[humanmarkup-comment] ARTICLE: GSA: Human Markup Language -- as Dataand Within Data

[Ranjeeth Kumar Thunga <rkthunga@humanmarkup.org>]

I wanted to share an article draft I wrote for the GSA Newsletter (part of
the US Gov't).  I didn't get a chance to incorporate many of your input's
during time constraints, but they had mentioned it should only be an
introductory essay anyway.



I've been away from discussion for a while, and probably for some more time
due to non-HumanMarkup constraints, but I do have some important ideas to
share (and some past work to assemble).  I do invite all of you to develop
formal documents which may explain aspects of our work to this date, or
ideas you may have, in a summarized format.  This will be of tremendous
help.



Ranjeeth Kumar Thunga



----------------



Title: Human Markup Language--as Data, and Within Data



As a society, and within our government, we have celebrated having immediate
access to data.  Recent technical advances have enabled us to transmit and
exchange information instantly and reliably.  XML standards are starting to
become very promising as universal formats of data representation and
exchange.   We may conclude that these proposed solutions will finally
eradicate our data communication problems.  Yet, something essential still
appears to be missing.



Systems based on current XML formats are able to deliver data in a manner
all platforms can readily interpret, transfer, and store.  Yet, when all is
said and done, we must sit back and answer the following questions as well:
Is the data clear, not only in its content, but also in its context?  Does
this data adequately represent the individuals it is supposed to?  Is this
data meaningful for you, for them, for us?  And, does this data truly
reflect a system of government "of the people, by the people, and for the
people"?



Very often, this isn't the case.  Very often, fundamental human
characteristics within data are ignored, abandoned or even deliberately
removed.  Thus, Human Markup Language (HumanML) has been launched as an XML
specification designed to represent and embed human information explicitly
as data and metadata.  What is this human information?  This is any
characteristic that uniquely describes aspects of us as human beings.  This
can include our basic physical characteristics, as well as our intentions,
motivations, socio-cultural descriptors, beliefs, values, and behaviors.
Further, constructs representing our mindsets, or psychological schemas, can
be made explicit through a combination of several different HumanML modules.



Anecdotal summaries, which have so far been the only means of qualitatively
representing human characteristics, cannot be interpreted or parsed by
applications.  Thus, it is now up to us to design systems that can make
human information available as explicit, parsable data.  Data representation
of human characteristics would have to be fluid, flexible, and contextual,
yet at the same time adhere to universal standards.  All these needs can be
met through XML.  Thus, by designing an XML framework of human
characteristics, we can greatly clarify the scope, purpose, and context of
information by effectively representing human information as data.



There are numerous ways in which HumanML can improve current systems.  An
obvious example is in the enhancement of electronic records management.  In
current systems, a minimal set of human information may be described.  These
may include the name of the person, appropriate contact information, and
perhaps geographical address.  However, these systems may be neglecting
other human characteristics that may be absolutely essential for proper
interpretation.  These characteristics could include the individual's
background, motivation, intention, attitude, psychological state, personal
perspective and belief.  Although these human details may never have been
assembled before as explicit data, they may very well be essential in
interpreting records within their proper context.



Human information may be used not only as metadata, which provides context
for preexisting data, but also as data itself.  We can adapt current
systems, which already have the function of representing human data, towards
XML based representations of the same data.  For example, public safety and
social services databases, which currently may store human information,
could soon adopt a universal format for exchanging data based on XML,
through HumanML.  This would allow far more precise descriptions of
individual characteristics and far more efficient exchange of data between
one another.  Culturally and contextually specific frameworks can be easily
customized and elaborated on to the degree they are needed, in an explicit
fashion, through HumanML.



When these systems are developed, they would have to inevitably address some
important concerns-namely, the privacy, invasiveness, and security of such
information.  It would be up to each application and agency to determine how
to best gather, represent, and utilize this data in a manner that is
representative and helpful, without being obtrusive.    We are continuing to
explore a means by which we can provide individual access to a unified
repository of human data, which can then be shared with appropriate
agencies.



An example of using a HumanML based data repository could be in allocating
resources to individuals for various types of government assistance,
including disaster relief.  Unique human considerations need to be taken
into account in all solutions provided.  For example, during tragedies such
as the WTC incident, custom tailored action plans could be automatically
generated based on a HumanML representation of the cognitive framework of
individuals.  For refugee relief or foreign aid, cultural context and
individual specifics could be embedded directly within data.  This would
allow us to accurately develop solutions that automatically take these
factors into account.  Ultimately, a system of describing human
characteristics through XML could allow for far more precise, targeted
appropriation of government efforts than either impersonal statistical
information or anecdotal summaries could provide on their own.



How can such a system be technically developed?  Since HumanML is in fact
XML, we can use all of the related XML technologies to help build this
infrastructure.  Current XML based technologies can be used to transform
(XSLT), message (SOAP), represent (XSLFO), query (XQuery), validate (XML
Schema), and annotate (RDF) this information.



Our effort, which had begun in early 2001, is being sponsored by OASIS as an
official HumanMarkup Technical Committee
(http://www.oasis-open.org/committees/humanmarkup).  OASIS is an
international Internet consortium working with several hundred individual,
corporate, government, and NGO members, as well as the United Nations, in
developing universal non-proprietary open standards of information exchange
through XML.  Additionally, we have recently organized HumanMarkup.org, Inc.
as a separate non-profit company designed to support this work, by
developing solutions and applications based on these specifications.



We are currently in Phase 1 of the effort, which involves identifying the
requirements and frameworks we will represent within HumanML.  This is an
open source effort, and thus everyone is invited to participate in the
discussion.  Additionally, we are hoping to develop formal relationships
with government agencies especially interested in implementing these
solutions by further inquiring about the needs as they currently exist.
HumanMarkup.org, Inc. has been designed exclusively for this function.
Thus, feel free contact us if you are interested in exploring these
solutions.  With human specifications in place, we can ensure that our
government truly does represent "We the People".



Ranjeeth Kumar Thunga