Re: Luxasm news

Johannes Kroll <j-krollNO-SPAM-BLAH-BLAH@xxxxxx> écrivait news:d5b5tp$ald
$1@xxxxxxxxxxxxxxxxxx:

> On Thu, 05 May 2005 16:44:05 +0200
> "\\\\o//annabee" <http://www.TheWannabee.Org> wrote:
>
> I'm not sure why you post this here. What I meant was that software
> generally isn't "bullet proof", i. e. programs can still crash no
matter
> how careful you design/implement them. So a crash handler that tries to
> save a backup of the current document is not a bad idea.


The fact is that we do effectively have a "crash handler",
in RosAsm, that saves the user's Source before the shut-down.

But i was not much in favour of this implementation: The best
way to destroy a Source is with the hands of the Programmer.
So, better is not re-using a Source saved at the hang time,
in a questionable hope of saving some much probably defective
typing, that will very unlikely be longer than a couple of
Lines.

Considering the way a Programmer writes, when he is using
a Tool that can re-compile his app in the click time, and
considering the reasons why a Programming Tool might hang,
once made "pretty stable"..., in practice, in between the
writing time, and the end of the Compilation, there is no
reason, at all, to save anything. All the contrary: Saving
in between could put the source in danger, as the reason of
the hang is much probably inside the Source itself - even
if this shows a real bug in the Programming Tool -, locking
the developement into a vicious circle, until the programming
Tool's bug could be fixed.

A pretty good example where paranoïa could create problems
that do not exist, without paranoïa. Think of this, Troll,
you are on the right track, and try to convince Beth of
taking her medication, instead of believing Randall Hyde
bullshits.

:)

Betov.

< http://rosasm.org >



.

Quantcast