Re: 128 bit integer software
On 7/21/2014 1:04 PM, jacob navia wrote:
Le 21/07/2014 17:28, Victor Bazarov a ?crit :
You mean, you wrote a library and didn't write any test driver for it?
What criteria do you use to judge that it's ready to be tested? That it
compiles?
This is a good sample of someone trying to build a polemic from nothing
at all.
"Polemic"? You wrote "I need to test my 128 bit integer library."
Nothing about what it does, nothing about what its purpose might be,
nothing about what language it's written in, etc. Not even a link to
the web page that has those things, a description, a spec, maybe an
interface... I am simply trying to find out more about your "library".
There is no polemic.
"... you wrote a library and didn't write any test driver for it"
First off, you decided to omit *my* question mark after this. Good one.
Mr Bazarov doesn't know anything.
For somebody who doesn't know anything Mr Bazarov has had a rather long
career... Could be a coincidence, of course. There have been examples
of people having a long career without knowing squat, but it's more an
exception than the rule, I think.
> He never saw my test driver nor my
tests, but he confidently writes nonsense since he is the "guru" of this
group... or one of them.
*I* write nonsense? And what do you call "Has anyone any software I
could compile to test it?" Your software actually can be *tested* by
*compiling* some other software? Is your library a _compiler_? If so,
you forgot to mention it.
Besides, if you have your test driver and your tests, why are you so
defensive? Could you not simply say that you have tested your library
with a bunch of tests already (if that's true, you shouldn't be ashamed
to say so) and you are now ready for a more thorough testing, akin to
alpha testing usually done in-house, by trying to adapt others' code to
use your library instead of <somebody else's>?
You can never write a good test for a software yourself since
unconsiously you will avoid the parts of the software that you know are
weak.
You know nothing about me (aside from my contribution to this forum),
yet you assert that *I* can never write a good test. You presume too
much, Jacob.
> Besides, you have a limited imagination, other people writing
software will come with completely different tests!
Presuming you actually meant to write: "_I_ can never write a good test
for a software _myself_ ...", the only logical conclusion (albeit a bit
exaggerating) would be that, knowing that *you* can't write a good test,
*you* are not going to even try. Right?
Then he goes on in the same vein:
..." What criteria do you use to judge that it's ready to be tested?
That it compiles?"
Obviously it would be difficult to test it if it doesn't compile isn't it?
And yes, the main criteria for begining the tests is that it compiles.
Is that the only criterion?
After it compiles you can start testing it!
.... with other people's software? Seriously?
V
--
I do not respond to top-posted replies, please don't ask