Guy,

What about using the O/S level folder compression instead? Unless each memo is itself fairly compressible then you might not gain much benefit on a memo by memo level compression. And there is CPU cost of decompressing it. It just doesn't seem like compression is as big a deal as it used to be when disk capacity was in the MB range.

>OK, let's forget the indexes and focus on data itself.
>
>If I want to pack data tightly in memo fields (for example) for the sole purpose of reducing file size (and transport across http), is binary compressed data a smart thing to do? Or are there other gotchas?
>
>I realize I'd be giving up text searching, etc. But if I only need a memo field uncompressed at the time I view or print a record, why not store the data compressed and blow it back up a record at a time (when I want the text available)?
>
>All this is hypothetical at this point. I'm just wondering if a future version of VFP could make good use of such a compression function? (e.g., shrinking my memo fields by 80-90%)