rmarks@KSP.Unisys.COM (Richard Marks) (02/21/90)
Many of you are using my UUENCODE/DECODE (latest version 3.16). I am now finishing the latest version which will include an XXencode/decode capability (as well as UU..). I am having some problems with automatic detection of XXencoded files as compared to error conditions. I expect the new version will be available in a month. The advantage of my UUdecode is that it will accept multiple files and has all sorts of code to accept the little glitches that all the varities of UUencode throw out. It is written in Turbo Pascal for MS-DOS. My UUdecode also does the crc checks compatibly with c.b.i.p postings. The advantage of XXencode over UUencode is that XXencode uses a character set that does not get corrupted on ASCII to EBCDIC conversion. This character set is the upper and lower case characters, the numbers, and the plus and minus sign. This creates problems when trying to programmatially distinguish text at the start of a file from encoded data. The best "encode" solution is possibly Brad Templeton's ABE. (I have not used this yet, I have only read about it.) Regards to all, Richard Marks rmarks@KSP.unisys.COM
swh@hpcupt1.HP.COM (Steve Harrold) (02/22/90)
>>> Many of you are using my UUENCODE/DECODE (latest version >>> 3.16). I am now finishing the latest version which will >>> include an XXencode/decode capability (as well as UU..). Will you be able to supply an equivalent for the Unix environment so that we can shuttle files back and forth, like "arc" and "zoo" between environments?
w8sdz@smoke.BRL.MIL (Keith Petersen) (02/23/90)
[Richard Marks comments about UU/XX encode/decode programs]
Richard, one of the nicest variations I've seen on Usenet is the one
that includes a header just before the begin 644, called "table". It
uses that key word on the left margin and the next two lines include
the actual ascii characters used for the encoding, in progressing
numerical order. This is used by many binary newsgroups on Usenet.
The advantage is that the decode can automatically select the character
set, and error traps can be put in to exit if a character appears more
than once in the table (a sure indication of EBCDIC->ASCII conversion
errors).
The multi-part combining and error checking features of your Turbo
Pascal MSDOS program might become more popular if you were to release
a C language version which could be ported to Unix. Even the TP sources
would be helpful.
This is one good example why withholding sources is detrimental to a
program becoming a popular world standard. Example: it's likely that
Zmodem will become extinct because the sources for the latest version
are no longer freely distributable. Users will not accept being locked
in to a propriatary protocol. That means there is no "second source"
available in case the primary supplier is unable to deliver or goes out
of business.
Keith
--
Keith Petersen
Maintainer of SIMTEL20's CP/M, MSDOS, & MISC archives [IP address 26.2.0.74]
Internet: w8sdz@WSMR-SIMTEL20.Army.Mil, w8sdz@brl.arpa BITNET: w8sdz@NDSUVM1
Uucp: {ames,decwrl,harvard,rutgers,ucbvax,uunet}!wsmr-simtel20.army.mil!w8sdz