Thanks Al and Florent for your feedback. It's good to see that
we can resolve some issues/implementation differences before the PlugFest. I
will create a JIRA entry to continue the discussion.
@Florent: Our client also does the lookup in the property
definition. However I am not happy with this approach as it causes an
unnecessary burden to the client to cache in some form the property
definitions. In a large repository there easily can be tens of thousands of
properties. Also if I consider a mobile device as a potential CMIS client I would
like to avoid this additional lookup.
I had the same idea as Al. It would be helpful and nice to
deliver the display name everywhere where we get a property value. I will file
a JIRA change request to change the spec in this aspect. I also suggest to add
a comment to the spec that a property name can be a technical name and that
only the display name should be presented to the user. A similar situation is
if the client wants to create a query: The spec enforces that the property name
is used as a column name in a query. If we get query name and display name with
a property value we should be fine for both cases.
The confusion with property id vs. property name was discussed
on the f2f meeting. It seems that the issue is that the property id "is
unique" (without mentioning in what space) and the name is unique within
the type hierarchy and to be used in a query. So both are unique but one is
more unique than the other and this more unique is repository specific. If I
remember correctly there was a use case for this that nobody exactly could remember.
Summarizing the discussion here the use case looks like to expose "the URI
or un-mangled form of the property". In this way a repository can expose
something that is unique in whatever scope it considers useful. This makes
sense to me. However I am still not sure what a client should do with this
information. My suggestion would be to add this information to the spec where
the property id is introduced. This seems to be a separate JIRA issue.
I filed those under:
http://tools.oasis-open.org/issues/browse/CMIS-112
http://tools.oasis-open.org/issues/browse/CMIS-113
Of course feel free to extend and improve the proposal.
Jens
From: Al Brown
[mailto:albertcbrown@us.ibm.com]
Sent: Mittwoch, 4. März 2009 17:50
To: Jens Hübel
Cc: cmis@lists.oasis-open.org
Subject: Re: [cmis] unreadable property names and current clients?
That is an interesting issue. I thought clients would
display the name directly. Please open an issue for this. It might be better to
leverage the property id and include the display name in atom
entry/getProperties call.
-Al
Al Brown
Emerging Standards and Industry Frameworks
CMIS: https://w3.tap.ibm.com/w3ki07/display/ECMCMIS/Home
Industry Frameworks: https://w3.tap.ibm.com/w3ki07/display/ECMIF/Home
Office 714 327 3453
Mobile 714 263 6441
Email albertcbrown@us.ibm.com
CONFIDENTIAL NOTICE: The contents of this message, including any attachments,
are confidential and are intended solely for the use of the person or entity to
whom the message was addressed. If you are not the intended recipient of this
message, please be advised that any dissemination, distribution, or use of the
contents of this message is strictly prohibited. If you received this message
in error, please notify the sender. Please also permanently delete all copies
of the original message and any attached documentation.
Jens Hübel ---03/04/2009 03:24:36 AM---Hi all,
|
|
|
|
|
|
|
|
|
|
|
|
Hi all,
I have a question how the
typical clients behave in regards of property names today.
Properties have a name, an id
and a query name. A property definition also has a display name. In a
getChildren() or getProperties() call a property is identified by its name.
Property names have certain syntax constraints. So it might happen that a
repository needs to map its internal property identification to something that
fulfills the CMIS syntax. This could be a generated, meaningless string
like PROP_4711. Only the property definition then contains something
readable in the displayName like "Invoice number". All the examples
in the spec however contain readable property names.
So my question is: Will
existing CMIS clients today display the property name directly in the user
interface or will they typically do a look-up in the property definition for
the display name before presenting this in the user interface?
Any feedback is welcome how
you have implemented this today. Would I break clients by using a generated,
unreadable property name like PROP_4711? Do we need to define a recommendation
in the spec (in this case I would create a JIRA issue).
Thanks Jens