Re: Fwd: Re: [office-metadata] naming schema

[Patrick Durusau <patrick@durusau.net>]

Florian,

I think Bruce was suggesting that we recommend a "best" practice and I 
rather like the idea of using only one way to reference an object, at 
least as far as Bruce's citations are involved.

I don't think that would prohibit you from reusing some other means of 
referencing objects.

Bruce?

Hope you are having a great day!

Patrick

Florian Reuter wrote:

> 
>
>
> ------------------------------------------------------------------------
>
> Subject:
> Re: [office-metadata] naming schema
> From:
> "Florian Reuter" <freuter@novell.com>
> Date:
> Wed, 21 Feb 2007 13:43:39 +0000
> To:
> <bdarcus@gmail.com>
>
> To:
> <bdarcus@gmail.com>
>
>
>Hi,
>
>makes sense.
>
>However ...  :-)
>
>There are already a lot of naming schemes in ODF which establish relationships: Style, Names of tables, names of
>bookmarks, etc... My idea was to resuse them.
>
>Sure we can add a UUID. However then there would be at least two ways to address an object. E.g. consider a graphic with
>a name "Mona Lisa" and the UUID 24905720957209.
>
>So why would we like to have two different ways to reference to this object?
>
>This is where the idea of using "graphic::'Mona Lisa'" comes from since we already have an "identifier" set by the
>user...
>
>I have the strong feeling that if we introduce an additional reference schema, e.g. based on UUID and meta:class that we
> will end up allowing an ID/meta:class everywhere we can set a display-name or a style name :-)
>
>So I just tried to reuse the referencing schema present.
>
>However if this "feeling" is wrong there clearly is a need for an additional referencing schema.
>
>~Florian
>
>
>
>  
>
>>>>Bruce D'Arcus <bdarcus@gmail.com> 02/21/07 2:23 PM >>>
>>>>        
>>>>
>
>On Feb 21, 2007, at 8:11 AM, Florian Reuter wrote:
>
>  
>
>>When I understood you correctly you prefer the ID way.
>>
>>What I tried to do is to present a ways to "generate unique IDs". 
>>Sure, we can simply say the a ID is a URI. But why not use a naming 
>>schema to generate unique IDs?
>>
>>I understand that using the "opaque identifiers" identifiers give you 
>>a simply direct relationsship:
>>
>>ODF entity + URI id   <====> RDF statement about URI id
>>
>>My goal is to establish relationships differently:
>>
>>ODF enitity + style name <====> naming schema which maps between style 
>>names and URI <===> RDF statement about URI
>>
>>I understand that this is no longer an "opaque identifier". What I 
>>don't get is why this is soooo bad :-) In fact I thought that would be 
>>particulary clever :-)
>>    
>>
>
>I think you've walked yourself through most of an explanation: there 
>are already good URI (and other ID) schemes that are known to work well 
>(UUIDs and such), so why would we invent some new one, particularly one 
>based on natural language strings and such?
>
>But then you shift gears and want to say "no, identifiers should not be 
>opaque; they should be meaningful". It's actually more than that they 
>are "no longer 'opaque identifiers'" (as if it's just an accident).
>
>So I would turn this around and ask why you would possibly want to do 
>this? We can achieve the exact same thing by allowing an optional 
>meta:class attribute on styles.
>
>But that presumes we allow it elsewhere too, which is why I wanted us 
>to solve the core proposal first.
>
>Bruce
>
>
>  
>

-- 
Patrick Durusau
Patrick@Durusau.net
Chair, V1 - Text Processing: Office and Publishing Systems Interface
Co-Editor, ISO 13250, Topic Maps -- Reference Model
Member, Text Encoding Initiative Board of Directors, 2003-2005

Topic Maps: Human, not artificial, intelligence at work!