Re: Commit: import architecture now uses complex heuristics

From: Rui Miguel Seabra (rms@multicert.com)
Date: Tue Feb 26 2002 - 10:12:01 GMT

  • Next message: Rui Miguel Seabra: "Re: Text at bottom of the license agreement"

    This may be related to problems I have experienced with the AbiGdkPixBuf
    plugin and bz2abw plugin.

    Maybe other were affected, I can't tell.

    Hugs, rms

    On Mon, 2002-02-25 at 20:01, Dom Lachowicz wrote:
    > Well, almost - it's about 95% done. I basically just need to rewrite 1
    > more method which works right now just as if my heuristic code hadn't
    > been committed.
    >
    > I'll clean up my code shortly to use a Confidence datatype instead of a
    > UT_uint8.
    >
    > Basically, everything that imports returns a normalized number between
    > [0,255] with 0 being "I'm not at all confident", 127 being "I'm so-so"
    > and 255 being "I can totally handle this file type". Applies to both
    > recognizeContents and recognizeSuffix methods.
    >
    > What I'm going to do is heavily weight the recognizeContents method
    > (maybe 85-15) and apply the following heuristic:
    >
    > my_match = heuristic(contentsConfidence, suffixConfidence);
    > if ( my_match > best_match )
    > best_filetype = my_match_filetype;
    >
    > This will fix a few bugs in bugzilla.
    >
    > Dom

    -- 
    + No matter how much you do, you never do enough -- unknown
    + Whatever you do will be insignificant,
    | but it is very important that you do it -- Ghandi
    + So let's do it...?
    




    This archive was generated by hypermail 2.1.4 : Tue Feb 26 2002 - 05:23:04 GMT