Re: comedi as a POSIX standard

Frank Mori Hess wrote:

> This [POSIX standardization] seems premature to me.  

It may be. I'm more interested in the general question of whether it 
would be a good thing. If it makes sense to wait, that's OK.

> Also, there doesn't seem to be any pressing need since there
> aren't any competing, 'almost compatible' implementations I'm aware of.

The purpose of standards isn't to arbitrate between competing 
implementations. It's more to declare a certain class of problem to be 
adequately addressed by a certain approach, where adequate really means 
agreeable to a reasonably large cross section of interested parties.

The POSIX pthreads standard is a good example. Pthreads is by no means 
the only way to achieve multithreading, but the fact that it's a 
standard means that platform vendors are more willing to implement it 
than they would be otherwise. If you want to write portable 
multithreaded programs, this is a Good Thing.

The main benefits I can imagine from a POSIX Comedi standard are:

1. Implementation on more platforms, including non-Linux OSes, embedded 
systems, etc. My own open source application runs on pretty much 
anything Unix-like. If the Comedi API were supported on OS X, FreeBSD, 
Solaris, etc., the data acquisition functions would too. (They're 
currently disabled on non-Linux platforms.) Portability is good.

2. Support from a/d and d/a board vendors. It's easier to justify the 
cost of developing a driver if you can make some marketing claims to go 
with it. And a stable, well-documented API is a much lower risk for the 
vendor.

Steve

Received on 2003-04-13Z19:42:34