Database format

[Navdeep B. <ba...@ma...>]

In a recent post, Rik mentioned an idea for a new database format that he's been thinking about for awhile(<150...@li...>).

He mentioned the following fields(I'm using my own names, see his message for what he actually said):

Unicode target, ASCII target, start, length, match mode, and reported target

If anyone ever actually implements a new format, I'd like to make another suggestion: entry IDs. IDs would be stored in another file:

id<tab>reported target<tab>start<tab>length

and the index file would use the following format:

ASCII target<tab>Unicode target<tab>match mode<tab>id

This would allow database authors to include multiple targets to the same entry with ease. This would also allow the implementation of a command to bookmark & retrieve specific entries rather than just those that match a certain query.

I usually don't receive any replies to my posts, so I'd appreciate some kind of response, positive or negative.
--
Navdeep Bains
ba...@ma...
--
4/29/01, 1:25:17 PM