equal 01-rstudio-intro about r-novice-gapminder HOT 7 CLOSED

Hi @orchid00 - I know it seems crazy that things which are not equal should be called equal, but I disagree with you here that we shouldn't be encouraging people to use all.equal(). It actually is best practice to not use == in non-integer comparisons, because otherwise you risk returning a non-equality when in fact there should be one, simply because of how floating point numbers are computed/stored. There's a much better explanation of the issue here on Stack Overflow.

Unless of course I misunderstood, and you were instead recommending that we change the comparison function to identical() instead of all.equal(), in which case I do agree with you, as long as there is also information in the tip that highlights the difference between exactly equal and almost equal.

I think the real key here is to point out how people can get caught up in the distinction between these different ways of testing for something that seems like it should otherwise be completely straightforward.

from r-novice-gapminder.

Maybe this point should be moved to a later discussion on testing your
software?

On 29 September 2015 at 02:54, Naupaka Zimmerman [email protected]
wrote:

Hi @orchid00 https://github.com/orchid00 - I know it seems crazy that
things which are not equal should be called equal, but I disagree with you
here that we shouldn't be encouraging people to use all.equal(). It
actually is best practice to not use == in non-integer comparisons,
because otherwise you risk returning a non-equality when in fact there
should be one, simply because of how floating point numbers are
computed/stored. There's a much better explanation of the issue here on
Stack Overflow
http://stackoverflow.com/questions/9508518/why-are-these-numbers-not-equal
.

Unless of course I misunderstood, and you were instead recommending that
we change the comparison function to identical() instead of all.equal(),
in which case I do agree with you, as long as there is also information in
the tip that highlights the difference between exactly equal and almost
equal.

I think the real key here is to point out how people can get caught up in
the distinction between these different ways of testing for something that
seems like it should otherwise be completely straightforward.

—
Reply to this email directly or view it on GitHub
#65 (comment)
.

from r-novice-gapminder.

Maybe the best thing will be to add in the tip use isTRUE(all.equal(....)) and also recommend people to read ?all.equal and ?identical

from r-novice-gapminder.

Hi @orchid00 - want to give improving it a shot in a PR?

from r-novice-gapminder.

@orchid00 interested in pushing this one forward?

from r-novice-gapminder.

If not probably should close - it has been a while.

from r-novice-gapminder.

Closed because of no activity in over a year

from r-novice-gapminder.