1)chk_xmltexts.php lists up translatable texts 
  from xml files under target path.
  You can specify the path in chk_var.php .
  
2)chk_translations.php lists up translation candidates 
  from Babel database.
  Babel database requires the "latin1" encoding,
  not "UTF-8". And You can specify user id in chk_var.php .
  
  ("in my dev" at line 109 means I use tables removed some lines)

3)
  Each upd_*.php script is related to following plug-in.
  These scripts replace texts in xml files under conv_in folder,
  and output to conv_out.

  upd_atf_xmls.php  : org.eclipse.atf.doc
  upd_cdt_xmls.php  : org.eclipse.cdt.doc.user
  upd_git_xmls.php  : org.eclipse.egit.doc
  upd_mylyn_xmls.php    : org.eclipse.mylyn.wikitext.help.ui
  upd_platform_xmls.php : org.eclipse.platform.doc.user
  upd_svn_xmls.php  : org.eclipse.team.svn.help
  upd_wst_*.php  : org.eclipse.wst.*

Where you can take Babel data dump?
  see Babel Server weekly mysqldump on
  http://www.eclipse.org/babel/downloads.php
  
  or, Amazon Machine Image for Babel in Tokyo region.
  (cf: http://d.hatena.ne.jp/satoruyoshida/20120929/1348883183 )

Thanks!

Satoru Yoshida
@strysd