From: ag@unido.UUCP (Anke Goos) Newsgroups: eunet.general Subject: (long) text of e-mail intro Summary: Constructive Comments wanted! Date: 15 Jan 90 20:57:59 GMT Reply-To: Anke Goos Organization: University of Dortmund, W-Germany Hello EUnet-landers, Your comments on the following EUnet documentation are welcome: A User Introduction to Electronic Mail Daniel Karrenberg Anke Goos (c) Copyright Dec 1988 Right for copy permitted, if for no direct commercial use. Support the European E-Mail Directory, published by EUUG by ordering the book! 1. A User Introduction to Electronic Mail - - ---- ------------ -- ---------- ---- This chapter is an introduction to electronic mail message formats and addressing in wide area networks in the research and development community. We will focus on 'domainaddressing' worldwide taking a quick look at European ------ conventions. Finally, we glance at future X.400 addressing. This document is not a formal specification for the expert but a guide for the user. The style is therefore deli- berately informal and sometimes redundant with cute exam- ples, with the aim of combining pedagogy and attractiveness to the newcomer or even run-away user in computer networks. 1.1. Message Format - - ------- ------ Almost every computer system has its own electronic mailing system and often the formats are quite incompatible. How- ever, there is one format which is widely used in many com- puter networks. This particular message format is often iden- tified by the name of the document defining it: RFC 822. Although RFC 822 + originated as a standard for the US research network Internet, it is now very widely used throughout the world's research and development (R&D) net- works. This "RFC 822" message format has become the de-facto transfer standard of today's mail systems. The format has become particularily important where messages cross from one network into another through gateways. Such gateways will understand not only the message format but also the address syntax described in RFC822. Even if your mail system cannot use the RFC822 address syntax, read this chapter anyhow because you are very likely to come across these addresses in messages from other networks or on letterheads and busi- ness cards. We will use the RFC822 format as a reference in this guide. Later sections will discuss the peculiarities of different networks and how to deal with them. ------------------------- + Apropos RFC822: RFC is short for Request for Com- ment. The full name of RFC822 is "Standard for the Format of Arpa-Internet Text messages" and cited in our bibliography. Arpa-Internet is some sort of "meta- network" of several US networks, in first instance the former (D)ARPANET. This is named after the Defense Ad- vanced Research Project Agency, which originally ini- tiated it and is part of the US Department of Defense. 1.1.1. Header and Body - - - ------ --- ---- Let us briefly examine a hypothetical message: From: Molly Cule To: Alec Tron Cc: ATOM32@DHDCHE11.BITNET (Germanium), oxygen-lovers@chemistry.mit.edu Subject: new bindings Date: Sat, 26 Nov 88 15:13:38 +0100 Dear Alec, I find your ideas quite interesting and have forwarded your outline to the oxygen-lovers group in the US. Maybe you can get some more feedback from there. Regards Molly Mail messages consist of a header part and the message itself, which is often referred to as the body. The message starts with the header which is separated from the body by exactly one empty line. The header lists information about the sender, the recipients, the posting date, the subject of the message, etc. 1.1.2. Header Lines - - - ------ ----- Most mail systems only offer the most essential header lines for To: From: Date: and Subject: , where From and the date is often inserted automatically. But you may give your mail a lot more attributes. Here are the most important header lines and their meanings. From: Sender's address There may only be one of these lines in the header. Date: Date and time the message was sent. It is important that the notation of the date follows a common form, because many programs must be able to compare dates, sort messages in your mailbox and the like. Most mail systems will automatical- ly insert a correctly formatted Date: line for you, like: 12 Sep 88 02:17:38 GMT , some will not even allow you to insert it yourself. To: Primary recipient(s) of the message. This line can specify more than one desti- nation, in which case the addresses are separated by commas. Subject: Subject of the message. This is free text. Choose a short, mean- ingful subject and don't forget punctua- tion. Free telephone calls? is different from Free telephone calls! . Although the presence of this line is not mandatory, it is highly recommended to always include it, if only for politeness to the reci- pient who prefers to have a short summary of the contents. Furthermore a good choice of title eases the filing and searching facilities in many mail systems. Cc: Copy recipients. This is the carbon copy of the electronic age, means a copy just for your information. The example shows how data, that does not fit on a single line (like the second addressee in the CC-list) can be put on subsequent lines beginning with white space. There may also be something like BCc: for "Blind" carbon copy if you want to send a copy to a third party without notice to other recipients of the message. Message-Id: Unique identifier of this message given by the sending host. Something like <8808140800.AA1302@unido.irb.informatik.uni- dortmund.de> . Used to make cross references between mes- sages automatically and inserted by most mail software. Received: Trace information used to analyze mail delivery problems. Usually multiple lines showing which machines the message passed through at what times and date. Resent-From: Address of the person or program where the message comes from. Labels starting with Resent- indicate that the message has been forwarded to you by the person identified in the Resent- From: line. There may also be such lines as Resent-To: and Resent-Cc:. Reply-To: Obviously the address of the person whom to reply to. In most cases this should be the address of the sender. But this is an opportunity to automatically forward all replies to another place or person without having to ask for it in the message itself. There may also be some other lines like Sender: to identify the actual sender if this is different from the author(s) shown in the mailbox address in e.g. From: postmaster . Maybe you are surprised whether your own mail system is able to offer all these intelligent facilities? Read the docu- ments of your mail systems and ask your local mail guru what is possible and what is not. 1.1.3. The Message Body - - - --- ------- ---- The message body contains the text of the message itself in the format of ASCII (or EBCDIC for BITNET). The empty line separating header and body is an essential part of the for- mat. Most mail systems ensure its presence automatically. Most mail systems do not permit the sending of binaries without taking some special action. This sounds ridiculous for networks initially made by and for computer people. But it's the problem of the "eighth bit" which might get lost when transferred in ASCII or EBCDIC. To save all control codes or special characters bound to the eighth bit, you are supposed to encode your binaries into pure ASCII or EBCDIC before the transfer and decode it at the receiving end. In most mail systems, some software exists to do this. Note that there is a convention in EUnet not to ensure a transfer exceeding 100 kilobyte in the message body. 1.2. Address Syntax - - ------- ------ Let's look at three forms of a hypothetical sender's address as you might see them in the header of a mail message: From: Cule@chemistry.oxbridge.UK (Molly Cule) From: Molly Cule From: Cule@chemistry.oxbridge.UK The string Molly Cule -- although probably giving the sender's name -- is not part of the address used by the sys- tem to route messages. The mail system will treat all three addresses exactly the same because it only looks at the string Cule@chemistry.oxbridge.UK . This is what we call the mail address. Molly Cule in this example is regarded as a comment which is passed on untouched by the mail system. You may either put the comment in () or the address in <> according to taste. If the comment includes punctuation characters it is more correct to use the () form as in burner@chemistry.uni-goettingen.de (Dr. Bunsen). And if there is a blank in the comment it may be even better to insert double-quotes like "Dr. Bunsen". Let's look at the mail address itself now. First of all split the address at the @ character - there may only be one: Cule @ chemistry.oxbridge.UK The part to the left of the @ is called the local part or mailbox name, the part to the right the domain. The domain specifies where that particular mailbox is located. If the address has no @ at all, like From: Cule , it should be a local address, means a local mailbox in your organization. Most mail systems allow you to omit the local domain as a typing shortcut when sending mail to a local mailbox. But the full domain should be added by the system before it is sent out. 1.2.1. What's in a Mailbox Name ? - - - ---- - -- - ------- ---- The mailbox can belong to a user or a group of users, or it can be the place to put mail for someone with a specific function such as postmaster. In general, it is very hard to guess the mailbox name of a particular person. However there exist some conventions about the form of mailbox names used for special purposes: x-lovers Mailbox names with dashes in them are likely to be special. If the name looks like belonging to a group of people it is probably a distribu- tion list or mailing list. In BITNET/EARN a list which discusses FOO is named FOO-L@host.bitnet. Mailing to such a mailbox will redistribute the message to all those subscribed to the list. It is generally a very bad idea to send your subscription to the list ad- dress itself since your request is not really meant to be seen by pos- sibly hundreds of members but only the list maintainer. x-lovers-request Names ending on -request are admin- istrative addresses of distribution lists. This is where you send re- quests to subscribe or unsubscribe. You may even send your contributions to this mailbox if the mailing list is moderated. The moderator is some sort of volunteer editor giving the mailing list a better quality as a "human filter" for relevant topics. postmaster Every domain and site is supposed to have a postmaster or postman mail- box in BITNET where you can address questions and report problems con- cerning the mail system. MAILER-DAEMON This is the agent of your friendly mail system itself. Messages coming from a similar address are most likely reporting problems with one of your messages or they will notify you of its delivery. Don't be intimidated if you can't understand all details of these messages which may show strange faults in the mail system. If however you cannot make sense of such a report at all, be sure to mail or show the Bwhole message to your local mail expert who might need Ball of the information in it to diagnose the problem. local%domain If you see % signs in the mailbox name then this is most likely not a real mailbox name at all. Most mail systems will interpret mailbox names with % s in them as full mail addresses. More about this later in the section about domain addressing. John.Mailuser A growing number of mail systems let you address users by their full names. First and last name are usu- ally separated by a dot ". " or an underscore sign " " like in - "John Mailuser " because it is - unhealthy to have blanks in mailbox names. This is good practice because it makes mailbox names somewhat guessable. xyz123AB If you see cryptic mailbox names they are most likely from an organi- zation where they use computer account numbers as mailbox names. These are found mainly on EARN and BITNET. In these cases it is good practice to include a comment in the address giving the person's real name. dmr On Unix systems it is widespread style to take user names as mailbox names. Many Unix users prefer short user names, and initials in lower- case letters are widely used. After all dmr alias Dennis M. Ritchie and ken Ken Thompson created Unix. Unfortunately, we are stuck with the problem of guessing the mailbox name for a given person until there is a universal user directory service like some sort of electronic tele- phone book which can be used to have "a look" for an unknown mailbox name. Some organizations or countries and even mail systems already tend to enforce Firstname.Lastname as a mailbox name. Note: maibox names are case sensitive. Although our examples showed a mixture upper and lower case in the mailbox name Cule@ it might be a local bad decision not to allow the flat form cule as well as there are still mail systems like that of BITNET which are not able to address mailboxes in mixed cases. 1.2.2. What's in a domain ? - - - ---- - -- - ------ The part to the right of the @ is called a "domain" and denotes the place where the mailbox is located, which usu- ally is an organization. In some older forms it can also denote one or several computers. To find out, you split the right hand part at the "."s into so called subdomains. The rightmost part is called the top-level domain. The syntax for a mail address therefore is: mailbox@subdom-n. ... .subdom-2.subdom-1.top-level-domain Our example Cule@chemistry.oxbridge.UK gets split up as follows: mailbox name Cule subdomain-2 chemistry subdomain-1 oxbridge top-level-domain UK In the above example there is a chemistry department. Which one of the hundreds of chemistry departments worldwide is determined by the next subdomain oxbridge . Even if you don't know oxbridge you can infer from the top-level domain UK that it must be in the United Kingdom. This way you may see the domain address as a hierarchical set of geographical or organizational units. It is irrelevant whether you use capital letters or lower case letters or a mixture of them in the domain part of an address. You see that the higher the number of the subdomain the more detailed the information about the destination. Subdomain-n like chemistry is always to be interpreted as being "inside" subdomain-n-1 like oxbridge . Each subdomain is considered to be "inside" the one to its right in the address. The one on the far rights can be considered the "top-level" domain. In fact these addresses are often referred to as "domain addresses". This term is a synonym for "RFC822 address". Many people and mailer programs will read domains just the other way around starting with the top-level domain. From this point of view the domain address shows some sort of its internal "routing" information. Your mail to Cule@chemistry.oxbridge.uk will find its way by a mailer system looking at the rightmost part UK. Arrived in UK the mail would be forwarded or "routed" to subdomain-1 of oxbridge, from there to subdomain-2 for chemistry. It might be a local feature of a domain address to show something like local%domain@ like Cule%hotspot@chemistry.oxbridge.UK to the outside world. To understand this you have to imagine that each local domain like the chemistry in oxbridge prob- ably has one gatekeeper host for incoming and outgoing mail. If Gatekeeper has no directory to identify on which host user Cule may be reached, this has to be given within the address. Thus the Cule%hotspot just indicates that hotspot is the local or even distant host where to forward the mail. To do that Gatekeeper would transform the adress Cule%hotspot@chemistry. ... UK to Cule@hotspot . This way you may even address one or more hosts behind an official domain, like Cule%hotspot%gatekeeper@chemistry.oxbridge.UK . This stuff is not pure "holy" RFC822 and not handled the same throughout the network world. But at least you should know that there is no real mystery behind such strange "mailbox names" which are no real mailbox names at all. Considering such local variations it's no wonder that there is some diversity in the worldwide use of the domain. For- tunately, there are only a few top-level domains used in the world of networks. These fall into three classes: countries, networks and special cases. 1.2.3. Country Toplevel Domains - - - ------- -------- ------- The countries are denoted by the two letter abbreviations specified in standard ISO3166 of the International Standards Organization (ISO). As an example here are the European countries listed in this directory: AT - Austria BE - Belgium CH - Switzerland DE - W. Germany DK - Denmark ES - Spain FI - Finland FR - France GR - Greece GB - Great Britain (ISO-synonym for UK!) IE - Ireland IS - Iceland IT - Italy LU - Luxembourg NL - The Netherlands NO - Norway PT - Portugal SE - Sweden TR - Turkey UK - United Kingdom YU - Yugoslavia To prove that we are not eurocentric: Note that there are still some more country top-level domains, such as NZ for New Zealand or AU for Australia, JP for Japan and US for United States. Be warned however of a British peculiarity. You may come across addresses like Cule@UK.oxbridge.chemistry The British drive on the left hand side of the road, so you shouldn't be too surprised that they chose to write their mail addresses "backwards". Outside the UK you should turn it back into the "normal" continental ordering Cule@chemistry.oxbridge.UK if it suits you or your mailer better. Your mailer may correctly handle the address the wrong way round like this. Actually, the gateways between the UK networks and the rest of the world turn around the domains for you. But unfortunately sometimes this does not happen on letter- head and business cards as they cross the Channel. From time to time a left handed address might escape in the body of a mail message, too. Budejovice, + Czechoslovakia? :-) ++ Normally both domain orderings will be routed correctly by clever mail software looking for an innermost subdomain UK . But beware - some gloating people are already waiting for the moment when Czechoslovakia or domain .cs join the net- work! As the subdomain cs for computer science frequently occurs at the lowest level of a UK-address this will prob- ably cause problems. Just imagine the address uk.ac.bude.cs Is this the computer science department of Bude-Stratton University or the faculty for academic study of brewing in ------------------------- + Budejovice is the home of the real "Budweiser" beer ++ BTW: The "smiley face" :-) is used in electronic mail to mark something said in a humorous mode. You don't know why? Turn your head to the left and look again. It's very advisable to use it to make your in- tentions clearer as written words don't pass the into- nation and mimic of the spoken word. The human kind showing a lot of emotions there exist a lot of so- called "icons". I'm sorry :-( , even unhappy :-(( for not to introduce ALL of them, but I'm *sure* you will work out the irony >-) or the wink ;-) in the elec- tronic eye for yourself next time. 1.2.4. Network Toplevel Domains - - - ------- -------- ------- The network top-level domains are a trick to incorporate ex- isting networks and their mailboxes into this naming struc- ture. Historically a lot of networks have developed elec- tronic mail systems based on their own internal naming schemes. These naming schemes can usually name mailboxes and computers such as in host!user user at host host::user To integrate these existing networks into the domain naming scheme special "network" top-level domains have been intro- duced temporarily until everyone uses domain addresses. These top-level domains only have one subdomain and that denotes the host computer on which the mailbox resides. For example: host!user -> user@host.uucp user at host -> user@host.bitnet host::user -> ? user at a Decnet host The later ? points to the fact that you cannot predict how such a DECnet address of VMS mail can be transformed to a domain address - if at all - as it depends on the nearest gateway in its particular environment. This network form of naming is not as easy to interpret as the organizational names with countries as top-level domains. Most of these networks are currently converting to country top-level domains but as this is a slow process you can expect to see these top-level domains for some time to come. If it is possible to use both addresses with a network top-level domain and addresses with a country top-level domain the country based address is usually preferred. But there are countries with no national naming authority at all so that naming is a matter of the different national or sometimes international networks. In different networks even the organizational subdomain concept can be very different. In HEPnet for example, a worldwide network of High Energy Physicists the country domain is subordinated to the top- level domain of hep. Therefore Hepnet people might think user@nikhefh.nl.hep should be the correct structure for the Dutch colleagues of the universal HEP community and the computer people would think that hep is nothing more than one of the components of the Dutch name space i.e., being user@nikhef.hep.nl. People with some experience think that division of subdomains by discipline is even more difficult to manage than by geography ... The most important worldwide network top-level domains are: uucp is used for UUCP based networks worldwide (includ- ing EUnet in Europe and Usenet in North America). An internal address in these networks is converted to a domain address as follows: host1!host2!target-host!user <=> user@target- host.uucp More about this later in the EUnet section. bitnet is often used as a synonym for one worldwide net- work system including BITNET, EARN and NETNORTH. These academic and research networks technically form a single network based on the IBM RSCS proto- cols but differ in their management and geographi- cal extension. Thus EARN is limited to the Europe and the Middle East and Netnorth to Canada. An internal address in these networks is converted to a domain address as follows: XYZ0123@DDOHRZ21 <=> xyz0123@ddohrz21.bitnet Most domains of EARN show a mixture of letters and numbers in their addresses as ccs021@ddohrz21.bitnet This might look a bit irritating for EUnetters but these names are as systematic as 8 symbols can be: D - Germany DO - the city of Dortmund HRZ - the institution "Hochschulrechenzentrum" 1 - the operating system of host (1 - IBM VM 2 - IBM OS/MVS 5 - DEC VMS 6 - Unix) 1 - the number of the machine With some experience you will see that HAMCWI6.bitnet should denote a Unix machine at the CWI in Amsterdam in Holland. This particular machine however happens to be known only by its alias name: MCVAX.bitnet . earn is a synonym for the European Academic Research Network. Its members all belong to academic insti- tutions like universities or research centers. EARN is part of BITNET but is not as widely known by gateways. Since bitnet is widely known and includes earn on the technical level of networks it is generally safer to use bitnet . csnet is an obsolete top-level domain used by hosts in the Computer Science network CSNET network which are not yet converted to other addresses. Founded in 1981 with financial help of the National Sci- ence Founcation to support research and develop- ment of US computer science it is now integrated into the Internet Domain addressing scheme. You may seldomly find one or the other strange domain which does not fit into all the above networks. For this rare case you may take a look at this list of other minor and major networks from all over the world which may be reached through links or gateways from EUnet. aristote - French academic network arpa - ARPA-Internet, an US research network cdn - Canadian Reserch network cern - Swiss CERN chunet - Swiss University network dec - Dec's Easynet DFN - German X.400 network dunet - Denish university network funet - Finish university network HEPnet - High Energy Physicist network iris - Spanish X.400 Research Network mlnet - Network of the university of West Ontario osiride - Italian research network oz - Australian university network sunet - Swedish university network surfnet - Netherlands X.400 research network uninett - Norwegian research network 1.2.5. Special Toplevel Domains - - - ------- -------- ------- You will also come across special top-level domains that don't fit into the either "country" or "network" classes: com - commercial gov - government edu - educational mil - military net - networking organizations org - other organizations These top-level domains belong to subnetworks of the North American research network Arpa-Internet and still exist for historical reasons due to the major influence of the former Arpanet on the evolution of networks and networks today. These domains should be subdomains of top-level domain US but at the time the domain system was "invented" the inven- tors didn't think of network life outside the US. So we are stuck with this colonialism :-). 1.2.6. Domain Naming Conventions - - - ------ ------ ----------- In 1988 the U.S. UUCP network began to introduce a top-level domain .US . Therefore, you may today also see addresses like user@organization.city-abr.state.us or user@foohost.city.state.us like ewing@killer.dallas.tx.us . In the above example tx stands for Texas and dallas , no doubt, for the city "Dallas". You may not be surprised to find a host killer in this area famous for the Southfork Ranch, oil, money, scandals. As you see there are different interpretations even in the near-standard of domain address- ing. But subdomains of countries are usually organizations, wether a whole state or companies with their own subdivi- sions. In Europe the notation of subdomains is subject to common agreement with the other networks in each country. In most of the European countries EARN has not yet moved to domain- addressing. Some countries have a general distinction between academic members like universities and research and the companies by using first level subdomains like ac and co. Note that this is not edu or com as these parties would be named in United States ! The only common agreement seems to be a great variety within limits. You may find domain addresses like texas.dk - no, not a town or state but just Texas Instruments in Denmark. There are solutions including site.organization.country , e.g. csinn.capsogeti.fr a French company named Cap Sogeti Innovation Gemini with the sub-domain csinn. The introduction of domain-addressing on the British isles is already history, showing addresses like (in continental notation) music.ed.ac.uk for the Music Faculty of Edinburgh (=ed) University (=academic) in the United Kingdom. In UK they did not agree whether long or short names were better. Thus all systems can have two names, e.g. music.edinburgh.ac.uk . Is there any message in your mailbox where a very long address is running out of your screen? This probably is mail from Germany showing the whole name of the host, one to two organizational subdomains, and the organization plus town, for example: anke@laura.irb.informatik.uni-dortmund.de . They are verbose those Germans :-) , unrelenting in their efforts to be most logical. On the other hand, however, paper mail bearing those e-mail addresses has already been properly delivered by the German surface post office! 1.3. Network peculiarities - - ------- ------------- What to do if your mail system does not understand or pro- duce domain addresses? Well, first of all, ask the person responsible for the (mail) system to provide this func- tionality along with some form of automatic routing. They will not have much excuse not to do it, since there is "free" mailer software that does support this for almost any computer environment presently in use in the R&D world. In the meantime, refer your local administrator to the national coordinators of your network whose addresses you will find in the chapter 2 about "networks and contacts". We know that screaming at systems people often takes some time to have an effect. So the next sections give a few rules on how to cope until they do something. You can go directly to the section dealing with your type of machine and network: 1.3.1. EUnet - - - ----- It should be extremely rare to find a Unix system which doesn't understand domain addresses. If you are connected to EUnet ask your systems manager to speak to the manager of the national EUnet site about what better software your sys- tem could run. In the meantime find out (from your systems manager) how to address the national EUnet backbone node or another "smart" machine nearby and address your mail as follows: host1!host2!smart!chemistry.oxbridge.UK!Cule This message will be sent to host1 which has to have a direct link to the sender's machine. host1 will send it on to host2 which in turn will send it to smart which will know how to route to chemistry.oxbridge.uk . If this fails to work you can try as a last resort host1!host2!smart!Cule@chemistry.oxbridge.UK The last form is an extremely dangerous kind of address, mixing ! and @ . Keep your fingers crossed as it is not at all clear whether any given mail system will send this mes- sage to host1 -- according to UUCP-style -- or to chemistry.oxbridge.UK . There's a general rule, however, that @ has priority over % and !. You should only use such mixed addresses if there is really no other possibility. Apropos UUCP: The so-called bang-addressing with host1!host2!host3!user is a reminiscence of the original UUCP store-and-forward transport mechanism where addresses are routes of host names read from left to right and separated by ! . This sort of addressing had the obvious disadvantage that an address for a given user was different relative to the actual position of the sender. Furthermore, the address might be very long and provoke mistakes if you have to type your way through to someone on the West Coast of the United States :-). Thus the Unix People have found better routing systems based on so-called "maps" of direc- tories with all officially registered sites. In communicating with EUnet participants you may sometimes face such long lists of hosts in the From: line of your header. Normally you might ignore it. If you want to pull any knowledge out ot the lines, take them as a very rough information on which path a mail had taken and how far the sending host might be! But there might be the rare case that you want to turn this sort of address into a domain address. Let's assume there's a mail From: mcvax!ukc!gatekeeper!hotspot!Cule and no better domain address in sight. Transforming a bang address into a domain address just means deleting all the path information but the two ending terms for the host and user. In this simple example hotspot!Cule would have to be turned into Cule@hotspot.uucp to form an acceptable uucp-address. But there is an exception to any rule: The above rule only applies to officially registered hosts! A mail to Cule@hotspot.uucp would most certainly result in a reply a mail-daemon telling "host or domain unknown". With a look in the domain index of the European E-mail directory - or in the map data itself - you may check whether domain hotspot or Gatekeeper is a known site. In any case you have to know one official UUCP host out of the row. Taken for granted that ukc is official you could form an address like Cule%hotspot%Gatekeeper@ukc.uucp. 1.3.2. EARN - - - ---- Before sending a message into other networks from an EARN machine it is most important to know that you should never need to use a different address if your host supports a modern mailer. It should be the exception to route your mes- sage "by hand" to the next possible gateway between two net- works. Your communication with Non-BITNET networks may be eased by free software like "Crossnet". Apart from this very simple tool to deal with gateways other Message Handling programs like "Crosswell" or UCLA (University of California, Los Angelos) might provide a comfortable mail system. Con- vince your local node administrator to install one. The reason for this special gatewaying lies deep in the heart of the addressing scheme of simple BITNET mailers: They accept only addresses of the form USER@HOST consisting of a maximum of eight upper case letters or numbers in either user name or host name and have their own difficul- ties with more than 80 characters in a line. (BTW: This is the reason why the local postmaster (10 characters) of the Unix domain is only a postman at any EARN node.) This does not leave enough room to represent domain addresses at all. Therefore each mail from BITNET destined for other networks like EUnet has to be explicitly or automatically routed via a gateway. Any gateway is located at a host belonging to the two networks and transferring all mail between these net- works with help of special programs. These gateways reside at normal EARN addresses such as MAILER@MCVAX . At the time of writing there are two official gateway addresses in Europe: MAILER@MCVAX in the Netherlands and MAILER@CERNVAX, reserved for the High Energy Physics Commun- ity, BSMTP@UNIDO for German EUnet people and some unofficial gateways in other countries. Your local EARN postmaster should know which one to use. In any case MAILER@MCVAX should work as a gateway from all EARN/BITNET domains to all EUnet domains. If you don't have a modern mailer to enable automatic gatewaying you will have to read on how to do this "by hand". Once the message arrives at one of the above gateway address, it is read by a program called the "BSMTP". In order to find out the addresses of the recipients on the other networks, the gateway reads a special part of the mes- sage it receives. This part is called the BSMTP envelope, which means a specific format for the Batch Simple Mail Transfer Protocol. It is wrapped around a normal RFC822 message by the sender or her mail system like a paper envelope around a letter. It contains the sender's and the recipient's addresses just like a normal envelope do. The gateway unwraps the RFC822 message from the BSMTP envelope and sends the message to the recipient on the foreign net- works. Usually the gateway program does not have to look at the message itself, it gets all the information it needs from the BSMTP envelope. As an example here is a possible input for mail from YOURNAME@YOURHOST on EARN to hername@herhost.uucp : HELO yourhost.BITNET VERB ON TICK xyz1234 MAIL FROM: RCPT TO: DATA Date: Fri, 02 Sep 88 22:44:01 GMT From: yourusername@yourhost.bitnet To: hername@herhost.uucp Subject: This is the message body separated from the header by an empty line. In networks like EUnet there is a common agreement to limit the message body to 100 Kilobyte. The body is terminated with a single dot, by itself, at the beginning of the line! o QUIT The lines that begin with words in all capitals plus the line with just a dot on it form the BSMTP envelope. The top-level domains bitnet and uucp or country codes are very important and must be included. Case is ignored in all EARN and BITNET addresses. The blank line between the message header and its body is also required, not to forget the single point at the beginning of a line to terminate the body! You may leave out Subject: and Date: if necessary. The same applies for the the BSMTP extensions VERB ON if you don't want to have long "verbose" records for a failed mail from the Gateway and TICK for later identification of the corresponding reply. All these are not essential but are useful in most cases. The assembled file should be transmitted to your next gate- way address on EARN using the appropriate commands depending on the operating system. The basic formats used for BSMTP-gateways are described in the SMTP document (RFC821) and a short paper by Alan E. Crosswell of Columbia University, who adapted the Simple Mail Transport Protocol SMTP into BSMTP to be run on BITNET. You are however not supposed to read these documents any further. Instead you should now already be a run-away user in sending a mail through the next gateway. 1.3.3. X.400 - - - - --- X.400 is the common standard by the International Standards Organisation ISO and the International Telegraph and Tele- phone Consultive Committee CCITT (Comit Consultatif Interna- tional Tlphonique et Tlgraphique) for Message Handling. Many European networks have plans to use the standards of X.400 eventually and it is used by the RARE MHS service already as a common standard. RARE is an acronym for Reseaux Associes pour la Recherche Europeenne and is a meta-network of European academic networks including Surfnet, Uninet, Funet, Janet, Sunet, Heanet, Aconet, DFN, Switch, Reunir, Fundesco, Centernet. At the moment existing RARE services live up to the name: rare. But in future it might offer a common and unified format for communication among different systems. In addition to normal text, X.400 messages can contain other formats, i.e. fax messages and even voice recordings. The addressing is potentially very user friendly. You will even- tually have to know only a unique set of so-called attri- butes describing the recipient. These attributes will then be used for an automatic search in an electronic directory, which may roughly be described as some sort of electronic telephone book for e-mail domains and addresses. Related to these directories you may stumble about the cryptic code of X.500 which determines the convenient ISO-standard. Unfortunately directories of user names are not yet here and a mapping to RFC822 addresses is also needed to address non-X.400 users. So RARE MHS uses a hierarchical set of organizational attributes to address its users: Country specifies the country of the recipient. This uses the ISO country codes similar to that of RFC822. Remember? UK for United King- dom, DE for Germany, (not for Denmark as some may think). Instead DK for Denmark and so on ... AMD The Administration Management Domain speci- fies the public X.400 carrier. In UK for example the name of the national PTT British Telecom. PRMD Private Management Domain specifies which private X.400 carrier is used. That might be BITNET, EARN, EUnet, CSnet or a private organization. Organization specifies the organization of the recipient. This could be a company or a university. The fictive university Oxbridge would probably fill in here. Org. Unit specifies sub organization(s). There can be more than one organizational unit. In our example however there is no further subunit like laboratory 1 or 2, just chemistry. Surname should obviously be the family name, thus Cule. Maybe someone who wants to have a nick- name in the address would have to cheat a little bit to fit into this scheme. Givenname if Surname should be obvious .... all Mollies and Johns of the world should happily fill in a firstname here. As this example may prove it might be not impossible to find a X.400 equivalent for any original RFC822 address. But reality will be a bit more complicated than our simplified example. Nevertheless there may already be the rare case that a X.400 mail might reach your mailbox. To reply to this message you could simply take the address string of the From:-line. This might look like Firstname.Surname@orgunit.orgunit.org.pmd.amd.country - or more probably a subset of this like Surname@orgunit.org.country. If the address is not yet extended by a gateway address it may be more secure to do this. For UUCP-participants this might end into something like Firstname.Surname%orgunit.orgunit.org.pmd.amd.country@gateway.uucp. EARN people will probably have to send their mail to the X.400-participant by a gateway mechanism similar to that described in the EARN section. Fortunately there is a specification (RFC 987) how to map X.400 messages and addresses automatically into the formats of RFC822. You might even need less attributes than demanded by RFC822. Unfortunately the mapping of addresses is still done differently at different places in Europe and different mailer systems. This may cause some confusion but is han- dled by gateway mapping tables. Unfortunately there is not yet any standardised short way to show each X.400 address as a text-string like an address on business cards, as different X.400 systems may use different representations. RARE is working on a standard notation for the European Research & Development community. Copyright 1988 permitted by Daniel Karrenberg & Anke Goos, if for no direct commercial use and for EUnet purposes