Makey@LOGICON.ARPA (Jeff Makey) (01/07/89)
In article <6182@hoptoad.uucp> gnu@hoptoad.uucp (John Gilmore) writes: >(Referring to a standard for sending databases over the net): >> =- Format can be ASCII or UUENCODED binary. > >We have two new versions of netnews coming out within the month -- C >news and TMNN (News 3.0). I believe both of them are 8-bit-clean, that >is, the data part of a message can have any 8-bit characters in it >(including nulls and very long lines). Most of this support is >necessary for "un-American" :-) character set support anyway. If this is >really true, we should consider simply sending binaries as binaries. Even if it is true, posting binary files in binary format is a bad idea for several reasons: 1) Not all machines have 8-bit bytes. Multics, DEC-10, and DEC-20 machines (there may be others) all have 9-bit bytes (and 36-bit words). Some of the major archive sites on the Internet still use this ancient 36-bit hardware, and they can't be ignored. 2) My terminal does strange things when it sees certain escape sequences. The last thing I want is to be reading an article that suddenly includes a binary section. 3) Many sites will continue to run obsolete news systems for years. Why don't these wonderful new news systems have builtin binary file encoders/decoders (similar to the rot13 support in rn)? I think that would solve the problem. Note that I have directed followups to news.software.b (for lack of a better place). :: Jeff Makey Department of Tautological Pleonasms and Superfluous Redundancies Department Disclaimer: Logicon doesn't even know we're running news. Internet: Makey@LOGICON.ARPA UUCP: {nosc,ucsd}!logicon.arpa!Makey
david@ms.uky.edu (David Herron -- One of the vertebrae) (01/09/89)
In article <271@logicon.arpa> Makey@LOGICON.ARPA (Jeff Makey) writes: <In article <6182@hoptoad.uucp> gnu@hoptoad.uucp (John Gilmore) writes: <>(Referring to a standard for sending databases over the net): <>> =- Format can be ASCII or UUENCODED binary. <>We have two new versions of netnews coming out within the month -- C <>news and TMNN (News 3.0). I believe both of them are 8-bit-clean, that <>is, the data part of a message can have any 8-bit characters in it <>(including nulls and very long lines). ... <Even if it is true, posting binary files in binary format is a bad <idea for several reasons: And not just the reasons that Jeff states .. Not all links are "8-bit-clean" ... the one we run across BITNET is an example. -- <-- David Herron; an MMDF guy <david@ms.uky.edu> <-- ska: David le casse\*' {rutgers,uunet}!ukma!david, david@UKMA.BITNET <-- Now I know how Zonker felt when he graduated ... <-- Stop! Wait! I didn't mean to!
ljdickey@water.waterloo.edu (Lee Dickey) (01/17/89)
In article <271@logicon.arpa> Makey@LOGICON.ARPA (Jeff Makey) writes: |In article <6182@hoptoad.uucp> gnu@hoptoad.uucp (John Gilmore) writes: |>(Referring to a standard for sending databases over the net): |>> =- Format can be ASCII or UUENCODED binary. |> |>We have two new versions of netnews coming out within the month -- C |>news and TMNN (News 3.0). I believe both of them are 8-bit-clean, that |>is, the data part of a message can have any 8-bit characters in it |>(including nulls and very long lines). [...] |> ... we should consider simply sending binaries as binaries. | | [...] posting binary files in binary format is a bad |idea for several reasons: | 1) Not all machines have 8-bit bytes. ... | 2) My terminal does strange things when it sees certain escape | sequences. ... | 3) Many sites will continue to run obsolete news systems for years. ... I agree that binaries should be encoded into visible (graphic) characters. UUENCODE is good enough for some. However some problems arise when files go through gateways. The UUencode scheme includes the tilde (~), and some EBCDIC machines still have translate tables that get it wrong. (Example: My most recent bout with this was in trying to send a file to a friend in the UK, and we made the mistake of sending the files through UKACRL.) Another problem with UUencode is that the blank character is used, and again, some EBCDIC software smashes trailing blanks. The scheme used by "BTOA" and "ATOB" avoids these problem areas, and is more efficient, since it is base-85 instead of base-64. I would not object to doing some file compression before encoding. At least one binaries groups on the net already have conventions like this, using ARC and then UUE. -- L. J. Dickey, Faculty of Mathematics, University of Waterloo. ljdickey@WATDCS.UWaterloo.ca ljdickey@water.BITNET ljdickey@water.UUCP ..!uunet!watmath!water!ljdickey ljdickey@water.waterloo.edu