FW: [tm-pubsubj-comment] Good PSIs never die

["Bandholtz, Thomas" <thomas.bandholtz@koeln.sema.slb.com>]

Kal cannot post to this list. He has interesting contributions. Isn't tm-pubsubj-comment public?

 

-----Original Message-----
From: Kal Ahmed [mailto:kal@techquila.com]
Sent: Thursday, March 21, 2002 3:10 PM
To: Bandholtz, Thomas
Subject: RE: [tm-pubsubj-comment] Good PSIs never die

Thomas, 

My last email was 
bounced by the list as well... :-(

At 13:54 21/03/2002 +0100, you 
wrote:
> [Kal:]
> No one has yet 
>   said that the documentation would be XML ! But even so which is more human 
>   readable:
>
> <record>
> <isbn>123456-09-23</isbn>
> <auth-code>AHM1298</auth-code>
> <pubdate>20011110</pubdate>
> <stock-code>98993939385402</stock-code>
> </record>
>
> <book>
> <book-title>XML 
>   Meta Data</book-title>
> <authors>
>   <author>Kal 
>   Ahmed</author>
>   <author>Danny 
>   Ayers</author>
>    
>   ...
> </authors>
> <published>2001-11-10</published>
> <description> 
>   -- blurb about the book goes here 
>   </description>
> </book>
>
> I would suggest that XML of the 
>   first form is "machine-readable" and XML of the second form is 
>   "human-readable". But depending upon the system(s) involved, the first form 
>   might be the only form that can be automatically generated for the subject 
>   indicator.
> [Thomas:]
> We have been talking about XTM, RDF, XHTML, customized 
>   XML so far - all this is XML. But you may be right - needs not to be XML. But 
>   I think it should not be binary encoded.

If this became a limitation for a PSI, it would 
restrict PSIs to being a much smaller subset of all subject indicators. The ISO 
and XTM specifications do not specify a format for a subject indicator. I guess 
it would be a shame if that flexibility had to be sacrificed for PSIs.

> Readability only depends on the specific intelligence implemented 
>   in the machine/human.
> If I (human, 
>   hopefully) understand the encoding of <auth-code> etc., I can read 
>   it.
> If a machine doesn't, it cannot read 
>   it neither.
>  

This is true within the definition 
of "readable". But should a PSI resource be "readable" or "understandable". If 
it is the latter, should that "understandability" be dependent on other 
knowledge external to the resource itself ? In limited circumstances (e.g. 
intranet or extranet environments) it could be argued that all users of the PSI 
would know what auth-code indictaed (esp. if using a documented schema). But in 
generalised internet solutions, surely a subject indicator that relies on 
knowledge of yet another schema would be flawed 
?

Cheers,

Kal