MapForums

This morning’s news [SearchEngineLand | Chris P Blog] that Ask.com’s maps and directions site has migrated from an application built on deCarta’s (formerly Telcontar) mapping platform to Virtual Earth got me wondering about the future of self hosting of mapping infrastructure for high volume sites. The question isn’t if there is a future here at all, there is. In fact the Virtual Earth appliance is offered to our customers with really specialized needs (off the grid or lots of custom pre-rendered data are a couple of good cases)* My question is how big this already niche market will become.

If your web based map app calls for supreme control and customization of cartography you historically would build your own cluster around ESRI’s universe of software and get to coding. For small to mid sized apps this was OK assuming you could make the development investment, but it broke down when scaling forced you to build out that cluster. this is where hosted solutions like Virtual Earth come in – you trade off a level of control (you don’t like our highway shields? sorry) for infinite scalability and performance.* Both are viable models for building your app and you could go with whichever worked for you. But Telcontar was offering a third option that always seemed to me to be the worst of both worlds – the limited control of hosted with all of the hosting and maintenance costs of building your own. Not everyone agrees with me of course and that’s the beauty of our free market, but i still don’t understand why someone would go this route for core maps/routes/geocoding. I’m biased as i want to see everyone building with VE of course, but if you have some insight into the deCarta model clue me in with a comment.

Click here to view the full post.

Anyone using the Manifold web server? I wonder how it compares to DeCarta's offering..

Could Ask.com have used something like Manifold or UMN MapServer?

Eric

__________________