Path: utzoo!utgpu!watmath!clyde!att!osu-cis!tut.cis.ohio-state.edu!unmvax!unm-la!lanl!beta!ttp From: ttp@beta.lanl.gov (T T Phillips) Newsgroups: comp.binaries.ibm.pc.d Subject: Re: Thanks to everyone who helped with SIMTEL20 Message-ID: <22924@beta.lanl.gov> Date: 7 Dec 88 20:23:00 GMT References: <1383@virginia.acc.virginia.edu> <5616@sdcsvax.UCSD.EDU> Organization: Los Alamos National Laboratory Lines: 36 In article <5616@sdcsvax.UCSD.EDU>, holtz@odin.ucsd.edu (Fred Holtz) writes: > In article <1383@virginia.acc.virginia.edu> lhb6v@watt.acc.Virginia.EDU (Laura H. Burchard) writes: > >my local UNIX mainframe, and from there to m PC. For anybody > >else who has the same problem, the procedure is this: > > > >ftp to SIMTEL20 > ... > > This procedure will work, but sending a file using binary kermit is slower > than uuencoding the file, sending it via text kermit, then uudecoding on > the PC side. The uuencoding increases file size about 50%, but it has been > my experience that binary kermit file transfers are > 90% longer than text > mode transfers. I don't know why kermit has such a large overhead for binary > file transfers; any kermit gurus out there that can clue me in?? I have had the exact same experience, but I always assumed the slow transfers had something to do with the elaborate multiplexing, encryption, decryption, de-multiplexing link up that our communications go through. Perhaps the slow transfer is a characteristic of all error checking protocols. The uuencoding-ascii transfer scheme works great for downloads, but I haven't been able to find an equivalent scheme for uploading. It would be very nice to have a small unix routine that would set up a large memory buffer area to accept high speed uploads of ascii files. I know that if I try to upload this way into a text editor, I quickly overload the input buffer and start losing characters. Obviously this would only work for highly reliable links, but that would be fine for me. I think that at least 99% of the binary files that I download as uuencoded ascii arrive perfectly. Does anybody out there have an uploading program such as I described? Terry Phillips Los Alamos National Laboratory ttp@beta.lanl.gov