On Thu, 16 Dec 2010, Gerry Creager wrote:
I'll try to figure out what we did wrong here. We're snagging the same
EMWIN feed Gilbert is, but could have fumble-fingered something, since
I'm still short of calibrated fingers since shoulder surgery...
Unfortunately, one can not seamlessly plex together the EMWIN/WeatherWire
with the NOAAPORT/SBN dataset using baseline software. I forget which
system does it, but one of them arbitrarily tacks on an extra DOS line
feed to every product, thus LDM sees the same product twice, since the
line feeds are used to compute the product checksums.
For NWSChat, I get a direct socket feed from the Gateway/TOC that contains
the plexing of, I think, 4 different internal NWS systems. This feed has
duplicates all day long. So when Gilbert was hoping for a direct backup
IDD feed from the TOC, this problem would have caused lots of pain for
others... Anyway, I ended up writing a pre-conditioner to my decoders
that striped extraneous line feeds and computed product checksums that
way. This works for 99% of the products, but there are a few more
products that have extra control characters inserted (namely from SPC)
into the non SBN feeds.
My impression is that the private sector weather folks have been dealing
with this issue for a very long time and it has been frustrating for them.
I've been bugging the TOC/Gateway folks about it, but they are extremely
short of resources and dealing with ancient IBM computer systems. This
issue is on their TODO list though.
I wish LDM had an option to arbitrarily strip trailing line feeds before
checksum computation. Before somebody says this would be a performance
killer, it isn't. :)
daryl