W3C home > Mailing lists > Public > www-dom@w3.org > January to March 2001

Re: candidate for errata?

From: K. Ari Krupnikov <ari@iln.net>
Date: Tue, 20 Feb 2001 12:53:26 -0800
Message-ID: <3A92D946.52C7EC86@iln.net>
To: Joseph Kesselman <keshlam@us.ibm.com>
CC: www-dom@w3.org
Joseph Kesselman wrote:
> 
> >"The DOM implementation may not put arbitrary limits on the amount of
> >data that may be stored in a CharacterData node."
> >The Java bindings specify that CharacterData.getLenght() return int.
> >Is there an implicit assumption that in Java a CharacterData would not
> >exceed 2Gb?
> 
> Yes, the standard Java bindings make that assumption. Other bindings may
> bind the IDL "int" to other datatypes, of course.
> 
> Frankly, if you've got a single CharacterData that contains more than 2GB,
> it is extremely unlikely that _any_ off-the-shelf XML software will handle
> it, and one can argue that you were foolish to embed that much unstructured
> data in an XML document in the first place.

I certainly agree. The reason I asked is that I'm working on an
SQL-based implementation with Java wrappers, and there 2 GB
CharacterData _are_ possible, if ill-advised. I was wondering if I
should limit the size of CharacterData to the maximum size of VARCHAR
(or clob) field (1GB in PostgreSQL), or normalize CharacterData into a
one-to-many with its parent. And then the int question came up.

Would limiting CharacterData to 1GB be reasonable? What would you
consider to be a 'practical' limit?

-- 
K. Ari Krupnikov

DBDOM - bridging XML and relational databases
http://www.iter.co.il
Received on Tuesday, 20 February 2001 15:52:52 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 22 June 2012 06:13:48 GMT