Date: Fri, 21 Feb 2020 08:59:21 +0100
On 2020-02-21 at 00:29, Shengdun Wang via Std-Proposals wrote:
> Recently I did a simple survey on google forms. My statistics show 48%
> of C++ projects banned iostream in part or whole of their project. The
> statistics look very similar to Herb Sutter’s Survey on C++ exceptions.
> Around half of C++ projects banned iostream.
>
IMO the survey result leaves much open to interpretation.
For example, "Allowed in some parts of code" could mean that all I/O is
performed in a separate part of the project, and disallowed in others.
Not that it is not used at all.
If so, iostreams might be used in 75% of all projects. Of the others,
some might be dialog based and just don't do stream I/O. Some might be
server apps using a database backend, and use no files at all.
Also, a survey with 98 responses might not be representative for the C++
community in general. :-)
Bo Persson
> Recently I did a simple survey on google forms. My statistics show 48%
> of C++ projects banned iostream in part or whole of their project. The
> statistics look very similar to Herb Sutter’s Survey on C++ exceptions.
> Around half of C++ projects banned iostream.
>
IMO the survey result leaves much open to interpretation.
For example, "Allowed in some parts of code" could mean that all I/O is
performed in a separate part of the project, and disallowed in others.
Not that it is not used at all.
If so, iostreams might be used in 75% of all projects. Of the others,
some might be dialog based and just don't do stream I/O. Some might be
server apps using a database backend, and use no files at all.
Also, a survey with 98 responses might not be representative for the C++
community in general. :-)
Bo Persson
Received on 2020-02-21 02:02:09