We use PathLoss (www.pathloss.com)
It is a powerful program, and comes with 3 second data of the entire US. This is a very powerful program, at a very reasonable cost, (approx $2,000).
An Excell spreadsheet could never model accurate coverage and propogation, like a full blown dedicated application can.
Als0, programs such as PathLoss allow you to conduct calculations several different ways (different algorithyms) as well as build complicated predictive models.

I would not rely upon a simple spreadsheet to accurately predict a wireless network.

Given the low power and typically limited range of wireless Ethernet and 802.11(x), the 'terrain data' would have to include every structure in the vicinity.  This is a different problem (MUCH more localized) than planning, for example, a typical state-wide paging network.  In other words, having the elevation data for your state, county and city isn't applicable to the installation of such a localized network as wireless Ethernet or 802.11(x).

There would obviously be a point where it would be much easier to ty it than to model it.  Just my opinion...

I use Comstudy from Radiosoft.  I use it for radio coverage for mountain top repeaters/base stations.  I do agree with VE1BLL that sometimes it's easer to try it than model it.  Some of the software vendors offer free demo versions (with limitations).  Comstudy does give you the ability to add "features" or obstrructions (buildings etc.)  Comstudy is not perfect but does have great tech support. And it's not cheap, so it depends on what you need it for and how often do you plan to use it.

Also visit:
http://www.rfcafe.com/vendors/components/software_home.htm
and
http://www.rfglobalnet.com/content/homepage/default.asp
(need to signup for rfglobalnet and click on the download library).

If you are deploying 802.11 in a limited fashion, I would probably not spend the money on a full blown propagation package. But, if you wish to deploy extensive services over a greater distance, (i.e. campus wide deployment, large roaming area, etc.), a lesser expensive propagation package could help make the deployment less stressful.
I have always been a proponent of planning any communications deployment, no matter how small.

For limited 820.11 your right, theres nothing like an Ipaq with a wireless card and a GPS plugged into the serial port. Quick easy and painless. But there are spread spectrum radios that can do 802.11 up to eight miles like any other licenceless shot. I was wondering mostly about the packaged softwares and dealing with the PHB mentality that "theres all these cheap options out there". I even had to fend off a suggestion that "Garmen and Magellis make topographical programs, don't they?" I came from the integration side of the tracks to begin with and I know theres nothing like a field test but now I see why those paper studies were so off. I was looking for anyone who had tested multiple programs and their insight. Thanks for the help and links gang.

I tried several radio propagation packages before I bought mine.

EDX Signal Pro  www.edx.com- very powerful, detailed analysis, will model propagation in metropolitan areas with building detail. This package appears to be geared towards commercial/service providers, and is more than the small guy would need. Have to purchase terrain data for your specific area, could get very expensive if your communications span several states.

Softwright Terrain Analysis Package (TAP) www.softwright.com- again very powerful, great models. Must purchase the base software and each module to conduct specific functions (i.e. coverage, database, project mapping, etc.), which drives the price up. Also must purchase terrain data. This package also seems to be geared towards larger operations.

I tried several smaller packages, priced between $500-$1000, most were barely more than a supped up Excel spreadsheet. The programs would only allow modeling of one path at a time, and would only give one terrain view. They also would not delve into the deeper reflection/refraction calculations. I found these too light for my requirements.

PathLoss 4.0 www.pathloss.com- this is the package I chose. The software is very powerful, but will not require a doctorate to use. Allows modeling of complete systems, can store site info in an internal database, and can display and print many different reports. The package comes with 3 second topo data for the entire US, (big plus), which has been accurate enough for the systems I have deployed. If I were to design a 12 hop microwave system, I would purchase more detailed topo data, but what is supplied works for me. One drawback, the documentation for the product stinks. It gives good examples, but doesn’t explain how to initiate specific functions. Luckily, the tech support for the product is great. The cost was also good, about $2000 (US)  for the whole package.

I have deployed large WLAN systems, utilizing 900 Mhz and 2400 Mhz spread spectrum systems, (proprietary protocols not 802.11), in several states for our company. Conducting terrain/propagation analysis has ensured successful implementations, even several states away. We are currently deploying wireless video surveillance throughout the Midwest and Eastern US utilizing 5.2/5.7 Ghz WLAN systems. Again, terrain/propagation analysis is a must for us.

If you would like more specifics, just ask.

I found a really nice free one on the net its not the caliber of radio soft or planet but it is an effective tool
it connecte to the government usgs sites to download terrain data

http://www.cplus.org/rmw/english1.html

www.ecommwireless.com has several simple point to point calculators for free.