Received: by nummer-3.proteosys id <01C19443.9963CF04@nummer-3.proteosys>; Thu, 3 Jan 2002 11:44:20 +0100 In-Reply-To: Your message of Thu, 30 Jan 92 17:05:07 MST. <9201310039.AA12563@enet-gw.pa.dec.com> MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----_=_NextPart_001_01C19443.9963CF04" Return-Path: <@vm.gmd.de:LATEX-L@DHDURZ1.BITNET> X-MimeOLE: Produced By Microsoft Exchange V6.5 x-vm-v5-data: ([nil nil nil nil nil nil nil nil nil][nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil]) Content-class: urn:content-classes:message Subject: Re: LaTeX trip tests and checksums Date: Fri, 31 Jan 1992 03:02:04 +0100 Message-ID: X-MS-Has-Attach: X-MS-TNEF-Correlator: From: "Leslie Lamport" Sender: "LaTeX-L Mailing list" To: "Rainer M. Schoepf" Reply-To: "LaTeX-L Mailing list" Status: R X-Status: X-Keywords: X-UID: 563 This is a multi-part message in MIME format. ------_=_NextPart_001_01C19443.9963CF04 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Recall the TeX trip and METAFONT trap diffs, where architectural differences can produce small differences in the last few digits of large integers, making tedious manual inspection of the diffs by an expert human essential. I thought that TeX's use of fixed-point arithmetic eliminated this problem. Anyway, it is a minor problem for the official distribution, which can always be tested on the same machine. It is more of a problem for the style-file writer. Perhaps he needs to maintain his own version of the trip output. When a new release of LaTeX changes the official version of the trip output, the diff between the old and new trip output would tell the style-file writer what changes to expect when he compares the output of his style running on the new release with what it produced on the previous release. Also, my idea is that (except for stuff like error messages), one should be able to look at the printed output to see if things are working right. A discrepancy between the old and new .log files would alert someone to examine the appropriate portion of the output carefully. Leslie ------_=_NextPart_001_01C19443.9963CF04 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Re: LaTeX trip tests and checksums

   Recall the TeX trip and METAFONT trap = diffs, where architectural
   differences can produce small = differences in the last few digits of
   large integers, making tedious manual = inspection of the diffs by an
   expert human essential.

I thought that TeX's use of fixed-point arithmetic = eliminated this
problem.  Anyway, it is a minor problem for the = official distribution,
which can always be tested on the same machine.  = It is more of a
problem for the style-file writer.  Perhaps he = needs to maintain his
own version of the trip output.  When a new = release of LaTeX changes
the official version of the trip output, the diff = between the old and
new trip output would tell the style-file writer what = changes to expect
when he compares the output of his style running on = the new release
with what it produced on the previous release.

Also, my idea is that (except for stuff like error = messages), one
should be able to look at the printed output to see = if things are
working right.  A discrepancy between the old = and new .log files would
alert someone to examine the appropriate portion of = the output
carefully.

Leslie

------_=_NextPart_001_01C19443.9963CF04--