On 4/2/07, Stewart Stremler <[EMAIL PROTECTED]> wrote:
begin quoting Lan Barnes as of Mon, Apr 02, 2007 at 10:16:58AM -0700: > A generally accepted standards document. Here's how it works: > > I figure out a pretty good way to do something in computing. Maybe it's an > email protocol, maybe it's a network file system, maybe it's TCP/IP. I > write it up and publish it, asking for feedback -- a "Request For Comment" > or RFC. Someone, I'm not sure who (IEEE?) assigns it a number. We wait a IETF?
It is a generic document name, but RFCs are in fact the accepted working documents of the IETF, and most of them come from there.
period of time. People comment. Some of the comments may be useful. I then > program it and release it. I thought RFCs came *after* the code, in practice. > People adopt it, but only if it's useful. When they want to write programs ...or they use the fact that it's an RFC to cram it down the throats of others. After all, if there's an RFC, it's a "standard", and they wouldn't have a standard unless it was a good idea, right?
Actually the IETF has several classes of RFCs. Informational, which can be submitted by anybody, and are sometimes a crock (http://rfc.net/rfc2324.html) Experimental, which is much the same, and then the rest are on the standards track: Proposed Standard, Draft Standard, and Standard. There's not that many that are "Stadnard" and most live in the Proposed standard, which should be, and often is enough to qualify for industry to implement. Yes, what you state is how most people may think of it, and often how industry may misrepresent something.
or libraries that interface with my protocol, they know where to go to > find out exactly how -- to the RFC. If I think of an improvement, I can > rev it. It becomes a standard by convention and because of its usefulness > - a tech meritocracy of ideas. There are some great ones in there, too, like RFC4824. > Now a Big Soulless Corporation comes along that wants to own everything. > It takes my protocol and adds features just to add features. It doesn't > bother to ask anybody. It copyrights or even better, patents these > additions. They don't even have to be particularly useful as long as they > make the BSC's rendition of my protocol incompatable with what everyone > else is doing. Like IE as a browser or HTML in email. Interoperability is at odds with market dominance.
Unless you have no choice. Are you telling me cisco is not a member of many standards bodies? I helped write Media Resource Control Protocol and there were many Cisco reps involved, and I can't say the really contorted the process. That's a very vague non adressable statment.
If the BSC can get enough people to buy into their extension like mindless > sheep, then they can drive competitors out of business. They become the > "de facto standard." We all end up paying for what used to be free. Don't forget that the BSC can provide their extension "for free" with their existing product, and third-party replacement extensions can be tagged with "nonstandard" and "non-free(-as-in-beer)" and thus crushed. > Pretty cool, huh? You let other people do the _real_ innovation and then > you "fence off the commons" and claim it as your own. Mucho dinero para > nada. > > Is this a great country or what! Heh.
Also, the IETF requires IP disclosure, so you'd at least know what you're getting in to, and have grounds for a suit if they didn't disclose their IP. Sometimes people even bring up unconcerned external IP to prevent similar messes. If you've never been involved in the IETF I hesitate for anyone to claim that they could co-opt the process to their wishes. If you aren't interested in the subject matter, all that will happen is people will call your bluff openly on the list, and people will know about it. That's no so bad is it? On the other hand if you want to blame companies for getting as close to the lines as possible, that's fine. I like to know that the companies I invest my 401K in are not worrying about politics, but the system as it is. -Tom -- [email protected] http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list
