From: BlueChip <cs_bluechip_at_webtribe.net>
Date: Thu, 24 Jul 2003 01:56:39 +0100

At 18:06 23/07/03 -0400, you wrote:
>Björn Stenberg wrote:
>>Feature request #613544 describes an ID3 database, which is precisely
>>what is necessary for browsing tags. It has raised priority to indicate
>>this is an issue we (core developers) feel would be valuable.
>>The trick is to design a database/file format that is quickly searchable
>>while requiring a minimum of ram used. (I.e. waste disk space if it saves
>>ram use and/or increases speed.)
>I was toying with something like this a while ago. By imposing the
>limitation that all database entries are the same size, the file on the
>disk can be binary searched without any buffering and with relatively
>little disk access. I kinda lost interest, having got it to the point
>where it found stuff, but didn't actually do anything with it.

would this not rely on a contiguous file?

if you use pascal strings, you can search quickly... the string is
prefixed with it's length - this also allows for variable length strings;
alternatively you can prefix a long stream of C strings with an offset
table ...this second method would be quicker, but require more memory.

If you restrict yourself to 26 characters, then you only need 5bits per
character; which will compress the text quickly and simply, leaving 5
characters over for say {' ', '\'', '_', '!', '?'}

food for though.
Received on 2003-07-24