[rfk-dev] androidfindskitten

[Peter A. H. Peterson]

Quoting Steve Pomeroy:
> That said, there is some value in adding simple metadata about the
> list/package. At the very least, it'd allow a friendly name and the name
> of the list author for each list (I'm imagining a list of NKI items in a
> GUI that you can select from. Maybe with a detail view to see who's
> responsible for the NKI list about rare cheeses).  I'd say something
> simple akin to HTTP headers:
> 
> For example:
> ------- cut ----
> Key: value
> key2: value2
> 
> nki1
> nki2
> nki3
> ------ cut -----
> 
> A validating header regexp would be: ^[a-zA-Z][a-zA-Z0-9-]*: *(.*)$
> And all keys would be case-insensitive, but would be suggested to be
> transmitted in Title-Case.
> 
> Keys to start could be: Title, Author, Soup

That sounds reasonable to me. We definitely want some kind of
header-like information.

> Also, I'd leave off the gzip and let the HTTP layer negotiate that. I'm
> thinking of a javascript client here, where it'd be impractical to have
> an ungzipper. 

Sure, that's easy enough to turn on in Apache.

As an example, see:

http://robotfindskitten.org/aw.cgi?main=nki.rfk

This is actually a somewhat older NKI file, but you guys can test with
it. 

Peter

-- 
Peter A. H. Peterson, technician and musician.
 ---=[ http://tastytronic.net/~pedro/ ]=---