[
Lists Home |
Date Index |
Thread Index
]
Surely data-context = a bunch of bits in relationship to one another,
but you don't know what the relationship is.
Lack of context doesn't make something random, it just makes it out of
context. Randomness is something very different, isn't it?
When you hear a word or phrase "out of context" is it suddenly random?
No-it's just out of context.
There are some texts on walls at archeological sites that nobody has
cracked yet... because they lack enough context to decipher them. This
is rare however--even a little context can make data far more
informative. But taking away that context doesn't mean the runes are
random.
[Data + context = information] actually seems a useful thought model
Let's take average temperature, for example, measured every day for the
last hundred years-- that is data.
If you knew nothing whatsoever about what the data referred to, or how
it could be manipulated, it might "appear" random, a stream of bits.
This is a lack of context problem.
Although obviously temperature data wouldn't really appear random at all
under inspection--patterns would quickly emerge. You might then inspect
that data to see if you could learn anything from the set in isolation.
Given the limited range of values, the data might even suggest a "use"
for itself....especially given the evidence of what looks like date
information, which provides yet more context. [Information - context =
data]
Unless I am mistaken you are positing context as something binary,
either on or off. Either you have it or you don't. [Forgive me if I
misunderstood]
But surely context, like evidence, is all about weighting. There can be
degrees of context. How much context is too much context (overly
verbose)? How much context is not enough context? How much context is
"just enough"?--that seems a very important question and will vary
according to use case.
If context really is binary then what does that say to the Semantic Web?
Don't we need to build mechanisms that can deal with some fuzziness?
Isn't that one of the reasons behind human readability - because we, not
machines, are the best fuzzy engines and context builders?
Marketshare numbers [data], for example, mean pretty very little in
isolation. But add context and the data becomes information. Add lots of
context and that information becomes something you might act on.
You say "Can data even exist without context? I don't think it's data if
there's no context to which it can be applied"
That looks suspiciously like reduction ab abdsurdam
Of course, if there is "no context to which it *can* be applied" then it
is not data. But that is not what owen suggested.
There IS a context to which it can be applied, we just don't necessarily
know what it is. That is why it is data, rather than "random bits" or
useless patterns.
Context can be introduced or taken away, just as owen stated. Context
can be provided at either end of the pipe, while the data in the middle
is "just bits".
As you say, using XML for MOP-ish descriptions seems like a good idea.
James Governor
RedMonk
(+44) 207 254 7371
-----Original Message-----
From: Jeff Lowery [mailto:Jeff.Lowery@creo.com]
Sent: 12 February 2003 18:26
To: 'Owen Walcher'; Jeff Lowery; xml-dev@lists.xml.org
Subject: RE: [xml-dev] web services stack
> XML is *more* than a data model, it is an information model, where:
>
> data + context = information
Yes, but
data - context = a bunch of random bits
So your assertion doesn't seem all that profound. Can data even exist
without context? I don't think it's data if there's no context to which
it
can be applied. Does an XML document, as a store of data, have a
special
ability to carry context? I don't see how it does.
> Adaptive application can be written (if certain XML modeling rules are
> followed), that allow Meta Object Protocols (MOP) and
> Reflection to be used
> to dynamically extend the runtime environment.
Ah-yup. But a specific application of XML technology does not by itself
constitute an proof that there is an underlying data model for the
syntax
that is XML. The fact that you can construct a data model (or
information
model, if you like) using XML syntax does argue for its flexibility in
being
adaptable to a wide variety of programming a modeling domains. But it
is,
by the same token, adaptable to wide variety of document domains, and
serialized representations of all sorts of things.
It think using XML for MOP-ish descriptions is an excellent idea myself,
and
have suggested as much this list in the recent past for describing how
to
represent XML data in objects and collections of objects. It doesn't
mean
that XML has a fixed underlying data model, though. You can make it
represent such data models, but that's different hill of beans.
-----------------------------------------------------------------
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://lists.xml.org/ob/adm.pl>
|
