Re: [boost] Sockets vs. net connection

From: Beman Dawes (beman_at_[hidden])
Date: 2000-12-28 18:04:45

At 05:20 PM 12/28/2000 -0500, Ronald Garcia wrote:

 >>>>>> "GS" == Gregory Seidman <gseidman_at_[hidden]> writes:
 >
 >    GS> The issue seems to be one of naming rather than
 >    GS> functionality. What is a "socket" to you? A socket in the Unix
 >
 >I would agree that "socket" can be considered to refer to multiple
 >concepts, but I think it's fair and in fact beneficial to use the term
 >"socket".  From my personal experience, I don't necessarily think of
 >sockets as the specific interface provided by unix, but instead a
 >generic term for network streaming capability.  In the end, I think
 >that a C++ interface for network communication would be grand. Any
 >time I look for such  an interface I use the keyword socket.  Is the
 >network API under win or mac given a different name such that it would
 >not be obvious to a windows or mac programmer what is being referred to 
by
 >a "socket" library that runs on multiple OSes?

Windows uses "Windows Sockets", often called "winsock".  See 
http://www.sockets.com/winsock.htm and http://www.sockets.com/winsock2.htm

Here is a quote from the first of those: "The Windows Sockets Specification 
has been built upon the Berkeley Sockets programming model which is the de 
facto standard for TCP/IP networking. It is intended to provide a high 
degree of familiarity for programmers who are used to programming with 
sockets in UNIX and other environments, and to simplify the task of porting 
existing sockets-based source code. The Windows Sockets API is consistent 
with release 4.3 of the Berkeley Software Distribution (4.3BSD)."

--Beman