RE: R-tree indexing, RA wrap and LISTENing

["Creager, Robert S" <CreagRS@LOUISVILLE.STORTEK.COM>]

> -----Original Message-----
> From: Chris Albertson [mailto:chrisalbertson90278@yahoo.com] 
> Subject: RE: R-tree indexing, RA wrap and LISTENing
> 
> Rob,
> 
> I think that's the way to go, create a new data type called 
> "Sky Coordinate" that is defined as an "RA" and a "DEC" and 
> then define the types RA and DEC and the operators that would 
> be required like <, >, +, -, *, "between" and so on.

I figured you'd say something like that ;-).  It's going to take some work
(ie I stopped after 10 minutes...).

> 
> Later you define a set of special astrometric operators
> over the set of (RA,DEC)  Certainly we'd need a "anglar 
> dstance" maybe a "precess" operator, The more functions you 
> can add to the database the more powerfull the querries can 
> be. You could even define "constellation" that returns the 
> name of the constallation that a given (RA, DEC) falls into 
> so you'd be able to querry on "all stars in Orion with mag < 
> 6" or whatever.

At what distance do you need the spherical operators over planar?  And would
they be used often enough to justify their creation?  I don't believe they
would be required for matching stars to each other (7 arcsec?).

>  
> Think carfully before signing up for e-mail notifications.
> If someone impoorts 15 million observation pairs it could 
> trigger a real flood of e-mail.  I would want something that 
> insures no more then say one e-mail an hour no matter how many hits.

Good point.  Maybe the LISTEN would populate another table, and a cron job
would send out all entries in the table once an (hour, day, ...).

Later,
Rob