plan

What should be trivial:
    * Finding tags for a given file.
    * Finding files for a given tag.
    * Adding a tag to a given file.
    * Removing a tag from a given file.

What should be possible but unimportant:
    * Finding files for a combination of tags.
    * Giving tags descriptions.
    * Marking tags as related.

Structure:
    .tigger dir in any tagged folder.
    'tags' dir in there, with each file in it being named for one tag
      - maybe also split into subdirs by first two chars of tag, so as to
      avoid having many files all in a single dir
    each tag file begins with a description, then a blank line, then one filename per line
    'index' file also in .tigger, alternates between filename lines and taglist lines
    'tags' used for quickly listing files by tag, 'index' used for quickly listing tags by file
    'index' automatically regenerated after any command that modifies a tags file

Commands:
    tigger tag $file $tag [$tag ..]
    tigger untag $file $tag [$tag ..]
    tigger files $tag
    tigger tags $file
    tigger mv $file $newfile  // does an mv, and preserves tags
    tigger cp $file $newfile  // does a cp, and duplicates tags
    tigger rm $file           // does an rm, and removes tags

Pseudocode:
    reindex:
        if index is nonexistent, create it, and set new_index=true
        for each tagfile in tags:
            if last modified date is more recent than index's l.m.d, or new_index==true, add $tag to the hashmap for each filename it contains, creating a new entry for file if needed
        iterate over the hashmap, dumping out in the index format specified above

    tag $file $tag [$tag ..]:
        if $file is not in current directory, move to $file's directory
        for each $tag:
            check existence of tags file for $tag, create if nonexistent
            append $file name to it
        do reindex

    untag $file $tag [$tag ..]:
        if $file is not in current directory, move to $file's directory
        for each $tag:
            check existence of tags file for $tag, skip $tag if nonexistent
            remove $file name from it
        do reindex

    files $tag:
        start with a list
        walk the tree from here, gathering a list of all subdirs plus .
        iterating over the dir list, for each dir:
            if $tag is in tags for dir, add files to the list, with path relative to starting dir
        dump the list to STDOUT, one file per line

    tags $file:
        if $file is not in current directory, move to $file's directory
        do reindex
        search index for $file, print out its taglist if found, otherwise "untagged"

    mv, cp, rm:
        mimic OS behavior for these, modifying tags afterwards as needed
        do reindex