home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Columbia Kermit
/
kermit.zip
/
newsgroups
/
misc.20020314-20021006
/
000046_fdc@columbia.edu_Thu Apr 25 15:23:41 EDT 2002.msg
< prev
next >
Wrap
Text File
|
2020-01-01
|
6KB
|
148 lines
Article: 13337 of comp.protocols.kermit.misc
Path: newsmaster.cc.columbia.edu!news.columbia.edu!news-not-for-mail
From: fdc@columbia.edu (Frank da Cruz)
Newsgroups: comp.protocols.kermit.misc
Subject: Re: a bug on GNU/linux: speed reset to unintended value occasionally.
Date: 25 Apr 2002 15:23:29 -0400
Organization: Columbia University
Lines: 131
Message-ID: <aa9l3h$4s4$1@watsol.cc.columbia.edu>
References: <3CAFF81C.8039CBF8@yk.rim.or.jp> <3CC6E9D7.F4F2C624@yk.rim.or.jp> <aa6sjh$1a9$1@watsol.cc.columbia.edu> <3CC84CA6.D49F85EC@yk.rim.or.jp>
NNTP-Posting-Host: watsol.cc.columbia.edu
X-Trace: newsmaster.cc.columbia.edu 1019762612 18841 128.59.39.139 (25 Apr 2002 19:23:32 GMT)
X-Complaints-To: postmaster@columbia.edu
NNTP-Posting-Date: 25 Apr 2002 19:23:32 GMT
Xref: newsmaster.cc.columbia.edu comp.protocols.kermit.misc:13337
In article <3CC84CA6.D49F85EC@yk.rim.or.jp>,
Ishikawa <ishikawa@yk.rim.or.jp> wrote:
: > If you have a truly transparent 8-bit path, then you can
: > use "set parity none" (which is the default anyway, or
: > "set parity hardware" if you need it) and "set prefixing
: > minimal" or "set prefixing none" and get very close to
: > the maximum physical speed of the port; sometimes faster,
: > because Kermit also does some compression.
: >
: I tried your suggestion, and found
: "set prefixing none" somehow produces slightly lower
: throughput than "set control unprefixed all".
: Hmm. Could be an artifact or something.
:
They should be equivalent. Probably a statistical fluctuation.
: The real kicker is the use of "/transparent"
: as in
: send /binary /transparent ./wermit ./wermit.tmp
:
: Using the 8N1 connection at 38400 bps, I got
: about 3800 bps. This is close to the bare line speed.
: Isn't this great?
:
: I have been unaware of "/transparent" for a long time, but
: then I was interested in using KERMIT as terminal emulator
: and a `reliable' file transfer engine.
:
It should not make any different. It applies only to text-mode
transfers (it means "don't convert character sets"), and is
therefore ignored when you give it with /BINARY. Any difference
you observed must be another statistical fluctuation.
: But having done some coding for Japanese character processing,
: I tend to shy away from the complex code conversion thingy,
: especially since the underlying code standards may change
: from now and then. You are probably aware of the tome by Ken Lunde:
:
Yes, I have it. Very thick! And yes, CJK standards change all the
time. Soon the world will be filled up by Kanjis :-) For example,
special Kanjis for the names of race horses in Hong Kong. And
what about DoCoMos -- is this a new kind of Kanji? :-)
: > set file character-set shift-jis
: > set transfer character-set utf8
: > send <filename>
: >
: > On the Linux system:
: >
: > set file character-set euc-jp
: > receive
: >
: Thank you for the tips, I will try
: this later on when I am in need of such feature, but
: are you sure that the second topmost command is
: to specify "utf8"?
:
Yes. Actually, it could also have been "ucs2" or "euc-jp". You
can use any standard character set as intermediate, that contains
the characters used in your text; thus "ucs2", "utf8", and "euc-jp"
are all suitable.
: In any case, the inclusion of these capability into
: KERMIT must have been deemed necessary upon popular demands,
: and I hope you had a nice feedback from the users of KERMIT.
:
It comes from a conference we attended in Tokyo in 1987. We worked
with people from NTT, DEC Japan, and KEK on this and it was widely
used in Japan for some years. It makes sense to convert the
character set as part of the file transfer, and to do so using an
intermediate standard character set. This way, each program needs
to know only its own local character sets and the standard ones,
and not the character sets used by every other kind of computer.
When you consider all the Kanji encodings used on IBM, Hitachi,
Fujitsu, and other computers, this begins to make sense. Kermit
solved that problem more than 10 years ago. See:
http://www.columbia.edu/kermit/papers.html
To illustrate, you can use Kermit 95 on a PC to send a Shift-JIS
Kanji text file to a mainframe that uses IBM, Hitachi, or Fujitsu
"EBCDIC" Kanji (three different mainframe Kanji codes). By the
same principal, you can use C-Kermit on Linux to send an ISO2022JP
Kanji file to any of those mainframes. Or vice versa. How else
could you do such a thing?
Perhaps the need is not so great now, because since 1987 many
platforms and character sets have disappeared, so the "n" in
"n * n" is much smaller than before (for all practical purposes,
only Windows and Unix survive), and because Unicode was developed
to address the problem in another way: over time, all text will be
converted to Unicode. Of course Kermit knows Unicode too, and
therefore is part of this process:
http://www.columbia.edu/kermit/unicode.html
: I have a nagging feeling, though, that KERMIT is
: now suffering from "Everything except for the kitchen sink"
: syndrome. Emacs is often cited for this
: creeping featurism.
:
Yes, this is the fate of all popular software. In fact, for a
while, the icon for the Windows version of EMACS *was* a kitchen
sink ;-)
I think C-Kermit handles this better than many other packages, by
allowing itself to be built with any desired set of features
included or excluded, as described here:
http://www.columbia.edu/kermit/ckccfg.html
And of course we also supply a very small, bare-bones "lean and
mean" alternative, G-Kermit:
http://www.columbia.edu/kermit/gkermit.html
: PS: I have a feeling that KERMIT will live on until
: Frank retires from Columbia university
: and/or the last KERMIT devotee will die on the earth.
: (But then someone comes along and find a copy of
: KERMIT being used for monitoring remote router/remote
: scientific station, etc. and then that someone
: might become hooked and become a KERMIT user again...)
:
I confess, it can be fun. It's like a laboratory in which we
experiment with protocols and algorithms, and at the same time
produce something useful that helps people in real life, and
grows with them and with the times so those who like it don't
have to leave it behind.
- Frank