See also: IRC log
saz: filecontent class and security have been
hot topics lately. They are separate issues.
... where does everyone stand on the fileConent class?
carlosI: fileContent class could be useful. It
is similar to webContent class but different.
... There are issues with the file protocol
saz: where the file protocol is applicable
rdf:about can be used
... the problem still stands where the resource is not a url
... local files are not universally available but for EARL they need to
universally identifyable
cv: use of reletive uri's is an option
<shadi> http://www.w3.org/2006/10/18-er-minutes
<Daniela> sorry for being late - have problems with my headset (it produced the echo some minutes ago)
saz: for the webContent class uri:uri is being
dropped
... use rdf:about in place of uri:uri
cv: concerned that makes it worse
... eliminating uri:uri eliminates identification of the resource
saz: summary of issue - any web content has a
uri so we use rdf:about for the uri
... the problem is when a different identifier is needed
... local IP address would be in the rdf:about
cv: the problem is rdf:about is then not unique
<scribe> ACTION: cv will expand this discussion to the mailing list [recorded in http://www.w3.org/2006/10/25-er-minutes.html#action01]
saz: file content on a local drive may not have a unque name
cv: we need to avoid properties that are the same
saz: earl:filename was to compliment the rdf:about or rdf:id
cv: sees no difference between using the uri or filename
ci: sees no difference also
saz: earlier it was proposed that filename would be a subproperty of uri:uri
ci: that still doesn't solve the main problem
saz: Is there any reason not to have filename as a file protocol
ci: wonders if we want filename property
saz: question to revisit - what to do when
there is a local identifier that may not be unique when published to the
world
... Do we really need the filename property? What is the use case?
cv: the file protocol is the uri
ci: still does not work between different organizations
saz: The problem is making something on a
personal hard disk unique when published to the world
... filename would provide additional information
ci: wonders how filename will look like
<shadi> <uri:uri>file://file.html</uri:uri>
saz: filename would not necissarily be
unique
... uniqueness would come from the id of the class
ci: how do we distiguish two earl reports when file paths are the same but the files are different
saz: wonders if confusion would really happen in practice
<shadi> <earl:FileContent rdf:ID="myIndex">
<shadi> <uri:uri>file://file.html</uri:uri>
<shadi> <earl:FileContent rdf:ID="yourIndex">
<shadi> <uri:uri>file://file.html</uri:uri>
saz: clarifies that the identifier is the id
not the filename
... is there any objection to a fileContent class?
... the silence must mean that there are no objections
RESOLUTION: Adopt a fileContent class
<scribe> ACTION: Shadi will send out a proposal of the fileContent class [recorded in http://www.w3.org/2006/10/25-er-minutes.html#action02]