RE: [boost] Formal Review Request: iostate.zip

From: Fernando Cacciola (fcacciola_at_[hidden])
Date: 2002-02-01 13:41:21

----- Original Message -----
From: Douglas Gregor <gregod_at_[hidden]>
To: <boost_at_[hidden]>
Sent: Friday, February 01, 2002 3:11 PM
Subject: Re: [boost] Formal Review Request: iostate.zip

> On Friday 01 February 2002 10:45 am, you wrote:
> > Without even looking at the library, I've really got a problem with
basing
> > design decisions on a compiler that is way out-of-date.
> [snip]
> > What do others think?
>
> I don't think we should be basing design decisions on any compiler,
period. I
> too would vote against an interface that represents a compromise due to
> current compiler technology.
>
> Doug
>
I agree.
I think that we should base design decisions on the current language
specification. Period.

The only exception should be those language features that are known to have
really poor support from most of the principal compilers,
like the __export keyword, or features that are still under consideration by
the committees (any open active issue)

In fact, the features that we better not use (in the interface) should be
identified and set aside explicitly in a guideline. Every other aspect of
the language not explicitly labeled as 'avoidable' should be used if it
leads to a better design.

I know that this requires boost developers to be up to date not only in
theory but also in tools (compilers), but I think that otherwise we would be
holding back the development of comforming compilers and knowing developers.

Anyway, as Gustavo Guerra just point out to be the case here,
compiler-specific constructs are free to be used in the implementation of
any boost library.

Fernando Cacciola
Sierra s.r.l.
fcacciola_at_[hidden]
www.gosierra.com