[
Lists Home |
Date Index |
Thread Index
]
winkowski@mitre.org wrote:
> Do you regard office file formats as binary? To me they are application
> specific file formats. Even so, is your comparison of these file formats with
> XML a lossy comparison? That is does the XML version include all the
> application specific information needed to recreate the application file or
> is it limited strictly to the data content?
I won't answer for Elliotte, but for me, yes, these are what I regard as
informationally impoverished documents. That is, they are impoverished *as
documents* because their creator relies either on specific schemas to be
understood when a consumer instantiates these documents or relies on a consumer
to use specific software to process them. In either case, the document itself
it bereft of content which might otherwise be included to guide a document
consumer to the semantics intended by the document creator. Whether the
consumer is at all interested in the intent of the creator, or in re-creating
specific semantics, is immaterial. Something has been left out of a document
which the creator might have included if not for counting on a specific schema
or specific processing software. That is, in fact, a binary format: an
'encoding' specific to one particular set of choices for its processing. It is
also unquestionably lossy--even before we talk specifically about
compression--because something has been left out of the *document* which is
required to recreate the same data structure that the creator used.
|
