LSTSRV-L Archives

LISTSERV Site Administrators' Forum

LSTSRV-L

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Topic: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
"Tansin A. Darcos & Company" <[log in to unmask]>
Fri, 3 Sep 1993 11:05:24 -0400
text/plain (78 lines)
From: Paul Robinson <[log in to unmask]>
Organization: Tansin A. Darcos & Company, Silver Spring, MD USA
-----
Eric Thomas <[log in to unmask]>, writes:
 
> It appears that high level managers are now making statements which
> one would ordinarily expect from young technical people who think
> they know the answer to every problem in the world,
 
You think this is new?  In the 1959 book "Starship Troopers" by
Robert A. Heinlein, there is a line where a sargent says to his
commander, "These kids are all fired up, and think they know
everything; that's okay, I used to think that myself."
 
> such as "it won't be useful unless it is written in C",
 
There are reports in the media that some of the worst time overruns
have been caused by development in the C language.  Lotus 123 had
one of its versions written in C - not sure if it was 2.2 or 3.0, it
was probably the 2.2 and sufferred from terrible schedule slippages
and cost overruns.
 
> "if it's written in PASCAL it can't possibly be any good",
 
You might want to ask some people how long it takes to debug a sendmail.cf
script and yet allow mail service.  Listserv and (the companion program
that handles mail, is it L-Mail?) have been in use at hundreds of sites
by hundreds of thousands of recipients of bitnet lists.  The requests
I've heard have been for upgrades, I don't remember hearing about errors
in the code; a reliability issue can be raised.   After all, it was the
beloved sendmail that had the debugging feature that allowed Robert T.
Morris Jr. to almost shut down the Internet.
 
How long does it take to setup a listserv?  Probably less time than it
takes to install and configure sendmail.
 
> "anything derived from  a mainframe-influenced  design  won't  be
> useful  to  the  Internet  community",  and so on.
 
IBM had TCP/IP running on 370 mainframes long before anyone had ever
even heard the name "Sun Microsystems."  There was a big discussion about
just this issue in the Com-Priv (Privatization of the Internet) List.
 
> As a computer professional I found this deeply depressing, but of
> course it is a serious business problem and it needs to be
> addressed as such.
 
It's not a "serious" issue.  A serious issue is one where there is
a reason for doing something.  This is an issue of stupidity and
ignorance.  Hopefully dinosaurs that think that way will become
extinct, when the asteroid of reality comes crashing down on their
environment.
 
> While there will always be people arguing to their management that
> LISTSERV should not be used because it is written in PASCAL, we are
> confident that proven quality will prevail over religious
> arguments.
 
Well, there is a translator program called "p2c" - I have a copy of
it myself - that translates Pascal code into the C language.  Some
hand tweaking has to be done, but most standard pascal will go through
with little or no trouble.  Try doing an archie search if you're
interested, or it might be on the SIMTEL20 group.
 
Here's one to answer people with: try setting up a unix-based sendmail
system, installed by an expert, versus a Listserv - converted to C
for this test if the site doesn't have a Pascal Compiler - installed by a
novice manager.  Now have both systems be bombarded with mail, and see
which of them has problems keeping up with the load.  My guess is that
sendmail would crack long before listserv does.
 
They did this before with competing TCP/IP products.  You could use the
same term, call it the "listserv vs. sendmail bake off."
 
 
---
Paul Robinson - [log in to unmask]

ATOM RSS1 RSS2