These answers are great, thanks to both Joe and Kirk Borne. More
inputs are welcome.
I should have mentioned that the alternative strategy is to specify
actual types of processes (services) that were performed, but for this
problem we are consciously speaking of levels.
I had to laugh at the wayback machine reference -- I remember spending
half a day looking for the NASA levels (virtually nowhere to be
found), which I thought was ironic given it's the one thing everybody
points to in this realm.
John
On Mar 26, 2009, at 11:35 AM, Joe Hourcle wrote:
> On Thu, 26 Mar 2009, John Graybeal wrote:
>
>> I am aware of one vocabulary for data processing levels, the NASA
>> Levels (1, 2, 3) list.
>>
>> Does anyone know of any other vocabularies that describe the level
>> of data processing embodied by a data set?
>
> There's actually multiple versions of that list.
>
> There's the original CODMAC (Committee on Data Management,
> Archiving, and Computing) list, which is 0,1,2,3,4 :
>
> http://ilrs.gsfc.nasa.gov/reports/ilrs_reports/9809_attach7a.html
>
> And then there's the EOS (Earth Observing System) list which
> differentiates between '1A' and '1B', and they put some extra
> qualifications on the other levels (eg, L0 mentions that the
> communications artifacts have been removed)
>
> http://observer.gsfc.nasa.gov/sec3/ProductLevels.html
>
> The only good reference that I knew of that was less than 10 years
> old I've lost track of when NASA redid a website and broke the
> link ... but it's in the wayback machine:
>
> http://web.archive.org/web/20070216014019/http://science.hq.nasa.gov/research/earth_science_formats.html
>
> (It's from the EOS folks, and it talks about both the levels, and
> other ancilary data products)
>
>
> ...
>
> The heliophysics community uses basically the same levels, with a
> few additions and modifications:
>
> Eg:
>
> '0.5' is the data that was downlinked from the satellite, but
> might have some form of lossy compression applied or missing
> blocks.
>
> '1' in heliophysics suggests that the data has been adjusted for
> known sensor issues, but not necessarily in physical units. They
> also don't use the restriction that it be at full resolution.
> (eg, "SOHO/LASCO, L1, 2x2 binned" suggests that they've adjusted
> for sensor effects, but it's at 1/4 the resolution)
>
> 'ql' (quicklook) is used to denote data that's had some form of
> sensor-specific processing applied, but was typically applied with
> speed in mind, not accuracy, so shouldn't be used for science.
>
>
> Some missions define specific reference levels. For example, the
> upcoming SDO (Solar Dynamics Observatory), defines 0, 0.1, 0.3, 0.5,
> 1.0q, 1.0, 1.5q, 1.5, 1.5p:
>
> http://hmi.stanford.edu/Status_Reports/Quarterly_April_2008_JSOC-SDP.ppt
>
> ...
>
> I also presented last year at CODATA, ASIS&T and AGU on the need for
> a reference model to describe data, especially the relationships
> between data (that we start seeing when we're storing multiple
> levels and editions of the data). I proposed using FRBR (Functional
> requirements for Bibliographc Records) as a base. (for which, I've
> realized some issues, and really need to make some updates), but the
> various posters, talks and the pre-print from the ASIS&T short paper
> are all at:
>
> http://vso1.nascom.nasa.gov/vso/misc/
>
> (see the file named '_README.txt' for a key to what's what)
>
> The quick summary, for those familiar with FRBR, is that I treated
> the different forms of calibration as different "Works", and the
> other forms of processing (compression, binning, other reductions,
> anything lossy) as different "Expressions", while tracking the
> packaged forms of the data as different "Manifestations".
>
> I was focusing more on the fact that there were relationships
> between the objects, not specifically that there were discrete types
> of processing that might be applied.
>
>
> -----
> Joe Hourcle
> Principal Software Engineer
> Solar Data Analysis Center
> Goddard Space Flight Center
John
--------------
John Graybeal <mailto:[log in to unmask]> -- 831-775-1956
Monterey Bay Aquarium Research Institute
Marine Metadata Interoperability Project: http://marinemetadata.org
|