OMB's technical direction
Eric C. Kansa
http://www.nextgov.com/nextgov/ng_20090623_2138.php
Apparently, OMB is dropping the idea of making "raw" data publicly
available for the Recovery. I guess that means the public will only get
"cooked data" (in some form of XML format). I can't understand why raw
or primary data won't be available, but there seems to be an underlying
impression that the public won't understand the primary data. Can't such
an issue be solved with some documentation? Doesn't independent public
oversight of the recovery require access to the primary data?
In the article, OMB seems to be making a number of incorrect and
seriously flawed statements about feeds (saying that they can't be used
with other XML data, and implying that they make reporting complicated).
This is simply factually wrong. It would be quite feasible to develop a
government run Web application that lets recovery grant recipients to
fill out some friendly reporting forms and then this application can
turn these into nice XML data published within easy feeds. Feeds simply
package the XML reporting data with some simple descriptive information
that makes it easier to find, get updates, and manage the XML reporting
data. Feeds are good because they are so widely used and supported. They
are easy for software developers to use for other applications,
including development of tools to enable independent oversight and
understanding of the Recovery.
This is in no way more complicated or difficult than OMB's proposed
Federalreporting.gov. Such a system can be run for groups who can't /
won't run it themselves (the Federalreporting.gov model). Or, this
reporting system could be cloned and run / deployed by federal agencies
responsible for managing the recovery, thereby distributing the job of
the reporting, verification, and dissemination of data. Instead, this
article makes it seem that OMB thinks the only way to use feeds was for
everyone getting a grant to somehow hand - code there own feed!
Anyway, if this article is correct in describing OMB's thinking, then
OMB has some serious problems in understanding basic technology issues.
Obviously not everyone needs to know about the details of web
technologies, but OMB officials have a responsibility to make use of
well-informed experts. I know several excellent federal technical
experts who worked directly with OMB on recovery issues in the spring.
It is so frustrating to see this expertise ignored and see such
fundamental errors coming out of the agency.
Sigh.
-Eric
--
---------------------------------
Eric C. Kansa, PhD.
Executive Director
Information and Service Design Program
Adjunct Professor
UC Berkeley, School of Information
http://isd.ischool.berkeley.edu/
Office: (510) 643-4757
Mobile: (415) 425-7380
Fax: (510) 642-5814
---------------------------------
Josh Tauberer
> In October, agencies will be required to start transmitting reports
> from recipients to Recovery.gov. The material that is made public
> will be XML-formatted, Gavin said
Is this a quibble about what flavor of XML?
I shuddered in the beginning when I saw "feeds" enter the picture of
reporting. Feeds generally mean RSS or Atom format, and these are at
best irrelevant or at worst distracting and inappropriate for the
purposes of reporting on government spending. (We had better see a field
for "dollar amount" somewhere --- RSS and Atom have nothing to do with
that.)
And if OMB wants to invest in cleaning up the data they receive from
agencies before they pass it on, that's great! I would rather get
cleaned-up data than a mess of formats for *us* to have to normalize.
(This is separate from the issue of the data format.)
- Josh Tauberer
- GovTrack.us
"Yields falsehood when preceded by its quotation! Yields
falsehood when preceded by its quotation!" Achilles to
Tortoise (in "Godel, Escher, Bach" by Douglas Hofstadter)
Jon Henke
and Transparency Board in charge of Recover.gov to "the OMB has
decided..."?
-------------
Jon Henke
(202) 595-4323
Eric C. Kansa
Good discussion. Let me clarify.
No I'm not quibbling about XML flavors. The current OMB guidelines use
NIEM, an XML standard for government use. That seems fine to me. I'm
sure there are other XML standards that would work too.
Now about the feed issue. Feeds (atom in particular) are extensible. You
add the additional special purpose XML (in this case, Recovery reports)
to the feed to communicate your special purpose data. The nice thing
about the feeds is that they support these special purpose use, and are
so widely supported and easy to use by developers. Think of feeds as a
standard shipping container, with some convenient descriptive labels, to
ship content (and that content can be complex, special purpose XML for
the recovery reports).
But! The real significance of this isn't around using or not using feeds.
It's about making recovery data available for third-party, citizen
software developers. The RFP for recovery.gov says nothing about how
this will happen. The new OMB guidance says nothing about now this will
happen, and that article in Next.gov now says that there will be no
citizen access to the primary data. That is a very disturbing development.
Best!
-Eric
Joshua Tauberer / GovTrack.us
> and that article in Next.gov now says that there will be no
> citizen access to the primary data. That is a very disturbing development.
Losing primary data from agencies merely shifts the burden from the
agencies to OMB to providing the data in a useful and timely format. Is
there reason to believe OMB is going to be doing a worse job at
distributing this info than the agencies? Maybe having some tech people
at OMB reviewing the data, even if it introduces a short delay, is on
balance a good thing to make sure all of the data coming out of the
government is sensible.
Josh
Eric C. Kansa
and the latest guidance from OMB, nobody will be providing primary data.
Look at the documents, and you will see no discussion on how to provide
public access to all records relating to the recovery.
But I think there is some nomenclature problem here. When I mean
"primary data", I'm referring to the complete detailed record of what is
reported. Having it verified and validated in a timely manner before
release to the public is fine. But we should have the complete
(verified) records in machine readable formats.
That is what is now missing from the data disclosure plans for Recovery.
-Eric