[occi-wg] Updates to OCCI core

[Ralf Nyren]

On Wed, 14 Mar 2012 11:39:33 +0000, <alexander.papaspyrou at tu-dortmund.de>
wrote:
> I think we could fix ourselves to the following normative namespaces:
> 
> occi.core.* for attributes from the core model

+1

> occi.infrastructure.* for attributes from the infrastructure model

Would require renaming all attributes in OCCI Infrastructure...
 
> and
> 
> com.example.* for our examples.

+1 for consistent naming of attributes in our examples!

> These three MUST NOT be used by third-party extensions; instead,
> third-party extensions SHOULD follow the reverse-domain-name pattern to
> prefix their attributes.

I like the whole occi.* namespace to be reserved for future OCCI spec
usage (as currently specified in OCCI Core). Otherwise I agree.

/Ralf

> Re: simple I would second what Ralf says. Properly implementing a type
> system is  a non-trivial task.
> 
> -A.
> 
> Am 14.03.2012 um 09:22 schrieb Ralf Nyren:
> 
>> 
>> On Sun, 11 Mar 2012 18:35:51 +0000, "Feldhaus, Florian"
>> <florian.feldhaus at gwdg.de> wrote:
>>> Here are some ideas for extending the description of attributes. To
more
>>> clearly specify attributes, a new new class 'attribute' should be
added
>> to
>>> OCCI Core. For the class diagram this would require to remove '+
>> attributes
>>> :Set<String>' from category and instead connect it with the attribute
>> class
>>> in a many to one dependency.
>> 
>> We actually had an Attribute class for I while and I guess it is time
to
>> re-introduce it.
>> 
>> However, we MUST NOT call it "Attribute". First we already have like 5
or
>> so different things we call "attribute" currently and adding another
>> "thing" with the same name is just too much. OCCI Core is hard enough
to
>> read as it is already. Secondly, an object of this class is not really
an
>> attribute since it has no value. It is a type definition of an
>> attribute, a
>> meta-attribute :D
>> 
>> But given a proper name which is not too confusing this is all great.
>> 
>>> Ideas for the attribute class:
>>> - the attribute class should consist of the attributes: mutable,
>> required,
>>> type, default
>> 
>> hehe, attribute class with attributes... you see why we need another
>> name?
>> ;)
>> 
>>> - the naming of attributes of the OCCI specification MUST start with
>> occi
>>> followed by term and the attribute name (e.g. occi.compute.cores)
>> 
>> No, this is too restrictive. Remember that each scheme is a separate
>> namespace. What we could do is say that attribute name must end with
>> .term.attrname, e.g. occi.[xxx.]<term>.<attr>.
>> 
>>> - the naming of attributes of user defined mixins MUST start with the
>>> reversed domain name of the mixin scheme followed by the term and the
>>> attribute name (e.g. com.example.my_mixin.my_attribute)
>> 
>> See above, same applies here.
>> 
>>> - the attribute type should contain restrictions on the attribute like
>> for
>>> example:
>>> * string: choices, min_length, max_length
>>> * number: bounds (one or more lower and upper bounds)
>>> - there might be more detailed attribute types such as URL or URI,
>> float,
>>> integer, ...
>> 
>> This must be detailed out properly. I would rather keep it on the
simple
>> side though...
>> 
>> /Ralf
>> 
>>> 
>>> --Florian
>>> 
>>> 
>>
-------------------------------------------------------------------------------
>>> GWDG - Gesellschaft für wissenschaftliche
>>> Datenverarbeitung mbH Göttingen
>>> Am Fassberg 11, 37077 Göttingen
>>> 
>>> Fon: 0551 39-20364
>>> Fax: 0551 201-2150
>>> E-Mail: florian.feldhaus at gwdg.de
>>> WWW: www.gwdg.de
>>> 
>>
-----------------------------------------------------------------------------------
>>> Geschäftsführer: Prof. Dr. Ramin Yahyapour
>>> Aufsichtsratsvorsitzender: Prof. Dr. Christian Griesinger
>>> Sitz der Gesellschaft: Göttingen
>>> Registergericht: Göttingen
>>> Handelsregister-Nr. B 598
>>> 
>>
-----------------------------------------------------------------------------------
>>> 
>>> _______________________________________________
>>> occi-wg mailing list
>>> occi-wg at ogf.org
>>> https://www.ogf.org/mailman/listinfo/occi-wg
>> _______________________________________________
>> occi-wg mailing list
>> occi-wg at ogf.org
>> https://www.ogf.org/mailman/listinfo/occi-wg