Re: RDBMS design issue

On 7/9/2025 12:44 PM, Joe Gwinn wrote:

On Wed, 9 Jul 2025 11:56:43 -0700, Don Y <blockedofcourse@foo.invalid>
wrote:
 

On 7/9/2025 11:40 AM, Joe Gwinn wrote:

But, an agent with access to the data AND THE RELATIONSHIPS
BETWEEN THEM can extract additional information to add value.

>
I think we've exhausted this SED thread, so I'll stop.
>
=====================================
>
But I did recall a key to the larger issue.  There are lots of ways to
organize library indexes.  The Savant of such things, Shiyali
Ramamrith Ranganathan, worked out the details in the 1930s, in India,
publishing in English.  He spent his entire career on this issue.
>
.<https://en.wikipedia.org/wiki/S._R._Ranganathan>

>
You're not trying to organize a library index.  You're trying to imbue
data with *meaning* that another agency can understand.
>
If some ill has befallen you and I can't get in touch with your
"emergency contact", should I give up?  Perhaps if I knew which
of the names in your phone were RELATIVES I might give one of them
a try...
>
Which of the 7 doctors in my phone should be called if I was
having a dental problem?  Medical problem?  What about those
"doctors" who have nothing to do with medicine??  Which are
my *current* providers vs. records of PREVIOUS (now retired!)
providers?
>
How do you (another agency) synthesize this information without
access to my grey matter?

 You are not the first to think of such things, by orders of magnitude.

Yet you don't see those solutions in the wild.   Look at your email
address book, your phone's "contacts", the index file you purchase
at the stationer's, Rolodex, etc.  They are all mundane in their handling
of *meaning* beyond the recording of raw data.
I have entries in mine for the receptionists (prior and present) at each of
the providers I've used (dentist, medical, HVAC, etc.), medtechs and nurse's
aids, the plumbers I've hired, the guy who painted the house decades
back, the guy who serviced the swamp cooler before I moved in (so we know
NEVER to hire him, again), etc.
But, still rely on my grey matter to recall who these people are/were
and their significance to me.  So, when I call my dentist, I can
address the woman who answers the phone by her first name -- without
asking her what it is!
I can do a brain dump onto index cards -- but that won't help an AI
sort out who these people are, "to me".

Another place to look are the various Secretary's "Handbook to Filing
Systems".  Google will bring up lots of examples.
 I was deep into such issues back in the days when DEC VAX/VMS was
being invented.  DEC eventually published a book on VAX/VMS Internals
and Data Structures:
 .<https://www.amazon.com/VAX-VMS-Internals-Data-Structures/dp/1555580599>
 It cited two books that were very influential.  The most important is:
  "Simon_Herbert_A_The_Sciences_of_the_Artificial_3rd_ed" - The
Architecture of Complexity.  MIT Press, still in print.
 The other, regarding file system organization,was Ranganathan.
 I don't recall where I got the Secretary's Handbook.  It may have been
my Department Secretary of that day; these no longer exist.
 Dig in if you wish.  Computers have changed nothing essential.

What has changed is they are ubiquitous.  So, you don't NEED to rely
on "scraps of paper", index cards, poorly designed "contacts"
datasets, address books, etc.
But, you still need a means of tagging information that imparts
meaning that something other than your own grey matter can comprehend.
The fact that a "computer" can now access that as a practical matter
means it is now a value-added proposition that an be exploited.
If you had the discipline to write everything about a contact on an
index card bearing their name, then the move to automating such
would be a piece of cake.  But, "you" (the populace in general)
don't do such things.  You, instead, rely on your mind to carry
much of that information -- largely because there is no structure that
made it easy for you to record it.