View Single Post

  #7 (permalink)  
Old 02-02-2007
Dimitri Dimitri is offline
Junior Member
White Belt
 
Join Date: Feb 2007
Posts: 1
Re: Building a GIS using mappoint technology

Quote:
Originally Posted by ericwfrost View Post
If you want to be different you could do the project in Manifold, but you might have to defend your use of it from professors who have not heard of it or think it's not as powerful as the ESRI platform.
That depends upon how much they know about GIS or modern computing technology.

Those who don't know much about modern technology will be happy if you use ESRI. Those who are up to date with modern notions such as 64-bit operating systems, multicore processors, .NET and other modern Windows notions will be surprised if you don't choose Manifold.

I realize some may think I am making a joke, but unfortunately it is frequently the case that some university departments lag many years behind progress in technology. To take just one example, in 2007 we take it for granted, even for our middle-school kids, that serious people understand the benefits of computers which run 64-bit Windows and include a modern, dual-core processor like the Intel Core 2 Duo or an AMD Athlon x64 x2.

Your 13 year old kids know all about that stuff and will insist at a minimum on that for gaming (it costs about the same as old-fashioned 32-bit stuff), yet you'd be amazed how many university folks doing GIS will have no idea what you are talking about when you talk 64-bit Windows and Intel Core 2 Duo.

Yet applying 64-bit Windows with a reasonable amount of RAM is by far the fastest performance boost you can do if you want to put together an Internet Map Server (IMS) enabled web site. It makes a huge difference in the scale and responsiveness of web applications.

Microsoft, Intel and AMD have spent billions of dollars to make such breakthrough performance available on the desktop, even down to almost the least expensive desktops out there, so it is not exactly something to be proud of if one is not aware of the incredible performance benefits of using Windows x64 for web servers.

I raise this as an example because ESRI has zero ability to run 64-bits in Windows x64, nor do they take advantage of dual core processors. Anyone who thinks a GIS that is limited to only 32-bit performance and can't do multithreading in an era of multiple-core, 64-bit systems, well, if they think such an antique system is more powerful than Manifold they are quite simply a dolt.

Another thing to consider is that almost all serious IMS-enabled web sites involve massive and sophisticated integration with a DBMS. Manifold supports direct connect to Oracle Spatial using Oracle's own native Oracle Call Interface (OCI). ESRI cannot do that, relying instead upon its own old-fashioned and nonstandard ArcSDE middleware. That's much slower than working through Oracle's own interfaces and it also traps your data within an ESRI silo. In contrast, by working direct with Oracle's own spatial standards Manifold reads/writes/edits data in a way that allows any of hundreds of other applications that use Oracle standards to interact with that same Oracle data, transparently.

If your taste runs to SQL Server or other DBMS, Manifold also connects to those using "open" standards, such as OGC WKB, instead of some proprietary mechanism such as ArcSDE.

Finally, the nice thing about using Manifold is that the object model is identically the same whether you are using Manifold as an interactive desktop application, as an Internet Map Server or as an objects library. In contrast, ESRI products use different object models, so what you program in a test bed case using some interactive desktop thing like ArcGIS will likely need significant reprogramming when you attempt to implement it in ArcIMS. That's a strikingly clumsy, old-fashioned and inefficient approach.

I don't get the sense from your original posting that you are very familiar with the technologies involved. Referring to a GIS-enabled web application as "a GIS" shows a somewhat dated conceptual approach that you might want to update in order to make the best use of contemporary tools. In general to create a sophisticated GIS-enabled web application there are three areas you need to master:

1. Webstuff. You must know how to create web applications using a programming technology appropriate to the web server involved. In the Microsoft world that means expertise in IIS and ASP or ASP.NET using appropriate Microsoft languages.

2. DBMS. Most web guys already have a DBMS they know and love. If you are choosing one ab initio, I'd recommend either Oracle or SQL Server, both of which are available in free Express editions. Oracle's Express edition includes "Locator" capability from Oracle Spatial, which enables fabulously performant and sophisticated geospatial data storage when used with a modern GIS like Manifold. That's a lot to get for free.

3. GIS. You must master whatever GIS package you intend to use. This is easier if the desktop package you will need to whip your data into shape is the same as the IMS package that works in conjunction with the web server to publish it on the web. Manifold has that advantage. Neither ESRI nor MapServer has that advantage.

By way of integration it also makes life much easier if your GIS package has modern, seamless integration with the Microsoft technologies used in your web server. I presume that since you posted to a MapPoint board you're not some sort of open source kook who wants to spend his time re-inventing the wheel and therefore you will be using Microsoft technologies. No other GIS comes close to Manifold's tight integration with Microsoft, whether it be deep technology like support for Vista and 64-bit Windows, or simple things like integration with MapPoint and Virtual Earth as geocoding data sources.

It's true that if your project is simple and your GIS/IMS, like Manifold, is well-integrated both vertically and horizontally with web environments, DBMS and Microsoft technologies, you could use Manifold to create a GIS-enabled web application with zero programming and relatively slim acquaintance with the DBMS and webstuff skillsets. That's certainly possible and people do it every day using default templates. But I don't get the impression that if you are doing this as a university project you intend such a simple thing and so I raise more explicitly the need to have both web programming skills and DBMS skills in addition to GIS skills.

I am, of course, a Manifold guy, but as you look closer you'll see what I write is true and, if anything, undersells the matter. It is no accident Manifold outsells ESRI by about 100 to 1 (on a unit volume basis) in Windows markets. See manifold.net for details.

Best regards to all,

Dimitri
Reply With Quote