[
Lists Home |
Date Index |
Thread Index
]
Elliotte Rusty Harold wrote:
> Alessandro Triglia wrote:
>
>> True, with XML 1.0 you can use any Unicode viewer (or any EBCDIC
>> viewer, or
>> any SHIFT_JIS viewer, or any xyz viewer, etc., depending on the
>> circumstances) -- you don't have to use a specific program like the
>> MS XML
>> 1.0 viewer that is built into IE. But still, if FI viewers became
>> ubiquitous, what would be the fundamental reason for concluding that
>> FI does
>> not comply with the "view source" paradigm?
>
>
>
> In the short term (by which I means a few years, maybe even a few
> decades) there's probably not a lot of difference. In the long term,
> i.e. centuries or more, the difference might become significant. Many
> of the NOT XML formats are much harder to decode without pre-existing
> knowledge of the format or even the specific schemas used to encode
> the information. Whether this is true of the FI version of NOT XML or
> not, I don't know. The real question is whether the full information
> content of the document is present in each instance. The level of
> redundancy also matters. Compression is the enemy of robustness.
>
A) The "Self Contained" property, which will have to be one of several
modes of operation, captures having "full information content of the
document ... present in each instance". A format that has externalized
structure, metadata, and redundancy of the data can be made
self-contained if it is paired with an interpretable version of the
externalized information. (An important distinction is how much of this
is interpretable metadata and how much is embedded in code.) I have a
strong preference that all externalized information be able to be
represented in a standardized interpretable form. This may need it's
own spec or be part of a "binary XML" spec.
B) Only self contained instances are suitable for archiving. As far as
interpreting the format, the reason we want an open standard is so that
the spec and code are available. I am confident that our distant
decendents will be able to read English and compile C code.
Robustness is important. A self-contained format need not be much worse
off than text-encoded XML. You could argue that a TNLV format is
actually a little more robust in some cases. If you know the length of
an element, for instance, you can deduce a little more when the end is
missing whereas with a missing text end tag, you have less of a clue.
The process of deduction and reconstruction might be slightly more
difficult and more likely to be done with a program than a text editor,
but it isn't that much of a stretch.
sdw
--
swilliams@hpti.com http://www.hpti.com Per: sdw@lig.net http://sdw.st
Stephen D. Williams 703-724-0118W 703-995-0407Fax 20147-4622 AIM: sdw
|
