bryan rasmussen wrote:
Hi,
I've just used cadaver(1) from a linux>
I believe it would be best to use WebDAV for that particular example. After
all, the Bookmarks UI just gives you a view of the user's DAV
home-collection.
I need to, hopefully, support a lot of users, suppose the level of
users of Del.icio.us. I had expected it was just basically a database
store, what does the DAV setup imply about performance/scalability?
Is the DAV collection a URL mapping of some sort to a database, in
other words it doesn't just save XML bookmark files in the filesystem
does it?
Our WebDAV is a filesystem in the database.
Performance is fine - we use it everywhere ourselves. Our public websites
are all stored in DAV (both those consisting of static HTML such as
<http://www.openlinksw.com/> and those coming from ODS-Wiki instances, such
as <http://wikis.openlinksw.com/> where both the text for each page and the
XSLT skins are files in DAV), etc.
This means you can set ACLs/permissions on various parts of the filesystem,
and yet browse it via HTTP in an ordinary browser, or mount it using the
Finder on the Mac ors davfs2 on Linux, whatever takes your fancy. WebDAV
itself uses HTTP GET and PUT and suchlike commands.
More interestingly, because I put it in the rdf_sink/ collection, the
Sponger will have been invoked to pull triples out of it.
What benefits will this have? (don't know Sponger, having just read
quickly on the page going over it. Seems actually like it could be
useful for my current work, but I didnt think it would be beneficial
for this project)
Saves you having to parse loads of file-types to extract RDF metadata, etc.
Sponger does it all for you automatically. Might not be relevant right now
but it's certainly designed to give you lots of RDF data with no effort :)
HTH,
~Tim
--
Tim Haynes
Product Development Consultant
OpenLink Software
<http://www.openlinksw.com/>
