had a system that
my
brother's pocket calculator could outperform in terms of price, memory,
computing power, size, power consumption, ease of programming,
everything
except I/O since his calculator could only support one terminal.
At some point you have to bite the bullet and overhaul your systems
rather
than demanding the rest of the world hold back.
Geoffrey
************************************************************************
**************
Dear Tomas and Geoffrey,
Although I don't infer the same holds for Europe, many organizations in
the US are
facing the same year 2000 problem by replacing hardware rather than
attempting to
fix all the software. For example, Tuesday the City of Chicago announced
it would spend $ 15m to replace all their computers and software rather
than invest $60m
in software fixes and upgrades to fix the 'Year 2000' and related
problems.. plus
the new hardware runs a lot faster and has much better reliability.
On computer industry critic has observed that the 'Year 2000' problem
may be the
best thing to happen for computer hardware sales since the invention of
the
microprocessor. And we don't even have the Euro problem...
If anything, the 'Year 2000' problem is forcing organizations to
reexamine
Information Technology STRATEGY AND POLICY, something most refuse
to do on a routine basis. They are getting some rude surprises.
As far modifying 7-bit ASCII, I think that you can safely forget it.
Note that there are
several characters which already have ISO-sanctioned substitutions. But
ISO 646
is unlikely to be reopened.
Microsoft's choice of 80H for the Euro in all Windows code pages 125x
(except
Cyrillic, where it is 88H) is rather important. Note it also picks an
unused position
in the Japanese Shift JIS table (JIS X208-1997; Annex 1) which meaning
that we
won't have a conflict between Western and Japanese Windows.
Alain's Latin 0 is the most important alternate proposal and has my
vote.
But until we have formal agreement on Latin 0, I think that you can
safely
bet that 80H is now the defacto standard for the Euro. After all, how
many
copies of Windows do you think exist in Europe? And what is the cost of
a
Windows NT upgrade compared to an O/S upgrade to a legacy system?
My contention is that Windows NT is much more likely to be around on
January 1, 2000 than a lot of the 7-bit legacy systems. Economics
usually wins over politics.
Clive :-)}
This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:37 EDT