On Sat, 26 May 2007 10:49:05 -0400, Gerhard Postpischil <[EMAIL PROTECTED]> wrote:
> >That's great if it works for you, but under most conditions the >ASCII-EBCDIC translation is lossy, and fails to preserve all >characters. Unless the text needs to be processed on the PC >side, downloading in BINary mode is preferred, and lossless. > >Gerhard Postpischil >Bradford, VT > Everyone has their own requirements. I find under most conditions people are just transferring normal text and the EBCDIC to ASCII translation loses nothing. And most transfers are so they can use the data on the PC so BINARY is not preferred. Fewer people transfer data with odd characters and need to be concerned. When you start looking at what system programmers and application programmers are moving around, much of that is still normal text. The oddball characters can sometimes be replaced. I stopped using the not-sign character back when I left my 3278 and started usng a terminal emulator. It was just too hard to find since shift-6 gave me a caret instead. A few others could be a pain and I developed an edit macro I would use to fix them. I have not had to do that for many years. If you can keep it simple then ASCII is fine. If you can't, then you must use BINARY. I even toyed with REXX code to convert an object deck into displayable characters that transfer just fine, then put it back into hex after I uploaded it. This is not something you want to do all the time. It doubles the size of the data you must move back and forth. ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
