From: gopher-bounce@complete.org
       Date: Mon Aug  4 10:18:32 2008
       Subject: [gopher] Re: Item Type Suggestions
       
       On Sun, Aug 03, 2008 at 09:45:03PM -0700, Cameron Kaiser wrote:
       > What text do you have indexed? I'd rather not have Veronica talking to
       > anything but its own internal data sources (nothing personal, I'm just
       > paranoid like that), but I'd love to see another search solution on the
       > scene.
        HyperEstraier isn't an internet service, it's a program you can install
       and use. Runs completely locally, sends no info to other parties etc
       etc. "Talking to it" would probably consist of using its command line
       interface, or perhaps the API that the builtin HTTP crawler uses (it
       might be necessary to look into the source code of that).
       
        It has it's own database format (stores in a directory structure on
       your HD) - it's much faster than the SQL-based engines I tried. Creating
       the database, indexing, and searching can be all done thru the CLI
       (it can yield search results in XML (which I transform and display in a
       gopher menu), or other various formats).
       
        I mentioned Veronica because I imagine it might be quite easy to give a
       series of gopher URLs and their content to "estcmd" somehow, and bam,
       you have a full textgopher  search engine.. (It might not be quite this
       trivial, but I'm sure it won't be impossible either.)
       
        The text is my private e-text collection :)
        (I run the search engine on my server, which has an AMD athlon 64 X2
       4000+ with 4 gigs of RAM and SATA II disks; Debian lenny, hyperestraier
       1.4.9-1.4; indexing 16 gigs of text was finished I believe within 6
       hours; the created index database takes 11 gigs of space.)
       
       Regards,
        Mate