Re: Do we need "Mission-Critical" software? Was: What to Do?

[Robert Dewar <robert_dewar@my-deja.com>]

In article <93eoku$cm2$1@nnrp1.deja.com>,
  n_brunot@my-deja.com wrote:
> I'm confident in Java to correct that, if users find it
> necessary, without listening to you even if you claim it's
> impossible :-)

Well it is certainly impossible with the current specification.
No one would disagree with this (it sounds like you do not even
understand the technical issue here, so I assume you are not
really disagreeing either). As to whether the Java spec can be
fixed, not at all clear. So far, the proposals to fix this have
been rejected. If you have a clever idea of how to fix this,
please speak up, otherwise your faith that this can be fixed is
rather unconvincing.

Of course Java runs fine on Suns, and perhaps in the long run
since it is becoming increasingly irrelevant on Windows
platforms, it really does not matter that it does not run
accurately on Windows.


> But, it was litteral translation from french, where (to my
knowledge,
> may be I'm wrong) it means that when you find a very small
number of
> exceptions to a rule, widely true otherwise, you'd better
trust the rule
> is most cases, rather than thinking you met an exception each
time rule
> goes against your convinctions.


Investigate:-) In English, most people also think, quite
wrongly, that this is what the saying means, because the use of
prove in the sense of test is obsolete.

After all an exception generally DISPROVES a rule, that is
what scientific progress is all about. One violation of a
theory in experiment is enough to reject the theory, or at
least require its refinement.


Sent via Deja.com
http://www.deja.com/