On Mar 10, 8:05*am, Juha Nieminen <nos...@thanks.invalid> wrote:
> Pete Becker <p...@versatilecoding.com> wrote:
> > On 2011-03-09 17:23:46 -0500, red floyd said:
>
> >> On Mar 9, 1:23 pm, Paul N <gw7...@aol.com> wrote:
>
> >>> And they'd probably suggest changing
> >>> <iostream.h> to <iostream>, which will mean putting a "std" somewhere
> >>> in your code. It's up to you how much of this you go along with.
>
> >> So you recommend using non-standard headers? *Even MSVC doesn't have
> >> <iostream.h> any more.
>
> > Um, no, the suggestion is just the opposite. The original code used
> > <iostream.h>, and the suggestion was to change that to <iostream>.
>
> * Read it again. It's not suggesting it, it's saying that "people
> will suggest it, it's up to you to decide". Writing non-standard code
> is not a good idea, so the "decision" is a no-brainer.
Perhaps I should have stuck to just answering the question, which the
OP seemed perfectly happy with...
I don't know the OP's circumstances. Possibly he is using an old
compiler which (like the older of my two compilers) simply doesn't
have <iostream>. Or perhaps he would be better off concentrating on
getting his code working on his system rather than worrying about how
it will port to other systems. Or, conversely, perhaps he would indeed
be better off learning about standard code at this stage. I was just
trying to give a pointer without being too prescriptive.
Incidentally, I hardly ever use either myself; for DOS programs I
either use stdio.h or access the screen memory directly, for Windows
programs I have to use other techniques anyway. I suspect I'm not
typical in this group, just as I suspect even more strongly that my
preference for C-type string handling over C++ strings is not typical
of the group.
Hope that clears up any doubt.
Paul.
|