[
Lists Home |
Date Index |
Thread Index
]
The same for services. It is the question I ask repeatedly:
what is the culpability for services created on top of services,
for example, using Google addresses and a separate geocoding
service?
I'd think long and hard before building an application over
unvetted and non-indemnified web services. That this is also
true of standards and specifications built by polite aggregation
and published prior to serious implementation goes without saying.
That is why committee work without the participation of the major
vendors in any market space can be hazardous. The days of 'let's
form an OASIS committee, get two buy-ins and start writing a
schema' produced monsters which no one in industry wants to
adopt and yet work their way into requirements without any
serious chance they will be adopted by more than a few
prototyping organizations or small wanna-be companies.
There is nothing new about this. Some of us are veterans from
earlier efforts. The difference is that there was no common
architectural platform such as the web in which to spread the
damage at light speed.
len
From: Ken North [mailto:kennorth@sbcglobal.net]
Michael Kay wrote:
> I've always thought that it's likely that in most incidents of people
being
> killed as a result of software bugs (or IT systems bugs), the software
> wasn't thought to be safety-critical at all.
>
> And as I said before, if your programming is negligent and it kills
someone,
> disclaimers are very unlikely to protect you.
This thread is about software licenses but it seems there's a similar issue
with
XML vocabularies, DTDs and schemas.
What if a flawed schema design results in an accident? For example, coding
incorrect constraints for allowable temperatures or pressures might cause an
instrument or a monitoring program to work from unsafe data.
|
