|
|
| GeoCommunity Mailing List |
| |
| Mailing List Archives |
| Subject: | Re: GISList: Project clairification |
| Date: |
11/22/2002 11:01:45 AM |
| From: |
Robert Heitzman |
|
|
Forgive the incomplete links - use google.com to find the sites...
>#1. I have a background in the web and network
>technology. The server we will host this on will be a
>clustered Compaq dl760 with 8 processors and 16 gig of
>RAM. The data will be housed on a SAN linked to the
>server on a fiber channel backbone...
>
That should do!
>#2. The GIS web interface platform must run on a J2EE
>engine.
Well, your making it pretty hard on yourself using this environment - good
job security however. You will pretty much have to build your own solution
based code source you acquire from somewhere.
>
>#3. The DB of choice is Pervasive SQL2000...
Ditto
>
>#4. The OS of choice in our minds is NetWare 6...
Ditto
Now that your job is secure and no one else in the world could work on your
system...
There are several source of GIS data that could be incorporated.
NOAA has raster images of many of their charts available. (search for noaa
chart maker) Unfortunatly they have been bought off and there is a middle
man that may prevent you from using the public domain data. A good lawyer
should be able to extract the data from NOAA if you are patient. Less
useful raster version of charts are available for some areas.
NOAA also has something called the 'coastline extractor'. This will give you
lines representing the shoreline that could be useful for a high level GIS
navigator. The current largest scale readily availabel is 1:70000 but I
think they may be working on a larger scale set for Florida and a few other
locations.
There are also so large scale raster maps around for ports.
Some USGS data (DRGs) may also be useful as well as whatever aerial photos
DOQQ are availabel for your AOI (area of interest).
A good GIS tool for working with all this data would be Manifold Enterprise
Edition ~$375 each with some good volume discounts if you will have a crew
working on this. Manifold is a full featured GIS and is inexpensive. The EE
edition has good data sharing via backed SQL servers. It may work with your
flavor of SQL server as it doesn't require esoteric spatial extensions as
some product do. (The coastline extractor exports in some strange formats -
I have a script in the works to import the extracted coastline. If you need
it drop by the Manifold-L list on the Directions Magazine site.)
Manifold can import a wide variety of formats (not the propritary version of
the NOAA charts but the version your lawyer would pry out of NOAA which has
been preverted by the vendor)
The final product can be exported in a format acceptable to whatever map
server you build.
Manifold has a web server built in but it won't run on your web server - IIS
is required. It might be useful for in house use or for some sidebar web
services if you have a IIS server available.
><snip>
>
>So if you folks tell me exactily what we need to do to
>get this project off the ground we can do it. A lot
>of your info has already been very helpful. I have
>two questions about the info:
>
>"Is this to be a temporal model -- i.e. will it
>account for changes in time? Remember, these
>waterways are dynamic systems. How do you want to
>account for time in this data model?" This is a
>Question we thought of but we're not sure of the
>answer. We would like it to have a temporal
>component, but we need more information.
Perfect data is great but I doubt you have the resources. The best you can
do is use what data is available. Most of the data mentioned above is on a
2-10 year update cycle.
The NOAA charts are probably the most current. It might be best to figure
out how to sign up for the data from the vendor that paid off NOAA for the
sole source license.
>
>"How much error are you willing to tolerate? Should
>your data be accurate to 1 meter? One centimeter?
>One kilometer? Your data will have errors. How much
>you will tolerate is a major factor in the time and
>cost." In terms of error, I never want to see any
>error. However, considering we will be gathering this
>data from a variety of sources, how much error is
>acceptable? What is the cost associated with reducing
>that error?
None of the data above could be used for navigation and you would have to
have disclaimers all over your site to clearly state that. That said
accuracy is more a data isssue than a position in space issue. That is if
the intent of the site is to provide travel info, like where one can find a
fuel dock, the probablem is mainly one of keeping that kind of data current.
Not a trivial task.
In rough GIS terms a 1:24000 scale (DRGs - topo maps) are accurate to about
50-100' for some map objects like coastlines - good enough to find a fuel
dock. Take a look at the NOAA sites, they probably state their accuracy. If
you want to pi
|
|
Sponsored by:
For information
regarding
advertising rates
Click Here!
|
