As I was kinda bored after work today, I had a closer look at what I saw during my fuckup in the morning. Well, Steve said, that when he looked at metadata.xml it’d be “really common” .. still that isn’t making it right ..

There is a reason we do have a herds.xml (exactly for the reason to associate people with packages, and that’s what the <herd> tag is for in metadata.xml) file. So after a preliminary look through the repository, here are the winners:
700 : perl
126 : xemacs
63 : haskell
47 : sound
32 : ha-cluster
31 : crypto
19 : desktop-misc
16 : netmon
15 : forensics
13 : web-apps
8 : mips
8 : app-backup
7 : kde
6 : tcltk
6 : net-im
6 : media-tv
6 : dev-embedded
5 : voip
5 : theology
5 : samba
5 : net-p2p
4 : sparc
4 : java
4 : graphics
2 : net-mail
2 : kernel
2 : fonts
2 : embedded
2 : cpp
1 : x11
1 : wxwidgets
1 : www-servers
1 : tex
1 : shell-tools
1 : sh
1 : sgml
1 : sci
1 : python
1 : proaudio
1 : php
1 : media-optical
1 : kerberos
1 : hp-cluster
1 : gentopia
1 : amd64

Don’t know how accurate that list is, but you can check it for yourself. The commands I’ve used are these:

for i in $( < ../herds.list ); do
    grep --exclude=eclass --exclude=CVS \
    --exclude=profiles --exclude=skel.* -R \
    "$i@gentoo.org" /cvs/gentoo-x86/* ;
 done > redundant-metadata-xml.list
for i in $( < herds.list ); do
    echo -e " $( grep "$i@gentoo.org" \
                 ~/public_html/redundant-metadata-xml.list | \
                 wc -l )\t: $i"; done | \
     grep -v "^ 0" | \
     sort -nr > public_html/redundant-metadata-xml.overview

While herds.list holds a list (separated by \n) with all the herds there are. The raw files are here and here and here. Knock yourself out!