TIC WG A, minutes from teleconference, April 26, 2004

Participants: Mitch Withers, Woody Savage, Harley Benz, Egill Hauksson,
Paul Whitmore, Won-young Kim, C.B. Crouse, Ron Tognazzini, and Rich Eisner

WG timeline states a draft spec by May 1.  This will not be met but
is not of great concern considering time remaining.

In attempting to create table of products and performance issues Withers
once again got confused and appealed to the committee for assistance.
Are these standards really product specific or is it more general such
as outcome specific.  Similar discussion took place within NEIC resulting
in two camps that are now coming together: a product oriented view and a
process oriented view.

Standards will necessarily speak to both products and processes.

National and Regional need to be integrated and trasparent to recipient.
To accomplish there must be adherence to minimum standards.

Agreement on standards is controversial and difficult to develop a
consensus.  While this group need not necessarily build a consensus
external to the group, this should be kept in mind when developing
standards.

Not all contributors are capable of meeting all standards or providing
all products.  For products that are provided, standards must be met.

Is reliability (e.g. redundant systems) a standard?  For the system yes,
but not necessarily for individual components.

Can we take appropriate tables from the TIC ofr that includes time and
just add on error bars?  Not complete but perhaps a good place to start.

Need to allow error/quality improve with time.
Better product is not always last arriving version.
Need to have limited release capability to assure quality is
tied to audience (e.g. EMA may be able to accept accuracy at the
expense of speed but media shouldn't).

Even if we agree on things such as amount of error, methods to calculate
error differ.  Hence also need to speak to the process and procedure.

Pay close attention to TIC ofr chapter 6.

Note that engineering data and processing is additional nearly parallel
but separate set of standards.

Don't forget standards for metadata.

Standards will depend on product classification: Rapid, Reviewed, Archive.

Mitch will move forward with table.  Needs to have 4 dimensions:
Products X Time X Quality X Reliability/Dissemination

Next conference call same time (1pm EDT) on May 26.