Re: Confusing language, was Re: Help help.. please.i am totaly new in ada programing

[Robert Dewar <dewar@gnat.com>]

In article <h3LU3.261$Ye5.10718@typhoon-sf.snfc21.pbi.net>,
  tmoran@bix.com wrote:
> Would there likely be more or
> fewer competitors if it were GPLed?


Well for sure there could not be fewer competitors with respect
to the particular problem at hand, which is fixing this
particular program for Y2K.

There are indeed a number of cases in which Y2K has been used
as a lever to get people to upgrade to new versions. There are
even cases in the Ada vendor community where people have been
forced to pay expensive upgrades to get Y2K certification for
Ada compilers, because the vendor refuses to certify an old
compiler as Y2K compliant. Now that *might* be legitimate if
there really were Y2K problems in the old compiler, and it
might be quite bogus if, as one suspects, compilers are not
likely to have Y2K problems in the first place. But if the
sources are closed, of course no one knows what is involved
in doing the fix.

In the larger software world, it is for sure the case that
there are instances of companies doing trivial Y2K fixes,
insisting that these are not just bug fixes, and charging
excessively large fees for the new versions, knowing that
their customers are over a Y2K barrel :-)

GNAT by the way is an example of a program that really does not
particularly use dates, so Y2K compliance has never been an
issue. We did change the format of file time stamps to have a
full 4-digit date, not because this is needed for any useful
purpose, but it only wastes a small amount of space, and we
got tired of having to explain to people why, in this particular
case, 2 digits dates were fine (*)

Robert Dewar
Ada Core Technologies

(*) FOr those interested, the "bug" with 2-digit dates was
that if in the year 2090 for instance, you accidentally try
to bind with a 100 year old object file, the binder might not
notice that the program was inconsistent. Yes, it could happen,
but we guess that the probability of this being a real problem
is somewhat less than say a malicious hacker going in and
deliberately changing dates :-) So now, you only have to worry
about binding with a 10,000 year object file (you know, people
are not really worrying enough about the Y10K problem :-)

P.P.S.
Tom Moran's basic point: that you cannot tell if this is a rip
off is quite correct. There are actually two kinds of Y2K
ripoff. The kind I mention above, where the change is trivial,
but you can overcharge because the user is between a rock and
a hard place.

The second kind is the insistence that Y2K is not a bug at all,
but some kind of mysterious act-of-god that could not have been
forseen by the vendor, and which therefore justified charging
for a new version, even in a situation where bugs are supposed
to be fixed at no charge.

But without the sources, you definitely do not know that the
charge in this case was in fact excessive, for all you know,
it is a bargain, and the real situation is that the vendor
has to put in a huge amount of work, far more than $8000 worth,
but figures the market will NOT let them charge for the full
work.

There is nothing magic about open source software or free
software that makes tough software problems disappear, or
suddenly cost nothing to fix!




Sent via Deja.com http://www.deja.com/
Before you buy.