If you feel so strongly about changes needing to be made, then why not make an official comment to the RFC and try to make things better? An RFC is, by the way, a Request For Clarification.
-----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Peter T. Breuer Sent: Saturday, December 04, 2004 9:17 PM To: [EMAIL PROTECTED] Subject: Re: How to add a field to the reply that contains data from the request? "Also sprach Michael Griego:" > All I have to say is that your attitude normally determines the response > you get. You came in here telling many people who have worked with > RADIUS for a long time how the specs are wrong and how you are much No I haven't. I'm sure radius is fine. OTOH I'm quite sure the rfc is probably a load of badly written rubbish, because they normally are. So? Is there something new? Have you read a rfc lately? I certainly haven't! ;) As to telling people? I am giving them the benefit of my judgement and appraisal. I've pointed out the things that are wrong. Now you get to act on it. That's cool. Go with it. > better than they. This is a fatally flawed approach when trying to > learn something. I'm not trying to learn anything! I would hope I never do. I hate learning anything. I avoid learning anything at all! I do it far too easily, thank you. I haveto work hard to avoid it. That's the point - I am not YOUR slave. I don't do what YOU want me to do. If you want ME to use YOUR tool then YOU must persuade ME to. And you do that by doing what *I* want, not the other way round. Got it? It's a market. You've got competition. I could have used gnu-radius. I could have used others. There has to be something about what YOU have that persuades me to use yours. As it is, I think it's a fine implementation. At least the config is good and comprehensible. Beats gnu-radius there! The weakness is clearly that it's been written by people who apparently don't know any computer science (is that true?), and who also aren't used to writing for others. That said, they've written WELL. They're just handicapped in their otherwise good and thoughtful writing by their own lack of abstract knowledge of what they're doing, so they can't explain themselves properly. That's clearly the fault, as far as I can make out. It's like watching some of those pop-artists who can make up songs but can't explain what they're doing. A musician comes along and says, oh, that's counterpoint, or something. Pling. Explanation. The concepts are lacking. Now, I've been kind enough to point out to you where the problems in the docs are. There's no need to go off the deep end - you just have to fix them. Yousay thanks, because fixing it lowers the market resistence to your tool, and enables you to make further inroads against your competitors. If I were to guess at causes, I would say it looks as though the authors didn't have the linguistic concepts in their heads with which to explain them. In particular, My Pet Theory, which is what it is, is that the authors don't understand the difference between syntax and semantics, so they keep saying VALUE when they mean TERM. Result, confusion. You try explaining grammar to a person who doesn't know that they are speaking a language. There is a difference between a THING and the NAME of a thing. Check out your Lewis Carrol. Does that bring it home? No? Then go "learn about it". > Now you are simply arguing semantics with people in an No, I'm not. They are. I _am_ a lingusitic semanticist, as you would find if you looked me up. I can't argue about it - I simply say what is right. > attempt to save face. RADIUS is truly not a complicated protocol. Why Sure - it looks easy. One packet in, one packet out, as far as I recall (and I only glanced at it). What has that got to do with anything? That's not in my competence to talk about and I've not commented on it. Nor do I care about it. > are you arguing over things that truly don't matter in the grand Because they ARE what matters. What I am interested in is the language used to describe the simple (one-action, repeated) state machine that you construct from the description. That's what *I* interface with. I give orders to freeradius. Freeradius understands and does what I mean. To do so, it constructs a state machine and runs it, thus implementing a protocol in conformance with the radius spec. Fine. Freeradius concerns itself with radius. I concern myself with the language needed to talk to freeradius. That's the idea. It doesn't take much to get it right. People have been doing this for nigh on 50 years now folks. Write the language in accordance with what people expect. > scheme? You started out by making assumptions based on previous > experience that were not correct. No, I have not! Where do you get this from? I am telling you what my expectations are, and my expectaions ARE correct, by definition. Just as my expectation that the steering wheel on a car turns the car right when you turn the wheel right are correct. Yes, there are vehicles where it works the other way round. They're wrong. > You then attempted to blast RADIUS > when your assumptions turned out to be incorrect. Poor etiquette. No, I have not! Will you stop this senseless insulting behaviour? Thank you! > Based on your original message, you still have a lot of reading and > understanding to do before making any more posts to this list. I would No I do not. I have no interest whatsoever in reading anything, and I most certainly have no interest in radius! Why do you think I should? Such arrogance! > imagine that you have already taken a look at the users file and the Of course I have. Somebody told me about that about 20 posts ago. That was the start of the whole thread. Their suggesion to write a Foo == BAR, Gum = BAR as a way of setting the field Gum to the contents of field Foo is incorrect, but the docs and examples all make it look as though it is correct. I had to experiment and find out that it was not the way to do it. I then postulated that contrary to all the appearance given in the docs, BAR is interpreted as a literal constant "BAR". That is correct. Somebody then pope up with (:-) piped) the statement that PPP is in the example given, and it is a VALUE. Wrong. It is a NAME. Failure to see the syntax/semantics divide strikes. And as a NAME it's lexically exactly the same as BAR. So I expected it to be a literal constant too, particularly as PPP is a well known string! It's the NAME of the protocol PPP. So the language makes no sense, from the linguistic point of view. One normally designs languages to be regular, not to have exceptional cases. But the situation here is that PPP is a NAMED constant (with value 27 or whatever) and BAR is a LITERAL constant (with value "BAR"). So the language is not regular. Tokens in the same lexical class have interpretations in different (linguistic, that is) semantic classes. (a linguistic semantics is a kind of abstract semantics - the usual domain for the semantics is just the "log" of what you tried to do, so it's all purely formal - the upshot of what I am trying to say is that PPP is a NAMED constant and BAR is a literal constant, and those fall in different classes of VALUES in the abstract domain, they're different kinds of linguistic object). Fine. Be irregular. But since we expect computing languages to be regular, you have to say so. Because otherwise peopel will make all sorts of wrong working hypotheses about your language. > "man 5 users" page, however assumptions are what got you in the position Thanks, I quoted that to you about 5 times in this thread already. Paying attention are we? > you are in now. So, start there, attempt to understand what it is > you're dealing when, and come back to us when you have something more > intelligent to contribute or ask. Look, I don't much appreciate ignorance - when it stands up and makes a noise, the best thing to say is "sit down and cease making that distressful noise". Do so, please. It's not my fault if you don't understand. Peter - List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html - List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
