home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Columbia Kermit
/
kermit.zip
/
newsgroups
/
misc.19971216-19980424
/
000191_news@newsmaster….columbia.edu _Fri Feb 6 10:05:54 1998.msg
< prev
next >
Wrap
Internet Message Format
|
2020-01-01
|
8KB
Return-Path: <news@newsmaster.cc.columbia.edu>
Received: from newsmaster.cc.columbia.edu (newsmaster.cc.columbia.edu [128.59.35.30])
by watsun.cc.columbia.edu (8.8.5/8.8.5) with ESMTP id KAA27018
for <kermit.misc@watsun.cc.columbia.edu>; Fri, 6 Feb 1998 10:05:54 -0500 (EST)
Received: (from news@localhost)
by newsmaster.cc.columbia.edu (8.8.5/8.8.5) id KAA05819
for kermit.misc@watsun; Fri, 6 Feb 1998 10:05:53 -0500 (EST)
Path: news.columbia.edu!watsun.cc.columbia.edu!fdc
From: fdc@watsun.cc.columbia.edu (Frank da Cruz)
Newsgroups: comp.protocols.kermit.misc
Subject: Re: Expect and Kermit (was: Re: frequent timeouts!)
Date: 6 Feb 1998 15:05:53 GMT
Organization: Columbia University
Lines: 137
Message-ID: <6bf8sh$5mf$1@apakabar.cc.columbia.edu>
References: <6ao98e$p4h$1@apakabar.cc.columbia.edu> <ncj3ei1redz.fsf@nytrdc058.eq.gs.com> <6b5asg$ia1$1@apakabar.cc.columbia.edu> <m3oh0oclbc.fsf@idt.net>
NNTP-Posting-Host: watsun.cc.columbia.edu
Xref: news.columbia.edu comp.protocols.kermit.misc:8374
In article <m3oh0oclbc.fsf@idt.net>,
Russell D. McManus <mcmanus@idt.net> wrote:
: fdc@watsun.cc.columbia.edu (Frank da Cruz) writes:
: > Also, libraries are not portable; the very library mechanism is not
: > portable.
:
: hmm. of course you are right that the library mechanism is not
: completely portable. then again, neither is the 'redirect' command
: that appears in c-kermit. this fact tells me that lack of portability
: is not a sufficient condition to doom an idea as far as kermit is
: concerned. the question is how 'unportable' must something be to be
: rejected?
:
Clearly there is more than one criterion for what gets into Kermit.
I don't think it will serve any useful purpose to try to formalize the
algorithm for how features are selected for adding to Kermit, because,
quite simply, it isn't very formal. But portability is always a major
factor -- portability not only of command and scripting features for
the end-user, but also portability (sharability) of code that we write
to implement those features, so it can be shared by Kermit programs for
many platforms. How many software makers these days pay attention
to any platform but Windows?
: > Increased use of Kermit overall does not pay the bills of the Kermit
: > Project, so what good has been done if everybody is using "Kermit" in some
: > form, but there is nobody here to answer their questions and keep up with
: > their demands -- which is pretty much what we do all day (and night).
:
: and this is the real reason. funding is a necessary condition for the
: kermit project to continue. that's why i own both editions of
: 'programming c-kermit' (excellent book, frank).
:
Thanks :-)
: a zillion non portable things must have been done for kermit-95, but the
: program has commercial value.
:
In fact, the Kermit 95 source code is portable to Windows 95, Windows 98,
Windows NT 3.51, Windows NT 4.0, and OS/2 2.0, 2.1, 3.0, and 4.0.
: ... that is of course what drives the decision making process.
:
Many things drive it. Demand, especially from the large installed base of
users who depend on us to continually adapt our software to new environments
and requirements as they appear, so they can keep using the communication and
automation tools they have become accustomed to.
: > The bare fact is, there *is* no "core" functionality. No two people will
: > ever agree on what constitutes the core; each one will want one more
: > function added in. So before you know it, the core is the whole thing.
: > And that begs the question of the "API" -- what is the API to a program
: > that has 100,000 functions?
:
: does kermit have 100,000 functions? somehow i think you are exagerating.
:
Busted! OK, I didn't actually count them. Type ? at the prompt. Then after
each top level keyword, type ?. Then after each second-level keyword, type ?.
And so on, as far as it goes -- sometimes pretty far, especially in the SET
command. And most especially in K95, when you factor in the 35 terminal
emulators and 43 character sets.
: ... besides, it doesn't matter that everyone doesn't agree
: what is in the core. the kermit project gets to decide; everyone else
: then becomes reality challenged if they disagree.
:
What's the point? If you don't like it, you won't purchase it, and we will
have wasted all that time for nothing. This is the 90's, the harsh decade of
"every tub on its own bottom" -- whatever we do has to pay for itself.
Anyway, think of all the other items that people want:
. GUI ("100,000" dialog boxes)
. tn3270 / 5250
. Tektronix emulation
. Built-in LZW-like compression
. Security (Kerberos, SSL, SSH, SOCKS, ...)
. Chinese/Japanese/Korean
. Learned scripts
That's just handful picked at random. Which of these should we sacrifice to
produce a "library" (in quotes because this word means something different
to each person) that few people will use or pay for?
Given the small number of people we can afford to pay to work on Kermit
software, we have to pick our work carefully. One of the great issues of the
age is "style versus substance" -- what the software actually does versus its
interface. No interface will please everybody; we understand that and live
with it. There is a neverending demand for different kinds of interfaces --
"Buzzword 1.0 compliance". We are not equipped to deal with that -- so
instead, we supply *one* interface. It's one that we can support. If you
have a problem with it, we can get you past it. If it has a bug, we can fix
it. We do not have the resources to support Perl, Tcl, Expect, Rexx, DCL,
etc, nor the numerous APIs that are requested from us every day (Delphi, DLL,
OCX, OLE, VBX, Active X, Netscape, and on and on).
: i don't think 'library' qualifies as a 'Buzzword', and if it does,
: then it would certainly have a higher rev than 1.0. the biggest
: platforms support the library concept: all flavors of unix, win95,
: win-nt, os2, and surely others that i am not personally familiar with.
:
But what is the interface to the library? Come sit in our chairs for a while
and listen to the conflicting demands. Company A wants "just the protocol"
because they already handle the communications functions themselves, whereas
Company B wants the communications and they're not interested in the protocol.
Meanwhile, Company C wants the VT320 terminal emulator (but not the Wyse or
Data General), whereas Company D wants the character-set support and could not
care less about the other stuff. Then Company E comes along and wants --
would you believe -- the command parser.
Meanwhile, to which "standard" must the library "conform" -- Delphi, DLL, OCX,
OLE, VBX, Active X, Netscape -- 1.0, 2.0, 3.0, 4.0, ... -- these all have
different structures, different conventions, different requirements.
: i'm still confused about one thing: why is it impossible for the
: kermit project to make design decisions about what should go in a
: library?
: ...
: how about the same amount as the current version of kermit; pay
: for a copy of the book if you want support?
:
The amount of work that would have to into a "library" that would satisfy
even 10% of the people who are asking for such a thing is staggering, and
there are no indications that anybody would pay the kind of money it would
take to cover the development costs.
: and there is something to be gained: a quick tour through the
: quotation rules in the kermit language demonstrates how difficult it
: is to have a single syntax that is useful for both interaction and
: programming.
:
Only when you need to refer to DOS pathnames :-)
: ... one encounters similar problems when programming in a unix shell
: language. don't fault me for dreaming of something better!
:
Of course not. But who will do the work, and who will pay them to do it?
- Frank