Clarifying DFDL vs. Binary XML (was RE: use cases: binary XML for scientifc computing)

I would also like to see use cases which clarify the differences between
binary-xml and DFDL so that we're not constantly revisiting that and we can
have a positive URL reference to some clarifying point that both Binary XML
and DFDL teams can depend on. To me the crux of the issue is that in the
DFDL/descriptive world, you have this problem of data format debugging,
i.e., what if the description is wrong? This is what you avoid with the
binary-XML/prescriptive approach, and it is very worth avoiding. This is why
I see the need for both and their overlap doesn't bother me.

To that end, here's some language revised from an earlier posting, which
makes this point about presecriptive vs. descriptive formats. You can adapt
it as you see fit.

Binary-XML is a prescriptive approach, that is, it specifies a universal
format that is used for data. Binary-XML shares this category with ASN.1 and
XDR, but leverages the popularity of the XML family of standards. The DFDL
approach is descriptive. That is, the data has some format. You describe in
DFDL the format the data is in. Use-cases where this approach is preferable
include high-performance programs which often want to arrange for data
structures to be aligned and directly mappable into memory layouts or
randomly accessible on disk. DFDL allows data to meet these requirements
while still being universally described for interchange with other programs.
Other important DFDL use cases include the broad array of legacy data
formats. DFDL also leverages XML technologies for describing the logical
structure of the data, but adds the ability to describe a different physical
realization. 

Pros and Cons: The Binary-XML prescriptive approach is preferable for new
programs which simply need the improved performance and density of a binary
encoding. The prescriptive approach is in a strong sense preferable to
descriptive as use of a prescriptive approach avoids the entire issue of
"data format debugging" which is insuring that the data and the description
of it match properly, and correcting any errors in the description. The DFDL
descriptive approach is preferable for pre-existing data formats, as well as
certain high-performance needs, but use of DFDL necessarily creates this
issue of data format debugging, which must be overcome by applications
developers.

Mike Beckerle
Co-chair GGF DFDL WG
Ascential Software



> -----Original Message-----
> From: Don Brutzman [mailto:brutzman@nps.edu] 
> Sent: Thursday, December 02, 2004 9:14 AM
> To: mike.beckerle@ascentialsoftware.com
> Cc: RogerCutler@chevrontexaco.com; aslom@cs.indiana.edu; 
> sdw@lig.net; whoschek@lbl.gov; xml-dev@lists.xml.org; 
> public-xml-binary@w3.org; kchiu@cs.binghamton.edu; 
> mgovinda@cs.binghamton.edu
> Subject: Re: use cases: binary XML for scientifc computing
> 
> mike.beckerle@ascentialsoftware.com wrote:
> > I do believe that GGF DFDL is relevant to the discussion here. 
> > https://forge.gridforum.org/projects/dfdl-wg/ is the site, and
> > 
> https://forge.gridforum.org/docman2/ViewProperties.php?group_id=113&ca
> > tegory_id=803&document_content_id=2973
> > 
> <https://forge.gridforum.org/docman2/ViewProperties.php?group_id=113&c
> > ategory_id=803&document_content_id=2973> (or
> > http://tinyurl.com/435j7 in case email clobbered the long 
> URL) is the 
> > most recent presentation. Around slide 7 is where you'll 
> find content.
> 
> I agree this is quite relevant, and hope that we at least 
> have a use case which compatibly describes the subset of 
> shared motivations between DFDL and the XBC effort.
> 
> all the best, Don
> -- 
> Don Brutzman  Naval Postgraduate School, Code USW/Br       
> work +1.831.656.2149
>                MOVES Institute, Monterey CA 93943-5000 USA  
> fax  +1.831.656.7599
> Virtual worlds/underwater robots/X3D/XMSF     
> http://web.nps.navy.mil/~brutzman
>

Received on Friday, 3 December 2004 15:32:12 UTC