jlb@apollo.uucp (Joel Breazeale) (08/17/87)
Could someone "in the know" tell how/where one obtains the
software which interfaces to BITNET?
I presume that BITNET software runs under several operating
systems. How about it running under a flavor of Unix and
VMS?
---
Also, I've been running into a problem where the address on
an incoming message has been upcased, presumably by a gate-
way between BITNET and the ARPA internet. Is this a common
problem? Is this gateway software standardized so that fixing
it in one place eventually propagates to all gateways? What
process would I have to go through to get this situation
resolved, contact the gateway site administrator?
Joel Breazeale
Apollo Computer, Inc.
UUCP: {decvax,yale,attunix}!apollo!jlb
ARPA: apollo!jlb@eddie.mit.eduehrlich@psuvax1.psu.edu (Dan Ehrlich) (08/18/87)
In article <36b9d33e.44e6@apollo.uucp> jlb@apollo.uucp (Joel Breazeale) writes: >Could someone "in the know" tell how/where one obtains the >software which interfaces to BITNET? > >I presume that BITNET software runs under several operating >systems. How about it running under a flavor of Unix and >VMS? > The Department of Computer Science here at Penn State offers a package for connecting UNIX systems to BITNET called UREP (UNIX(tm) RSCS Emulation Package). It is currently running under BSD 4.x, BSD 2.x, System V, and probably a few others that I do not know about. Currently a single cpu license costs $US850. To get information you should send a mail message to urep-license@psuvax1.{psu.edu,uucp,bitnet} or by US Mail to: Ms Romayne Bernitt Department of Computer Science The Pennsylvania State University 333 Whitmore Laboratory University Park, PA 16802 A company named Joiner Associates (spelling?) sell a package for VMS called JNET. Their address from the 1st quarter 1986 Data Sources directory is: Joiner Associates, Inc. 732 N. Midvale Blvd P.O. Box 5445 Madison, WI 53705 +1 608 238 8134 >--- > >Also, I've been running into a problem where the address on >an incoming message has been upcased, presumably by a gate- >way between BITNET and the ARPA internet. Is this a common >problem? Is this gateway software standardized so that fixing >it in one place eventually propagates to all gateways? What >process would I have to go through to get this situation >resolved, contact the gateway site administrator? Most reasonable mailer software is case insensitive. This is required by RFC822 which specifies message formats for the Internet. Unfortunately, IBM decided that host/user names must be eight or less UPPER CASE characters. This is from the people who brought you EBCDIC. :-) > >Joel Breazeale >Apollo Computer, Inc. > >UUCP: {decvax,yale,attunix}!apollo!jlb >ARPA: apollo!jlb@eddie.mit.edu -- --Dan Ehrlich <ehrlich@psuvax1.{psu.edu,bitnet,uucp}> The Pennsylvania State University, Computer Science Department 333 Whitmore Laboratory, University Park, PA 16802 +1 814 863 1142
lum%tut.cis.ohio-state.edu@osu-eddie.UUCP (Lum Johnson) (08/18/87)
In article <2854@psuvax1.psu.edu> ehrlich@psuvax1.psu.edu (Dan Ehrlich) says: >In article <36b9d33e.44e6@apollo.uucp> jlb@apollo.uucp (Joel Breazeale) says: >> >>Also, I've been running into a problem where the address on an >>incoming message has been upcased, .... Is this a common problem? > >Most reasonable mailer software is case insensitive. This is >required by RFC822 which specifies message formats for the Internet. >Unfortunately, IBM decided that host/user names must be eight or less >UPPER CASE characters. This is from the people who brought you >EBCDIC. :-) Sort of reminiscent of Breeze's comment on US foreign policy, at least without respect for third world nations, as expressed in their reggae song `Aid Travel with a Bomb'. TOPS-20 allows upto 19 arbitrary ASCII chars, or upto 39 alphanumerics in such things as directory, machine, and user names, and passwords. (Most non-alphanumerics must be quoted to get them to parse in properly; some control characters cannot be quoted for this purpose.) -=- Lum Johnson lum[%osu-20]@ohio-state.arpa lum@{osu-eddie|osupyr}.uucp ..!cbosgd!osu-eddie[!{osu-20|osupyr|tut}]!lum
ehrlich@psuvax1.psu.edu (Dan Ehrlich) (08/19/87)
In article <3952@osu-eddie.UUCP> lum%tut.cis.ohio-state.edu@osu-eddie.UUCP (Lum Johnson) writes: >In article <2854@psuvax1.psu.edu> ehrlich@psuvax1.psu.edu (Dan Ehrlich) says: >> >>Most reasonable mailer software is case insensitive. This is >>required by RFC822 which specifies message formats for the Internet. >>Unfortunately, IBM decided that host/user names must be eight or less >>UPPER CASE characters. This is from the people who brought you >>EBCDIC. :-) >... >TOPS-20 allows upto 19 arbitrary ASCII chars, or upto 39 alphanumerics >in such things as directory, machine, and user names, and passwords. >(Most non-alphanumerics must be quoted to get them to parse in properly; >some control characters cannot be quoted for this purpose.) What and operating system allows and what RFC822 specifies many times are not equivalent sets. It is usually hoped that the spcification will be flexible enough to encompass many different O/S's view of reality. In any event, RFC822 only requires that you support the printable ASCII character set. And that the ASCII characters specified as "special" in addresses are not allowed in names unless quoted. If you really wanted to it is supposedly possible to include any and all of the printable ASCII set in an address. Many mailers do not seem to handle quoting correctly so this is usually frowned apon. While '"This is a very weird address"@psuvax1.psu.edu' is valid according to RFC822 I do not think the version of sendmail we run here would be very happy with it. -- --Dan Ehrlich <ehrlich@psuvax1.{psu.edu,bitnet,uucp}> The Pennsylvania State University, Computer Science Department 333 Whitmore Laboratory, University Park, PA 16802 +1 814 863 1142
dboyes@uoregon.UUCP (David Boyes) (08/19/87)
In article <36b9d33e.44e6@apollo.uucp> jlb@apollo.uucp (Joel Breazeale) writes: >Could someone "in the know" tell how/where one obtains the >software which interfaces to BITNET? > >I presume that BITNET software runs under several operating >systems. How about it running under a flavor of Unix and >VMS? BITNET operates primarily on VM/SP CMS and OS/VS2[MVS] with TSO. There are 2 versions of the software out there -- RSCS V2r2 (IBM's latest) and V1R3 (what everyone with any sense runs if they don't have to deal with SNA). For either of those two operating systems, contact your local IBM rep. For non-IBM machines, your choice is a bit wider. Under UNIX, the people at Penn State University have written a package called UREP that works like a dream. I know it works under BSD 4.[23] and I'm pretty sure that Sys V people won't have too much trouble. You need a synchronous port to attach to the next host along the line, though. UREP was priced around $300/yr last time I checked -- call PSU for the latest details. I *like* this package -- it's well integrated into the Unix environment as well as providing full RSCS services including interactive messages. For VMS....well, it exists. jnet 3.0 seems to be the most popular and reliable. It's reliable and it works, but it will give you screaming nightmares installing it and routing mail, especially if you have a Vaxcluster and call the whole thing by one node name (DON'T DO THIS!!! It may look convenient, but do you REALLY want everything with an origin of POSTMASTER@foo?) Personally, I hate this program with a passion, but there's nothing better available and Joiner Associates do miracles within the rather limited framework of VMS. There are other products for CDC and Prime equipment. Send me mail if you really want to know the headaches involved. > >Also, I've been running into a problem where the address on >an incoming message has been upcased, presumably by a gate- >way between BITNET and the ARPA internet. Is this a common >problem? This is a peripheral "feature" of BITNET. VM/CMS does not accept lowercase userids as valid, so it converts them to upper case. Since BITNET began as the two VM machines at CUNY and PSU talking to each other, the restriction stuck. You'll find that userids longer than 8 characters will get chopped in most situations as well due to the same limitations. >Is this gateway software standardized so that fixing >it in one place eventually propagates to all gateways? The Internet - BITNET gateways do whatever they darn well please. With the demise of WISCVM (one of the most well-behaved and tolerant of the gateways) things are going to get really weird. I would say that some MVS machine somewhere along the way uppercased everything. MVS has a tendency to do that, even when told not to. >What >process would I have to go through to get this situation >resolved, contact the gateway site administrator? Invent a TCP that will run over RSCS protocols and we'll all join the Internet and then the problem will be solved. And hell will freeze over along the same timeline. It's just a matter of inertia. >Joel Breazeale >Apollo Computer, Inc. -- David Boyes ARPA: 556%OREGON1.BITNET@WISCVM.WISC.EDU Systems Division BITNET: 556@OREGON1 University of Oregon Computing Center UUCP: dboyes@uoregon.UUCP
marv@vsedev.VSE.COM (Marvin Raab) (08/21/87)
In article <495@uoregon.UUCP> dboyes@drizzle.UUCP (David Boyes) writes: >Since >BITNET began as the two VM machines at CUNY and PSU talking to each >other, the restriction stuck. You'll find that userids longer than 8 trivial correction: the two machines were cuny and yale -marvin -- Marvin Raab Arlington, VA 22202 ...!seismo!vsedev!marv 703-521-5449 (h) ...!verdix!vrdxhq!vsedev!marv (formerly MFRQC@CUNYVM.s bo5 ipsuv
lum%tut.cis.ohio-state.edu@osu-eddie.UUCP (Lum Johnson) (08/23/87)
In <2856@psuvax1.psu.edu> ehrlich@psuvax1.psu.edu (Dan Ehrlich) says: >In <3952@osu-eddie.UUCP> lum%tut.cis.ohio-state.edu@osu-eddie.UUCP (I) said: >>In <2854@psuvax1.psu.edu> ehrlich@psuvax1.psu.edu (Dan Ehrlich) said: >>> >>>Most reasonable mailer software is case insensitive [, as] required >>>by RFC822.... Unfortunately, IBM decided that host/user names must be >>>eight or less UPPER CASE characters.... >> >>TOPS-20 allows ... > >What an operating system allows and what RFC822 specifies many times are >not equivalent sets. It is usually hoped that the specification will be >flexible enough to encompass many different O/S's view of reality. Agreed, both should try to encompass at least common usage and historical necessity. The ARPANET is about fifteen years old; they've had more than enough time to make and correct all the mistakes one can reasonably make. TOPS is only the OS with which I am most familiar. >Many mailers do not seem to handle quoting correctly so this is usually >frowned apon. While '"This is a very weird address"@psuvax1.psu.edu' is >valid according to RFC822, I do not think the version of sendmail we run >here would be very happy with it. Indeed, I understand some BITNET software cannot handle addresses with "%" or quotes at all. Has this been properly implemented yet? I know of a site which could use this right about now. >--Dan Ehrlich <ehrlich@psuvax1.{psu.edu,bitnet,uucp}> -=- Lum Johnson lum[%osu-20]@ohio-state.arpa lum@{osu-eddie|osupyr}.uucp ..!cbosgd!osu-eddie[!{osu-20|osupyr|tut}]!lum