On Mon, 2002-01-07 at 11:08, Massimiliano Mirra wrote:

> All of this is done locally.  Such a system would be a real breeze to
> implement on top of RAA, *if* everything was kept at the same place
> and was organized in a well-thought out directory structure.  (Yes,
> there's more to APT than this, it also shines at dipendence handling,
> but let's KIS(S).)  Now, what about:
> 
> $ raa-get update
> 
> ...gets the list of available software.
> 
> $ raa-cache search ruby xml
> 
> ...displays a similar list as above.
> 
> $ raa-cache show xmlrpc4r
> 
> ...displays a description as above.
> 
> $ raa-get install xmlrpc4r
> 
> ...fetch and install xmlrpc4r, requiring you no extra step to use it
> than to read the manual and type require 'xmlrpc/...' in your scripts.

If you wanted to Keep It Simple, why have both raa-get and raa-cache?  A
single binary, if we're not worrying about dependency tracking, multiple
package backends, etc. could handle all of this.  It might be simpler
for the user to just have 'raa-get', or maybe just 'raa'.

> 
> 
> 
> Massimiliano