diff --git a/.gitignore b/.gitignore
index 5d91ad6e..2646fb98 100644
--- a/.gitignore
+++ b/.gitignore
@@ -33,6 +33,7 @@
 /info/Transportation.t
 /info/Updates.t
 /info/TOP.t
+/info/stamp-subj
 # $(distclean)
 /config.h
 /config.log
diff --git a/Make.mk b/Make.mk
index 6b14d04e..d34ad884 100644
--- a/Make.mk
+++ b/Make.mk
@@ -130,7 +130,7 @@ ttop := info/TOP.t
 # Formatted info:
 info.nr := $(addprefix info.nr/, $(info))
 info.html := $(addprefix info.html/, $(addsuffix .html, $(info)))
-info.all := $(info.nr) $(info.html) info.ps
+info.all := $(info.nr) $(info.html) info.ps info/stamp-subj
 # Tests
 # sandbox
 
@@ -322,16 +322,19 @@ $(libs) $(empth_lib):
 
 # Info formatting
 
-# FIXME Remaking subjects doesn't work correctly when info sources get
-# removed or subjects get dropped.
+# mksubj.pl reads $(tsrc) and writes subjects.mk $(tsubj).  The naive
+# rule
+#     subjects.mk $(ttop) $(tsubj): $(tsrc)
+#           COMMAND
+# runs COMMAND once for each target.  That's because multiple targets
+# in an explicit rule is just a shorthand for one rule per target,
+# each with the same prerequisites and commands.  Use a stamp file.
+subjects.mk $(tsubj): info/stamp-subj ;
+info/stamp-subj: info/mksubj.pl $(tsrc)
+	$(call quiet-command,perl $(srcdir)/info/mksubj.pl $(filter %.t, $^),GEN '$$(subjects)')
+	>$@
 
-subjects.mk: info/findsubj.pl $(tsrc)
-	perl $(srcdir)/info/findsubj.pl $(filter %.t, $^)
-
-$(tsubj): info/mksubj.pl
-	perl $(srcdir)/info/mksubj.pl $@ $(filter %.t, $^)
-
-$(ttop): info/mktop.pl
+$(ttop): info/mktop.pl $(tsrc)
 	$(call quiet-command,perl $(srcdir)/info/mktop.pl $@ $(subjects),GEN $@)
 
 info.nr/all: $(filter-out info.nr/all, $(info.nr))
diff --git a/info/findsubj.pl b/info/findsubj.pl
deleted file mode 100644
index 66275189..00000000
--- a/info/findsubj.pl
+++ /dev/null
@@ -1,204 +0,0 @@
-#!/usr/bin/perl
-#
-#   Empire - A multi-player, client/server Internet based war game.
-#   Copyright (C) 1986-2013, Dave Pare, Jeff Bailey, Thomas Ruschak,
-#                 Ken Stevens, Steve McClure, Markus Armbruster
-#
-#   Empire is free software: you can redistribute it and/or modify
-#   it under the terms of the GNU General Public License as published by
-#   the Free Software Foundation, either version 3 of the License, or
-#   (at your option) any later version.
-#
-#   This program is distributed in the hope that it will be useful,
-#   but WITHOUT ANY WARRANTY; without even the implied warranty of
-#   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-#   GNU General Public License for more details.
-#
-#   You should have received a copy of the GNU General Public License
-#   along with this program.  If not, see <http://www.gnu.org/licenses/>.
-#
-#   ---
-#
-#   See files README, COPYING and CREDITS in the root of the source
-#   tree for related information and legal notices.  It is expected
-#   that future projects/authors will amend these files as needed.
-#
-#   ---
-#
-#   findsubj.pl: Find info subjects, update subjects.mk
-#
-#   Known contributors to this file:
-#      Ken Stevens (when it was still info.pl)
-#      Markus Armbruster, 2006-2008
-#
-
-# Usage: findsubj.pl INFO-FILE...
-# Run it at the root of the build tree.  This updates the make include
-# file subjects.mk, which guides the remaking of info index files.
-#
-#       --- Global variables ---
-# @Subjects       Existing subjects
-# $filename       The name of the current info file
-# $filename{TOPIC}
-#                 TOPIC's info file name
-# $chapter{TOPIC} TOPIC's chapter (first arg to .TH)
-# $see_also{TOPIC}
-#                 TOPIC's SEE ALSO items (.SA argument)
-# $sanr{TOPIC}    Line number of TOPIC's .SA request
-# $subjfil{SUBJECT}
-#                 info files for SUBJECT separated by space
-#
-#     --- File handles ---
-# F               Filehandle for info page sources and makefiles
-#
-#     --- Functions ---
-#
-# read_make_var   Read a variable value from a makefile
-# parse_file      Read an info file
-# parse_see_also  Create %subjfil from %see_also
-# error           Print an integrity error to STDERR and exit with code 1.
-
-use strict;
-use warnings;
-
-use Errno qw(ENOENT);
-
-my (%filename, %chapter, %see_also, %sanr);
-my ($filename, %subjfil);
-
-# Get known subjects
-my @Subjects = split(' ', read_make_var("subjects", "subjects.mk", ""));
-
-# Parse the .t files
-for my $f (@ARGV) {
-    parse_file("$f");
-}
-
-# Create %subjfil from %see_also
-for my $t (sort keys %see_also) {
-    parse_see_also($t);
-}
-
-# Update @Subjects from %subjfil
-for my $t (@Subjects) {
-    print STDERR "WARNING: The subject $t has been removed.\n"
-	unless exists $subjfil{$t};
-}
-for my $t (keys %subjfil) {
-    unless (grep(/^$t$/, @Subjects)) {
-	print STDERR "WARNING: $t is a NEW subject\n";
-	my $fname = "info/$t.t";
-	if (-e $fname) {
-	    print STDERR "File $fname exists\n";
-	    exit 1;
-	}
-    }
-}
-@Subjects = sort keys %subjfil;
-
-# Update subjects.mk
-open(F, ">subjects.mk")
-    or die "Can't open subjects.mk for writing: $!";
-print F "# DO NOT EDIT THIS FILE.  It was automatically generated by findsubj.pl\n";
-print F "subjects := " . join(' ', @Subjects) . "\n";
-for my $t (@Subjects) {
-    print F "info/$t.t:$subjfil{$t}\n";
-}
-close(F);
-
-exit 0;
-
-# Read a variable value from a makefile
-sub read_make_var {
-    my ($var, $fname, $dflt) = @_;
-    my $val;
-
-    unless (open(F, "<$fname")) {
-	return $dflt if $! == ENOENT and defined $dflt;
-	die "Can't open $fname: $!";
-    }
-    while (<F>) {
-	if (/^[ \t]*\Q$var\E[ \t]:?=*(.*)/) {
-	    $val = $1;
-	    last;
-	}
-    }
-    close(F);
-    $val or die "Can't find $var in $fname";
-    return $val;
-}
-
-# Read an info file
-# Parse .TH into %chapter and .SA into %see_also, %sanr
-sub parse_file {
-    ($filename) = @_;
-    my $topic;
-
-    $topic = $filename;
-    $topic =~ s,.*/([^/]*)\.t$,$1,;
-    $filename{$topic} = $filename;
-
-    open(F, "<$filename")
-	or die "Can't open $filename: $!";
-
-    $_ = <F>;
-    if (/^\.TH (\S+) (\S.+\S)$/) {
-	$chapter{$topic} = $1;
-    } else {
-	error("The first line in the file must be a .TH request");
-    }
-
-    while (<F>) {
-	last if /^\.SA/;
-    }
-
-    if ($_) {
-	if (/^\.SA "([^\"]*)"/) {
-	    $see_also{$topic} = $1;
-	    $sanr{$topic} = $.;
-	} else {
-	    error("Incorrect .SA Syntax.  Syntax should be '.SA \"item1, item2\"'");
-	}
-
-	while (<F>) {
-	    error("Multiple .SA requests.  Each file may contain at most one.") if /^\.SA/;
-	}
-    } else {
-	error(".SA request is missing");
-    }
-
-    close F;
-}
-
-# Create %subjfil from %see_also
-sub parse_see_also {
-    my ($topic) = @_;
-    my @see_also = split(/, /, $see_also{$topic});
-    my $wanted = $chapter{$topic};
-    my $found;		       # found a subject?
-
-    $wanted = undef if $wanted eq 'Concept' or $wanted eq 'Command';
-    $filename = $filename{$topic};
-
-    for (@see_also) {
-	if (!exists $see_also{$_}) { # is this entry a subject?
-	    $subjfil{$_} .= " info/$topic.t";
-	    $found = 1;
-	}
-	if ($wanted && $_ eq $wanted) {
-	    $wanted = undef;
-	}
-    }
-
-    $. = $sanr{$topic};
-    error("No subject listed in .SA") unless $found;
-    error("Chapter $wanted not listed in .SA") if $wanted;
-}
-
-# Print an integrity error message and exit with code 1
-sub error {
-    my ($error) = @_;
-
-    print STDERR "findsubj.pl:$filename:$.: $error\n";
-    exit 1;
-}
diff --git a/info/mksubj.pl b/info/mksubj.pl
index 05d846ac..48e589e0 100644
--- a/info/mksubj.pl
+++ b/info/mksubj.pl
@@ -25,98 +25,108 @@
 #
 #   ---
 #
-#   mksubj.pl: Create the index for a subject
+#   mksubj.pl: Create the subject index pages
 #
 #   Known contributors to this file:
 #      Ken Stevens (when it was still info.pl)
-#      Markus Armbruster, 2006
+#      Markus Armbruster, 2006-2013
 #
-
-# Usage: mksubj.pl OUTFILE INFILE...
-# The INFILE... contain all the topics belonging to a subject.  Read
-# and check the information required for the index from them, write
-# the index to OUTFILE.
+# Usage: mksubj.pl INFO-FILE...
+#
+# Read the INFO-FILE..., read and update subjects.mk, create
+# info/SUBJECT.t for each SUBJECT.
 
 use strict;
 use warnings;
 use File::stat;
 
+use Errno qw(ENOENT);
+use Fcntl qw(O_WRONLY O_EXCL O_CREAT);
+
 # The chapters, in order
 my @Chapters = qw/Introduction Concept Command Server/;
 
 my @Levels = qw/Basic Expert Obsolete/;
+my @Subjects;
 
+# $filename{TOPIC} is TOPIC's file name
+my %filename;
+# $long{TOPIC} is true when TOPIC's page is "long"
+my %long;
+# $chapter{TOPIC} is TOPIC's chapter (first arg to .TH)
+my %chapter;
+# $desc{TOPIC} is a one line description of TOPIC (second arg to .NA)
+my %desc;
+# $level{TOPIC} is TOPIC's difficulty level (arg to .LV)
+my %level;
+# $see_also{TOPIC} is TOPIC's list of SEE ALSO items (.SA argument)
+my %see_also;
+# $sanr{TOPIC} is the line number of TOPIC's .SA request
+my %sanr;
+
+# current info file
 my $filename;
-my (%subject, %level, %desc, %long, %cnt);
-my $largest = "";
 
-my $out = shift @ARGV;
-$out =~ /([^\/]*)\.t$/
-    or die "Strange subject file name $out";
-my $subj = $1;
+# $subject{$subj}{$chap} = "item1\nitem2\n..."
+#                 Topics in that subject organized by chapter.
+my %subject;
+# $largest{$sub}  The largest topic name in that subject (used for
+#                 column formatting)
+my %largest;
+
+@Subjects = split(' ', read_make_var("subjects", "subjects.mk", ""));
 
 for (@ARGV) {
-    my ($topic, $chap, $lvl, $desc, $long) = parse_file($_);
-    $largest = $topic if length $topic > length $largest;
-    $subject{$chap} .= "$topic\n";
-    $level{$topic} = $lvl;
-    $cnt{$lvl}++;
-    $desc{$topic} = $desc;
-    $long{$topic} = $long;
-    $cnt{'long'}++ if $long;
+    parse_file($_);
 }
 
-open(SUBJ, ">$out")
-    or die "Can't open $out for writing: $!";
+for my $t (sort keys %desc) {
+    parse_see_also($t);
+}
 
-print SUBJ '.\" DO NOT EDIT THIS FILE.  It was automatically generated by mksubj.pl'."\n";
-print SUBJ ".TH Subject \U$subj\n";
-$largest =~ s/-/M/g;
-print SUBJ ".in \\w'$largest", "XX\\0\\0\\0\\0'u\n";
+@Subjects = create_subjects();
 
-for my $chap (@Chapters) {
-    next unless exists $subject{$chap};
-    print SUBJ ".s1\n";
-    for (split(/\n/, $subject{$chap})) {
-	my $flags = "";
-	$flags .= "*" if $level{$_} eq 'Basic';
-	$flags .= "+" if $level{$_} eq 'Obsolete';
-	$flags .= "!" if $long{$_};
-	$flags = sprintf("%-2s", $flags);
-	print SUBJ ".L \"$_ $flags\"\n";
-	print SUBJ "$desc{$_}\n";
+open(F, ">subjects.mk")
+    or die "Can't open subjects.mk for writing: $!";
+print F "subjects := " . join(' ', @Subjects) . "\n";
+close(F);
+
+exit 0;
+
+# Read a variable value from a makefile
+sub read_make_var {
+    my ($var, $fname, $dflt) = @_;
+    my $val;
+
+    unless (open(F, "<$fname")) {
+	return $dflt if $! == ENOENT and defined $dflt;
+	die "Can't open $fname: $!";
     }
+    while (<F>) {
+	if (/^[ \t]*\Q$var\E[ \t]*:?=[ \t]*(.*)/) {
+	    $val = $1;
+	    last;
+	}
+    }
+    close(F);
+    defined($val) or die "Can't find $var in $fname";
+    return $val;
 }
 
-print SUBJ <<EOF;
-.s1
-.in 0
-For info on a particular subject, type "info <subject>" where <subject> is
-one of the subjects listed above.
-EOF
-print SUBJ <<EOF if $cnt{'Basic'};
-Subjects marked by * are the most important and should be read by new
-players.
-EOF
-print SUBJ <<EOF if $cnt{'Obsolete'};
-Subjects marked by + are obsolete.
-EOF
-print SUBJ <<EOF if $cnt{'long'};
-Unusually long subjects are marked with a !.
-EOF
-close SUBJ;
-
-
+# Parse an info file
+# Set $filename, $filename{TOPIC}, $long{TOPIC}, $chapter{TOPIC},
+# $desc{TOPIC}, $level{TOPIC}, $see_also{TOPIC}, $sanr{TOPIC}
 sub parse_file {
     ($filename) = @_;
-    my ($topic, $chap, $lvl, $desc, $long, $st);
+    my ($topic, $st);
 
     $topic = $filename;
     $topic =~ s,.*/([^/]*)\.t$,$1,;
+    $filename{$topic} = $filename;
 
     $st = stat $filename
 	or die "Can't stat $filename: $!";
-    $long = $st->size > 9999;
+    $long{$topic} = $st->size > 9999;
 
     open(F, "<$filename")
 	or die "Can't open $filename: $!";
@@ -126,7 +136,7 @@ sub parse_file {
 	if (!grep(/^$1$/, @Chapters)) {
 	    error("First argument to .TH was '$1', which is not a known chapter");
 	}
-	$chap = $1;
+	$chapter{$topic} = $1;
 	if ($1 eq "Command" && $2 ne "\U$topic") {
 	    error("Second argument to .TH was '$2' but it should be '\U$topic'");
 	}
@@ -139,7 +149,7 @@ sub parse_file {
 	if ($topic ne $1) {
 	    error("First argument to .NA was '$1' but it should be '$topic'");
 	}
-	$desc = $2;
+	$desc{$topic} = $2;
     } else {
 	error("The second line in the file must be a .NA request");
     }
@@ -149,14 +159,136 @@ sub parse_file {
 	if (!grep(/^$1$/, @Levels)) {
 	    error("The argument to .LV was '$1', which is not a known level");
 	}
-	$lvl = $1;
+	$level{$topic} = $1;
     } else {
 	error("The third line in the file must be a .LV request");
     }
 
-    close F;
+    while (<F>) {
+	last if /^\.SA/;
+    }
 
-    return ($topic, $chap, $lvl, $desc, $long);
+    if ($_) {
+	if (/^\.SA "([^\"]*)"/) {
+	    $see_also{$topic} = $1;
+	    $sanr{$topic} = $.;
+	} else {
+	    error("Incorrect .SA Syntax.  Syntax should be '.SA \"item1, item2\"'");
+	}
+
+	while (<F>) {
+	    error("Multiple .SA requests.  Each file may contain at most one.") if /^\.SA/;
+	}
+    } else {
+	error(".SA request is missing");
+    }
+
+    close F;
+}
+
+# Create %subject and %largest from %see_also
+sub parse_see_also {
+    my ($topic) = @_;
+    my @see_also = split(/, /, $see_also{$topic});
+    my $wanted = $chapter{$topic};
+    my $found;		       # found a subject?
+
+    $wanted = undef if $wanted eq 'Concept' or $wanted eq 'Command';
+
+    for (@see_also) {
+	if (!exists $desc{$_}) { # is this entry a subject?
+	    set_subject($_, $topic);
+	    $found = 1;
+	}
+	if ($wanted && $_ eq $wanted) {
+	    $wanted = undef;
+	}
+    }
+
+    $filename = $filename{$topic};
+    $. = $sanr{$topic};
+    error("No subject listed in .SA") unless $found;
+    error("Chapter $wanted not listed in .SA") if $wanted;
+}
+
+# Add a new entry to %subject and possibly to %largest
+sub set_subject {
+    my ($sub, $topic) = @_;
+    my $chap = $chapter{$topic};
+    $subject{$sub}{$chap} .= "$topic\n";
+    $largest{$sub} = "" unless defined $largest{$_};
+    $largest{$sub} = $topic if length $topic > length $largest{$sub};
+}
+
+# Create a Subject.t file
+sub create_subj {
+    my ($subj) = @_;
+    my $fname = "info/$subj.t";
+    my ($any_basic, $any_obsolete, $any_long);
+
+    print "WARNING: $subj is a NEW subject\n"
+	unless grep(/^$subj$/, @Subjects);
+    sysopen(SUBJ, $fname, O_WRONLY | O_EXCL | O_CREAT)
+	or die "Unable to create $fname: $!\n";
+
+    print SUBJ '.\" DO NOT EDIT THIS FILE.  It was automatically generated by mksubj.pl'."\n";
+    print SUBJ ".TH Subject \U$subj\n";
+    $largest{$subj} =~ s/-/M/g;
+    print SUBJ ".in \\w'$largest{$subj}XX\\0\\0\\0\\0'u\n";
+    for my $chap (@Chapters) {
+	next unless exists $subject{$subj}{$chap};
+	print SUBJ ".s1\n";
+	for my $topic (split(/\n/, $subject{$subj}{$chap})) {
+	    my $flags = "";
+	    if ($level{$topic} eq 'Basic') {
+		$flags .= "*";
+		$any_basic = 1;
+	    }
+	    if ($level{$topic} eq 'Obsolete') {
+		$flags .= "+";
+		$any_obsolete = 1;
+	    }
+	    if ($long{$topic}) {
+		$flags .= "!";
+		$any_long = 1;
+	    }
+	    $flags = sprintf("%-2s", $flags);
+	    print SUBJ ".L \"$topic $flags\"\n";
+	    print SUBJ "$desc{$topic}\n";
+	}
+    }
+    print SUBJ ".s1\n"
+	. ".in 0\n"
+	. "For info on a particular subject, type \"info <subject>\" where <subject> is\n"
+	. "one of the subjects listed above.\n";
+    print SUBJ "Subjects marked by * are the most important and should be read by new players.\n"
+	if $any_basic;
+    print SUBJ "Subjects marked by + are obsolete.\n"
+	if $any_obsolete;
+    print SUBJ "Unusually long subjects are marked with a !.\n"
+	if $any_long;
+    close SUBJ;
+}
+
+# Remove the old Subject.t files and create new ones
+sub create_subjects {
+    my (@subj);
+
+    for (@Subjects) {
+	unlink "info/$_.t";
+    }
+
+    @subj = sort keys %subject;
+
+    for my $subj (@Subjects) {
+	print "WARNING: The subject $subj has been removed.\n"
+	    unless grep (/^$subj$/, @subj);
+    }
+
+    for my $subj (@subj) {
+	create_subj($subj);
+    }
+    return @subj;
 }
 
 # Print an integrity error message and exit with code 1