arch-artwork 
gimp-plugin-greycstoration  # (may actually not belong)
gkrellmpc 
gproftpd 
gq 
gstm 
hsoconnect  # This one may not belong.  SOmeone else should look.
icewm-themes 
icon-slicer 
inn 
libdlna 
libgdamm 
libgtksourceviewmm  # url from PKGBUILD redirects to real homepage [1]
libgtksourceviewmm2  # url from PKGBUILD redirects to real homepage [1]
libnewt 
lomoco  # url from PKGBUILD redirects to valid homepage [1]
man-pages-it  # Looks valid [2]
man-pages-ru  # See [2], but s/Italian/Russian/
mftrace  # See [1] below (redirects).
mpdcon 
nautilus-open-terminal 
openbox-themes 
openlh-client  # not sure.  looks valid to me.
openlh-server # as with openlh-client; looks valid
ozerocdoff 
prosper 
python-numarray 
python-pyparallel http://pyserial.wiki.sourceforge.net/pyParallel
python-pyserial http://pyserial.wiki.sourceforge.net/pySerial
python-pyxmpp 
qemu-launcher 
qmc2 
qt-recordmydesktop 
r8101 
recordmydesktop 
rexima 
rezero 
rlog 
scponly  # Redirects.  See [1]
sisctrl 
sniffit 
swarp 
terminus-font 
unshield  # a sourceforge project page, possibly valid
xchat-systray-integration  # someone should check me; url looks bogus
xdm-archlinux 
xmove 
xnc  # not sure; homepage *appears* valid

[1] Should this be considered invalid?  The link works, after all.
[2] The homepage is in Italian, which I don't read!  We can probably scratch
it from the list, but I'm being cautious.