Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Path: utzoo!utgpu!utcsri!utegc!utai!ubc-vision!van-bc!sl From: sl@van-bc.UUCP Newsgroups: can.general Subject: Re: Canadian UUCP Zone (was Re: Status of Canadian domain) Message-ID: <1267@van-bc.UUCP> Date: Wed, 31-Dec-69 18:59:59 EDT Article-I.D.: van-bc.1267 Posted: Wed Dec 31 18:59:59 1969 Date-Received: Wed, 26-Aug-87 02:28:43 EDT References: <8708190102.AA05431@ephemeral.ai.toronto.edu> <298@ncrcan.UUCP> <1251@van-bc.UUCP> <53@ncc.UUCP> <8708240437.AA05616@ephemeral.ai.toronto.edu> Reply-To: sl@van-bc.UUCP (Stuart Lynne) Distribution: can Organization: Public Access Network, Vancouver, BC. Lines: 98 In article <8708240437.AA05616@ephemeral.ai.toronto.edu> rayan@ai.toronto.edu (Rayan Zachariassen) writes: >In article <1262@van-bc.UUCP> sl@van-bc.UUCP (Stuart Lynne) writes: ># Other than that the UUCP Zone should be able to make it's own decisions on ># the structure of it's part of the .ca namespace (as long as it does not ># choose 2nd level names which collide with names already allocated in other ># Canadian zones). > >There are at least two big problems with this: > >1. Consistency. In an anarchy of that kind, there will be no guidelines > whatsoever for guessing an address. Once again you're imputing too much into a domain name. They are simply an identifier which enables people to send mail to you -- if they have it. Guessing an address is the domain (pun intended) of a Directory Service. You make a request giving as much information about an individual (name, sex, nick name, job description, previous address, employer, etc) and it spits out a list of domain names plus qualifing information about each, allowing you to either make a subsequent request giving more information, or identify the individual you wish to talk to. Heirarchical domain names are no more than a method of dividing a large name space up. Think of what your average unix system would be like to use if it had a flat file system. But does the fact that a file lives in /usr/lib instead of /lib have a great deal of signifigance (other than space availability on the underling disk partions, and the amount of use the system administrator thinks the file might get). > >2. There are a small but very important set of organizations that live > on more than one network. These are mostly the universities and R&D So what, they pick whatever one they want to deal with, and ask for a domain name. Once you have a domain name your mail is routed to you by whatever route specified by whatever name server the sender is using. This is totally independant of the domain name. We already are proving that. With the current Internet and uucp networks, I can send mail to user@xyz.com without knowing anything about what network that xyz is on. It could be arpa or uucp. People in the US on Arpanet can currently send me mail at sl@van-bc.uucp. They don't know that I'm in Canada. I could get a US domain name -- sl@van-bc.org -- and arrange to have my mail forwarded from UUNET to me via X.400 EAN links. Again the users on Arpanet wouldn't know the difference. The moral is I *already* have a good set of domain choices. I don't have to be in .ca. Of course being the flag waving patriotic citizen that I am, I would *like* to be in .ca. > organizations. If they are in N Zones, each with a different policy, > there'll be chaos. Why? Please elaborate. >Also, there may be a misunderstanding of what a "Zone" is in the context >of "The UUCP Zone". This is separate from the technical concept of a Zone >(which delimits the boundaries of delegated authority of a domain master). >The UUCP Zone doesn't set policy; it is an interface to the NIC so that They don't need much of a policy. The only requirement (that I'm aware of) is that you choose a unique identifier within the top level domain that you qualify for (ie. .edu/.com/.gov... ). >Of the three ways of dividing up .CA, there is no fundamental incompatibility >between 'geographical' and 'functional' or 'geographical' and 'organizational'. >However, 'functional' and 'organizational' cannot coexist, partially because >they overlap to a certain extent. Since geographical domains are so wonderfully >compatible with the alternatives, such domains will probably spring up at >some point (after all, there are many many organizations which stay put; >a condominium for example -- the one I'm in uses a computer for various >things - I wonder when they'll get on the net...). But do I really want my mail to go to my condominium for the next 25 years. I'll put it another way. If you where told that you could choose a domain identifier for your personal use, but that once you have choosen it you must use it for the next 25 years. Would you really want one that was based on your current geographical address. I personally can't come up with any suitable scheme that: a. keeps everyone happy b. will last a long time c. will suffice with exponential growth Other than, if you want a name, and it isn't in use (or seem likely to be used, i.e. large companies, cities, etc) then use that. This seems to work fairly well in the US. Why won't it work here. The argument that it works there because they use functional top level domains is not valid. Any of their top level domains has the potential for being larger than .CA. -- {ihnp4!alberta!ubc-vision,uunet}!van-bc!Stuart.Lynne Vancouver,BC,604-937-7532