One question: Would there be any problem of the name/nodeID being non-unique
when trying to look across SPDX docs if we went with B?

Hmm,  main concern I have is that the temporary license number used to
represent a non standard license is not a global value, but rather is
localized to a specific package (hence SPDXDoc).  For that reason, A
probably makes the most sense to be.

On Mon, Dec 6, 2010 at 8:45 AM, Kate Stewart <kate.stewart@...> wrote:

Hmm, main concern I have is that the temporary license number used to
represent a non standard license is not a global value, but rather is
localized to a specific package (hence SPDXDoc). For that reason, A
probably makes the most sense to be.

The non-standard licenses will be global regardless of which of those
examples we use. Rdf only has one level of storage. Everything ends
up in the graph next to everything else. Basically the only
difference between example a and b is that in a there is an addition
property connecting the spdx document to the non-standard licenses.

Peter

PS: The following is the ntriples version of the graph produced by example a.

_:bn1291605708126ai <http://www.w3.org/1999/02/22-rdf-syntax-ns#type>
<http://spdx.org/spec#SPDXDoc> .
<http://oss.net/foo-1.0.tar.gz>
<http://www.w3.org/1999/02/22-rdf-syntax-ns#type>
<http://spdx.org/spec#Package> .
<http://oss.net/foo-1.0.tar.gz> <http://spdx.org/spec#DeclaredName>
"Foo 1.0 Source" .
<http://oss.net/foo-1.0.tar.gz> <http://spdx.org/spec#DeclaredLicense>
_:nbn1291605708126ajNFullLicense-1 .
<http://oss.net/foo-1.0.tar.gz> <http://spdx.org/spec#Files>
_:bn1291605708126ak .
<http://oss.net/foo-1.0.tar.gz#contrib/iostream2/zstream.h>
<http://www.w3.org/1999/02/22-rdf-syntax-ns#type>
<http://spdx.org/spec#File> .
<http://oss.net/foo-1.0.tar.gz#contrib/iostream2/zstream.h>
<http://spdx.org/spec#Type> "source" .
<http://oss.net/foo-1.0.tar.gz#contrib/iostream2/zstream.h>
<http://spdx.org/spec#DeclaredLicense>
_:nbn1291605708126alNFullLicense-2 .
_:bn1291605708126ak <http://www.w3.org/1999/02/22-rdf-syntax-ns#first>
<http://oss.net/foo-1.0.tar.gz#contrib/iostream2/zstream.h> .
_:bn1291605708126ak <http://www.w3.org/1999/02/22-rdf-syntax-ns#rest>
<http://www.w3.org/1999/02/22-rdf-syntax-ns#nil> .
_:bn1291605708126ai <http://spdx.org/spec#AnalysisFor>
<http://oss.net/foo-1.0.tar.gz> .
_:nbn1291605708126ajNFullLicense-1
<http://www.w3.org/1999/02/22-rdf-syntax-ns#type>
<http://spdx.org/spec#License> .
_:nbn1291605708126ajNFullLicense-1 <http://spdx.org/spec#LicenseText>
"License 1 terms and conditions" .
_:bn1291605708126ai <http://spdx.org/spec#EmbeddedLicense>
_:nbn1291605708126ajNFullLicense-1 .
_:nbn1291605708126alNFullLicense-2
<http://www.w3.org/1999/02/22-rdf-syntax-ns#type>
<http://spdx.org/spec#License> .
_:nbn1291605708126alNFullLicense-2 <http://spdx.org/spec#LicenseText>
"License 2 terms and conditions" .
_:bn1291605708126ai <http://spdx.org/spec#EmbeddedLicense>
_:nbn1291605708126alNFullLicense-2 .
_______________________________________________
Spdx-tech mailing list
Spdx-tech@...
https://fossbazaar.org/mailman/listinfo/spdx-tech

On Mon, 2010-12-06 at 10:38 -0700, Peter Williams wrote:

On Mon, Dec 6, 2010 at 8:45 AM, Kate Stewart <kate.stewart@...> wrote:

Hmm,  main concern I have is that the temporary license number used to
represent a non standard license is not a global value, but rather is
localized to a specific package (hence SPDXDoc).  For that reason, A
probably makes the most sense to be.

The non-standard licenses will be global regardless of which of those
examples we use.  Rdf only has one level of storage.  Everything ends
up in the graph next to everything else.  Basically the only
difference between example a and b is that in a there is an addition
property connecting the spdx document to the non-standard licenses.

Are we talking about the same sort of global?

Non standard licenses are only what is found in the package, and should
have no meaning outside of it.


(ie. license 1 in SPDX doc for package a, is quite different from
licene 1 in SPDX doc for package b).

On Mon, Dec 6, 2010 at 12:00 PM, Kate Stewart
<kate.stewart@...> wrote:

Non standard licenses are only what is found in the package, and should
have no meaning outside of it.


(ie. license 1 in SPDX doc for package a, is quite different from
licene 1 in SPDX doc for package b).

In the scenario where one reads two spdx files each containing a
non-standard license node with @id='FullLicense-1' into a single graph
they will *not* collide. However, both of the non-standard licenses
will be top level nodes in the graph, each with their own unique uri.
This is true regardless whether the licenses are directly associated
with the spdx doc or not.

On Mon, 2010-12-06 at 12:30 -0700, Peter Williams wrote:

On Mon, Dec 6, 2010 at 12:00 PM, Kate Stewart
<kate.stewart@...> wrote:

Non standard licenses are only what is found in the package, and should
have no meaning outside of it.


(ie.  license 1 in SPDX doc for package a,   is quite different from
licene 1 in SPDX doc for package b).

In the scenario where one reads two spdx files each containing a
non-standard license node with @id='FullLicense-1' into a single graph
they will *not* collide.  However, both of the non-standard licenses
will be top level nodes in the graph, each with their own unique uri.
This is true regardless whether the licenses are directly associated
with the spdx doc or not.

Its licenses being top-level in the graph that feels wrong to me, its
like files in a package being top level as well.

Splitting out licenses for different treatment than files isn't
symmetric.