[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

Re: XML Compression - Java

• To: xml-dev@l...
• Subject: Re: XML Compression - Java
• From: Andy Greener <andy@g...>
• Date: Fri, 14 May 2004 09:19:35 +0100
• In-reply-to: <20040513220302.31270.qmail@w...>
• References: <20040513220302.31270.qmail@w...>

At 3:03 pm -0700 13/5/04, Landon Blake wrote:

>Any thoughts on the best compression method for XML data?

Well, LZ77 (used in zip and gzip) is good enough for most purposes.
In an application I'm intimately familiar with zip/gzip gets the
size of repetitive XML (roughly, repeating structures of around 1kb
in size) down to around 6-7% of original (but you've then got the
33% overhead of base64 encoding, so say around 8-9% overall).

If you want to improve upon the above take a look at Xmill:

	http://www.research.att.com/sw/tools/xmill

It's XML-specific, it's still in development, but it's open source.
My experiments with it have indicated that it is nearly twice as good
as zip/gzip (on the data in the above application).
-- 

Andy Greener                         Mob: +44 7836 331933
GID Ltd, Reading, UK                 Tel: +44 118 956 1248
andy@g...                       Fax: +44 118 958 9005

• References:
  • XML Compression - Java
    • From: Landon Blake <redefined_horizons@y...>

• Prev by Date: RE: XML moves COBOL into the .NET and J2EE arenas
• Next by Date: SAX Parser Raises Error Event Before startElement Event
• Previous by thread: XML Compression - Java
• Next by thread: Re: XML Compression - Java
• Index(es):
  • Date
  • Thread