An industrious monk could google mine perlmonks.thepen.com for science releated nodes and use his/her home node as an index for them.

This option has the benefits of

• not being application domain specific (if we have a bioinformatics section, why not a cgi section? ...ad infinitum...)
• being no work for pmdev since any user could do this today
• providing a litmus test for what interest there is in the use of such an index (provided there's any way of measuring node hits, referrals, and such)
• building and maintaining the index falls to those who would find it most interesting, with users /msg'ing maintainers not to miss an on-topic node, etc.

A really industrious monk could mine the Newest Nodes XML feed to help maintain the index for well-titled nodes.