Re: More namespacing woe ...

[Maciej Stachowiak <mjs eazel com>]

Miguel de Icaza <miguel helixcode com> writes:

> 
> Mhm, you raise an interesting point.  From the moniker perspective,
> using #gzip: should compress the parent, and return data compressed on
> the other end.  
> 
> So I think we should keep #gunzip: for gunziping, and #gzip for
> gzipping.
> 

Er, no. When you read you should uncompress, when you write, you
should compress. That way, access to the same resource (the compressed
file on disk) is using the same name whether reading or writing. In
particular, if you access the file through a read/write stream, you
definitely want to be uncompressing on read and compressing on write,
otherwise your mixed reads and writes will result in inconsistency.

The whole goal here is to transparently access compressed files, so I
see no value to compressing on read or uncompressing on write.

It would also be a huge mess if applications had to use a different
URI/stringified moniker to load a file and to save it. That would mean
the app would have to have detailed awareness of the syntax, and
couldn't just take a URI or stringified moniker passed to it by the
user or another app and use it blindly for both reading and writing.

A moniker represents a resource, not an action; resolving the moniker
against an interface and calling a method of that interface represents
an action.

Regards,

Maciej