[ale] Linux in libraries (a heads-up for June 2002)

[Joseph A. Knapka]

Ben Ostrowsky wrote:
> 
> >Can you explain why that is so?  Is it because the MARC data structure is
> >hierarchical?
> 
> I think it's got to do with repeatable fields.  Here's a sample MARC record:
> 
>   LEADER 00962nam  2200289Ki 4500
>   001    fla97047300
>   003    FLA
>   005    20010427183412.0
>   008    970901s1996    caua          001 0 eng d
>   020    1565921496 (pbk.) :$c$40.00
>   035    526575
>   040    FLA$cFLA$dWaOLN
>   245 00 Programming Perl /$cLarry Wall, Tom Christiansen, Randal L.
> Schwartz with Stephen Potter.
>   250    2nd ed.
>   260    Sebastopol, CA :$bO'Reilly & Associates,$cc1996.
>   300    xxi, 646 p. :$bill. ;$c24 cm.
>   490 1  A Nutshell handbook
>   500    Accompanied by booklet "Quick reference guide" (19 p. ; 21 cm)
> inserted at end.
>   500    "Covers Perl 5"--Cover.
>   500    Includes index.
>   650  0 Perl (Computer program language)
>   700 1  Wall, Larry.
>   830  0 Nutshell handbook.
>   920    970923
>   930    970923
>   945    997 FLA ils
>   949 10 $b3 2396 00329 1527$cFLAANF$d005.133P WALL$mTBLC$nFLA$p40.00$tFLABK
> 
> Note that there's a fixed-width leader, and that each field has two
> characters called indicators (such as "00" in the 245 field), which are
> often null.  There are also three 500 fields.  Each of these numbers has a
> specific meaning like "Main title", "Publisher", "Edition statement",
> "Additional author", "Series title", etc.
>
> Subfields are listed as $ followed by one character.  So the 300 field has
> two null indicators, contains the text "xxi, 646 p. :", and has two
> subfields, $b ("ill. ;") and $c ("24 cm.").

With a complete specification for MARC, designing a normalized
relational schema should be "not too hard". Repeating fields become
rows in an auxiliary table keyed to the master record. It's
pretty hard to come up with a data format that can't be
represented relationally (spatial data is the only example that
springs to mind); there may be performance issues
with a relational implementation, however. But if the current
state of the art is to search flat files that look like the
record above, it's hard to see how an RDBMS could be worse.

> It does get messier than this, of course...

Sounds like fun.
 
-- Joe Knapka
"You know how many remote castles there are along the gorges? You
 can't MOVE for remote castles!" -- Lu Tze re. Uberwald
// Linux MM Documentation in progress:
// http://home.earthlink.net/~jknapka/linux-mm/vmoutline.html
* Evolution is an "unproven theory" in the same sense that gravity is. *
--
To unsubscribe: mail majordomo at ale.org with "unsubscribe ale" in message body.