1997-09-19 - Re: Preparing the Remnant for the far side of the crisis

Header Data

From: Bill Frantz <frantz@netcom.com>
To: Tim May <cypherpunks@Algebra.COM
Message Hash: 690ee81c4bdd9670cc24ad7a9f7e9bfbfcea9f6043bdc39fd6b665422e4dfaae
Message ID: <v03007852b047c33e4e31@[]>
Reply To: <v0311070eb046c146ecf3@[]>
UTC Datetime: 1997-09-19 06:24:39 UTC
Raw Date: Fri, 19 Sep 1997 14:24:39 +0800

Raw message

From: Bill Frantz <frantz@netcom.com>
Date: Fri, 19 Sep 1997 14:24:39 +0800
To: Tim May <cypherpunks@Algebra.COM
Subject: Re: Preparing the Remnant for the far side of the crisis
In-Reply-To: <v0311070eb046c146ecf3@[]>
Message-ID: <v03007852b047c33e4e31@[]>
MIME-Version: 1.0
Content-Type: text/plain

At 10:03 AM -0700 9/18/97, Tim May wrote:
>There's simply no way that much code can be refactored to get the "00"
>problem fixed. It isn't, of course, just a matter of doing a
>search-and-replace on "00" and replacing it with "2000." For example, the
>date code was picked to be 2 digits (back in the 1950s and 60s and well
>into the 70s, 80s, and even 90s) to save space. Switching to 4-digit dates
>would require recompilation of the code (for which the compilers may not
>even run properly any longer), and the old hardware will of course not
>accept straightforwardly recompiled code (because part of the code "tucked"
>these compressed numbers into small registers, as but one of many examples).

Note also that the standard date format for OS/360 and its descendents is
the "Julian" date, a YYDDD format.  This means those old compilers have to
be changed to use a different system call to get the current date.

In 1969, my wife participated in a previous "expand the date" project for a
company in San Francisco.  They converted from one digit dates to two

Bill Frantz       | Internal surveillance      | Periwinkle -- Consulting
(408)356-8506     | helped make the USSR the   | 16345 Englewood Ave.
frantz@netcom.com | nation it is today.        | Los Gatos, CA 95032, USA