Add an origin information to GUID's

In #8 the notation of the GUID was discussed and changed for the future. I want to add another aspect to this.

As we know, a GUID is unique, regardless, where, when and who the GUID hat created. So it can be used to identify things in a secure way.

But sometime, the GUID is also used as a key, that can lead to more information, than the information, that is stored in an IFC file.

Imagine, you have a GUID for a door in an IFC file. The file contains some information about the door. But often, the system, that has generated the file has more information about the door, than the parts that where include while creating the file.

Now, I want to look up this additional information (e.g. in an external database) ; and yes, the GUID is good for that, because it identifies also the door in the database.

But in which database? Who is the owner/creator of the database that has created the GUID?

Let imagine whe have a publisher that can be identified by a `PublisherCode` and this publisher has create the GUID in a database that can be identified by a `SystemCode`.

We could have something like this:

`PublisherCode:SystemCode:GUID`

e.g.
`SAP:SYS1:21b1b279-269f-4df5-9a5f-455d65c0e4f1`
or
`BSI:BSDD:25abe353-29f6-45e9-b8ba-07a7daa2ac33`

What do we win:
We get the ability to resolve all parts of the models to the origins. This would allow the creation of models that are made up out of severall source, not only one.

What do we loose
Now, we woul loose nothing, since we want to change to GUID notation, which would be a breaking change. A new implemention could impelent such an GUID-resolver easily.

Schema impact:
We would have to define the notation of the resolvable GUID string.

Instance model impact:
Breaking change, as #8 is also a breaking change

Backwards compatible:
no

Automatic migration possible:
yes