|
Authored by: jesse on Friday, May 31 2013 @ 09:55 AM EDT |
Most of them provide the access at the system call layer.
Only when the system does NOT provide access is it done in a userspace
library... and when that happens, all bets are off as to whether it will
function properly or not.
As with the winsock problem - there is too much shared data necessary for
multiple applications to use the BSD sockets through libraries - and expect
things to work.
When only one application uses the network, with only one connection, then it
can be fairly reliable. More than one though, the thing tends to fail.[ Reply to This | Parent | # ]
|
|
Authored by: Anonymous on Friday, May 31 2013 @ 02:46 PM EDT |
If the BSD Socket API can be implemented without using any BSD code, it can
*also* be implemented without using any *other* code from any *other* project.
You keep talking about how APIs can be copyrighted somehow because the *source
code* which implements the API can be copyrighted. That's a complete
non-sequitor, akin to claiming that light can be patented because a lightbulb
can be patented. Or that iambic-pentameter can be copyrighted because a poem
written in that form can be copyrighted.[ Reply to This | Parent | # ]
|
|
Authored by: jjs on Saturday, June 01 2013 @ 04:53 PM EDT |
Sorry for shouting, but you just don't seem to want to
listen. I'll give you an example - TCP/IP. The API is
specified in IETF RFCs. One of the IETF rules is that no
RFC can become a standard without at least two INDEPENDENT
implementations. That means people writing to the API with
completely independent source code, done in a completely
independent manner, with no communications between them.
Yet they have to be compatible.
So no, APIs are not source code, and the copyright on the
BSD source code has nothing to do with the ability to
copyright the API.
---
(Note IANAL, I don't play one on TV, etc, consult a practicing attorney, etc,
etc)
[ Reply to This | Parent | # ]
|
|
|
|
|