In addition to what I've seen from the posts list upto now, most of which
are my wishes too, here are more items from me:
1. support some level of viewer personalization for things like a bookmark
function (like MG6.5's viewer bookmark but should include layer on/off
parameters), being able to save a user session map including redline/markup
work for later retrieval, a query-based thematic mapping function (like
DAT), etc. Yes, the user session repository and MGE APIs can do the first
two jobs, but what I prefer was some built-in commands or high level
components for us to easily incorporate this type of funtionality to support
end users' flexibility and personalization.
2. Towards/incorporate Model View Controller (MVC) model: MGE should be
enhanced and integrated with industry best practices and approaches such as
SOA, Model View Controller (MVC) model, 3-tier/n-tier application
development structure to separate presentation layer from business logic and
data access layers in order to promote the reusibility and interoperability.
More samples for implementing these standards and best practices are needed
and well documented.
3. Usage statistics functionality: MGE needs a usage statistics
functionality from the MGE Server Admin which is even a better one than MG
6.5. The function should provide the user access statistics and also the
access statistics on data resources in repository so that a MGE server admin
can easily produce a report on how many unique users are accessing over time
and by data resources.
4. Support load balancing and failover: MGE now does not support Windows
load balancing or its own solid load balancing and doesn't have a server
failover capability. Real load balancing and failover capability is really
important to a local government like us where our Web-GIS system VanMap has
become a mission critical one and the system performance as well as
availability is a crucial issue from a system architecture point of view.
5. Better support spatial analysis and data mining functionality from
existing open GIS databases: MGE as a WebGIS server should position itself
properly to take advantages of existing spatial databases (such as Oracle
Spatial and ArcGIS SDE) for spatial analysis and data mining. For doing so,
better interfaces like FDO APIs accessing the spatial databases are required
so that MGE (or FDO APIs) does not have to reinvent the powerful spatial
analysis and data mining functionality available from existing open spatial
databases. So the focus of MGE enhancements should be on those useful
packaged GIS functionality (like redline, routing, overlay analysis),
interoperability capability (like OGC Web services) and cartographic part
(like thematic mapping), as well as metadata management.