W3C home > Mailing lists > Public > public-xml-binary@w3.org > November 2004

Re: question: Increasing factor for XML vs Binary

From: Dennis Sosnoski <dms@sosnoski.com>
Date: Sun, 07 Nov 2004 01:38:33 -0800
Message-ID: <418DED19.5040102@sosnoski.com>
To: Silvia.De.Castro.Garcia@esa.int
CC: "Stephen D. Williams" <sdw@lig.net>, public-xml-binary@w3.org

Silvia.De.Castro.Garcia@esa.int wrote:

>
> Yes, of course.
> I could like to know more or less how much bigger is a XML file from 
> its equivalent in binary format.
>
This will depend on the data, as others have said, and on the packing 
efficiency of the binary format. My own estimate based on the work I've 
done for the XBIS (http://www.xbis.org) XML encoding format is that you 
could expect to see somewhere in the range of 6:1 to 15:1 for optimal 
binary representations over text; you can get higher, but only if the 
original text document has been basically designed to waste bandwidth 
(long element and attribute names with very short content, lots of 
indentation, short documents with lots of long namespaces, etc.).

Look at it this way: Binary formats will basically generally preserve 
text string data as text, while using binary representations for ints, 
floats, etc. For numbers this means you may save about half the space of 
the original text form, while general text will be the same size (except 
for enumerations, where you can use a much more compact index number in 
place of the text values). The rest of the savings just come from 
eliminating all the markup in the document.

  - Dennis

-- 
Dennis M. Sosnoski
Enterprise Java, XML, and Web Services
Training and Consulting
http://www.sosnoski.com
Redmond, WA  425.885.7197
Received on Sunday, 7 November 2004 09:48:11 GMT

This archive was generated by hypermail 2.2.0 + w3c-0.30 : Thursday, 1 December 2005 00:07:42 GMT