RE: Nautilus, metadata and extendet attributes

[raphael bosshard slm admin ch]

The use of metadata would also allow some quite funky things.

For example virtual folders which contain all my math files, wirtten by John
Anonymous, created between January 4th and February 15th, designated as
draft and with a filename that contains "Differentialequations" (or whatever
they are called in english).

This would allow a step far above the hierachical filesystem.

While this is quite a complicated task with current situation and requires
the creation of an extensive framework (Medusa) this would be quite trivial
with metadata.

Hmmm... Filesystems as semi-structured databases and Nautilus as a cool
interface to this database. Cool. :)


Raphael



> ----------
> From: 	Manuel Amador (Rudd-O)
> Sent: 	Thursday, January 29, 2004 21:00 PM
> To: 	Xavier Bestel
> Cc: 	raphael bosshard slm admin ch; alexl redhat com;
> nautilus-list gnome org
> Subject: 	RE: Nautilus, metadata and extendet attributes
> 
> El jue, 29-01-2004 a las 02:49, Xavier Bestel escribió:
> > Le mar 27/01/2004 à 09:32, raphael bosshard slm admin ch a écrit :
> > > As far as I know, EAs are cached. Some more
> > > information anyone?) it is definitely faster than sniffing the content
> of
> > > the file and try to match it against some regexes (This may be a bit
> > > oversimplyfied.)
> > 
> > What do you mean by "EAs are cached" ? They are all stored linearly
> > on-disk ? IMHO this is highly implementation-dependant, and in the case
> > where you have to seek to read each EA, you'll have exactly the same
> > delay as when you sniff all files.
> Reading EAs, while not as fast as simply reading a directory listing,
> would be much faster than opening each file sequentially.  The reason
> for this is that:
> 1) there is a higher probability of EAs being cached on memory than the
> file contents
> 2) the usage of EAs makes new things possible
> 3) the byte size of EA contents would be a tiny fraction (orders of
> magnitude less) than of the files they describe
> Once you've determined the file type and stored it in an EA, subsequent
> reads would be faster than sniffing the files, for all the
> aforementioned reasons.
> Nearly everyone here is forgetting about the exciting new possibilities
> and features EAs could enable or power.
> Why are people against using EAs?  Simply use them wherever they're
> available, and presto.  What's the big deal?  They're mature enough, a
> standard, and available nearly everywhere.  And where they're not
> available, there's always the old ways.
> > 
> > 	Xav
> -- 
> 	Manuel Amador (Rudd-O)
> 	GPG key ID: 0xC1033CAD at keyserver.net
>