[
Lists Home |
Date Index |
Thread Index
]
> -----Original Message-----
> From: Stephen E. Beller [mailto:sbeller@nhds.com]
> Sent: Monday, December 06, 2004 6:17 PM
> To: 'Peter Hunsberger'
> Cc: xml-dev@lists.xml.org
> Subject: RE: [xml-dev] Data streams
>
> As I said initially, larger data elements do change the
> ratios. To go to the opposite extreme, large blocks of text
> can actually be handled MORE efficiently with XML than CSV.
>
> On the other hand, the larger the attributes and other tag
> "labels," the greater the ratio, and visa versa.
>
> So, all I'm saying is that there are times when XML make more
> sense than CSV, and certain situations make CSV superior. No
> one solution is right for all circumstances.
>
> By choosing the method that fits most sensibly with the data
> will help alleviate some of the XML backlash. A good seems to
> be that, everything else being equal, (a) the longer the tags
> or the shorter the data elements, the less sense it makes to
> transport the data via XML and (b) the shorter the tags or
> the longer the data elements, the more sense it makes to
> transport the data via XML. Anyone disagree?
On the notion of transporting data via XML: I know that SOAP has not yet
been mentioned in this thread, and the point I am about to make is
somewhat tangential, but I thought some might be interested to see the
results of a study[1] done by some folks in Australia (one at U of Tech,
Sydney) that compared performance of real-time trading systems using a
text-based wire representation (FIX) and SOAP. From the intro:
"in realistic business application scenarios, SOAP's poor performance
cannot be adequately explained simply by the disadvantages of text-based
over binary wire formats. This also suggests that improvements in the
efficiency of SOAP encoders and decoders may enable its use in high
performance business applications."
Interesting conclusion...
[1] http://www2003.org/cdrom/papers/alternate/P872/p872-kohlhoff.html
Kind Regards,
Joseph Chiusano
Booz Allen Hamilton
Strategy and Technology Consultants to the World
> Steve
>
> -----Original Message-----
> From: Peter Hunsberger [mailto:peter.hunsberger@gmail.com]
> Sent: Monday, December 06, 2004 5:24 PM
> To: Stephen E. Beller
> Cc: xml-dev@lists.xml.org
> Subject: Re: [xml-dev] Data streams
>
> On Mon, 06 Dec 2004 16:35:48 -0500, Stephen E. Beller
> <sbeller@nhds.com>
> wrote:
> > In consideration of Elliotte's reply, I went back and looked at the
> > XML
> file
> > Excel generated. Here's what I found ...
> >
> > Every one of the XML data elements had this tagging structure:
> > <Row>
> > <Cell><Data ss:Type="Number">1</Data></Cell> </Row>
> >
> > In contrast, the CSV had this structure: 1,
> >
> > That's a 50 characters to 1 difference for each data element.
> >
> > I doubt that all those XML tags are necessary if you're
> rendering the
> > data in something other than a spreadsheet. But if you are
> planning to
> > use a spreadsheet, then the 50 to 1 ratio is valid, it seems to me.
>
> Use the number 10, now the difference is 51 to 2 or a ratio
> of ~26 to 1. Use the number 100 and the ratio is 52 to 3 or
> ~17 to 1. Six digits? 56 to 6 or ~10 to 1. Now add multiple
> columns of data (as any realistic example would do) and the
> ratio falls even farther.
>
> <snip/>
> >
> > So, this benchmark test still points to a huge difference
> in file size
> > and in unzipping and parsing time when you compare a large
> data array
> > in CSV compared to XML.
>
> Maybe, maybe not, the bench mark needs to be more realistic
> before you draw any conclusions about "huge".
>
> --
> Peter Hunsberger
>
>
>
> -----------------------------------------------------------------
> The xml-dev list is sponsored by XML.org <http://www.xml.org>, an
> initiative of OASIS <http://www.oasis-open.org>
>
> The list archives are at http://lists.xml.org/archives/xml-dev/
>
> To subscribe or unsubscribe from this list use the subscription
> manager: <http://www.oasis-open.org/mlmanage/index.php>
>
>
|