Search Mailing List Archives


Limit search to: Subject & Body Subject Author
Sort by: Reverse Sort
Limit to: All This Week Last Week This Month Last Month
Select Date Range     through    

[protege-discussion] IRI uniqueification ideas

Jim Tivy jimt at bluestream.com
Fri Nov 9 08:34:07 PST 2012


Hi Bradley

 

Nice to hear from you again.  Yes, I like the idea of
http://.../unqueId/Title that way you have solved uniqueness.

 

We have decided to use just http://.../unqueId and for now just use an
integer counter - GUIs are ugly.

 

We have a class annotation that specifies what property on individuals of
that class to use for name.  That way, we always have names when we have
individuals in the GUI.

 

cheers

Jim

 

 

 

From: protege-discussion-bounces at lists.stanford.edu
[mailto:protege-discussion-bounces at lists.stanford.edu] On Behalf Of Bradley
Shoebottom
Sent: November-09-12 5:17 AM
To: User support for Core Protege and the Protege-Frames editor
Subject: Re: [protege-discussion] IRI uniqueification ideas

 

Jim,

 

We at Innovatia have had a similar issue.

 

We have a adopted a IRI or filename convention that has

 

-Product acronym

-Topic/document type

-Random hexadecimal characters to ensure uniqueness

-Subject string that is a concatenation of the title of the document

 

We do not let the filename exceed 50 characters because of some CMS
limitations in this area. Also illegal character's like !@#$%^&*()?- are not
allowed. The last one - short dash - is important because SPARQL will not
query across a short dash even though W3C standards allow you to use it as a
URI/IRI. (I have had this battle with our OASIS DITA experts and the DITA
community and their view is that something better than SPARQL will come
along to solve this problem and they will keep their short dash, thank you
very much!)

 

We took this approach so that authors would feel confident they had the
correct files/results to a query as opposed to some seemingly random 16
character number like Nortel used to do.

 

If you are doing DITA sub topic reuse and need unique IRIs, ie at the
paragraph level, then the unique identifiers ones provided by XMetal are
sufficient so long as the content in your documentation set does not exceed
1000 pages. Over that and we found we needed to have more than
p_1234abcdef5678gh to help authors distinguish between many paragraphs
starting with the p character. Part of the issues is the difficulty in
assigning metadata to a reusable topic. Key refs and subject schema will
help, but so far only Oxygen offers decent Subject schema support and I do
not know of any DITA CMS capable of taking advantage of the subject Schema
feature to help sort through content.

 

 

Bradley Shoebottom
Senior Information Architect - Research and Product Development
Phone: (506) 674-5439   |   Toll-Free: (800) 363-3358
Skype: bradley.shoebottom

Email: bradley.shoebottom at innovatia.net 

 <http://www.innovatia.net/> Description: cid:image001.png at 01CD5EA2.5A7F56A0
 <http://www.innovatia.net/> www.innovatia.net

 

From: protege-discussion-bounces at lists.stanford.edu
[mailto:protege-discussion-bounces at lists.stanford.edu] On Behalf Of Jim Tivy
Sent: Thursday, November 08, 2012 9:26 PM
To: Protege List Server
Subject: [protege-discussion] IRI uniqueification ideas

 

Hello

 

This is not a direct Protege question but involves IRIs. I have alot of
individual documents that I wish to track as Individuals in OWL.  At first I
thought I could generate unique IRIs with meaningful names embedded in the
IRI.  Now I am thinking I want to generate unique numbers programmatically
to accomplish the need for unique IRIs and to retain the ability to change
the name of the Individual without deleting and adding it again.  There is
GUI to add these individuals so I can generate a next integer Id in the GUI
code.

 

I was curious, however, what people on this group consider best practice for
this problem of uniqifying IRIs and avoiding the embedding of possibly
changing semantic information in those IRIs.  

All the examples seem to show IRIs with cute names, like .../John and
../Mary.  My individuals will also have Name and Description properties, so
I will recognize them in the GUI by their name and description properties.

 

cheers

Jim

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.stanford.edu/pipermail/protege-discussion/attachments/20121109/f819e358/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image001.png
Type: image/png
Size: 5910 bytes
Desc: not available
URL: <http://mailman.stanford.edu/pipermail/protege-discussion/attachments/20121109/f819e358/attachment.png>


More information about the protege-discussion mailing list