Behind the sensation caused by Google Maps, image caching technology based on Client browsers is indispensable, I think this is a technological innovation-because no common WebGIS commercial product has such a map display method.
Go2map actually used this technology very early (when the new version of the VIP map service was launched last year), but the map service is still based on the Mapinfo product line, not very nice and high-quality (at least compared with Google, there is less gray border on the road), and the interface is a little complicated (personal feeling ). After the client image cache is used, the pressure on the central map server is relieved. You must know that the bus that used to query Beijing on go2map always waited for timeout.
Carefully analyze the internal mechanism of Google Maps (which can also be inferred from the Operation Experience on the Interface ), this WebGIS system is essentially different from the previously developed systems such as ArcIMS and mapxtreme:
Creating a map in advance is the first thing to do.
Map slicing, Which is downloaded by the client at the same time. When the map window is moved or scaled, You need to download a new image to fill the new area. Here, the multi-thread simultaneous download function of the browser (such as IE) is fully utilized. In addition, you do not need to access the server again to download the downloaded image. The disadvantage is that many scripts are required for client writing.
Because the server does not generate Map Images Based on Vector data in real time, the map data cannot be updated in time. However, if you look at it carefully, it is not difficult to find that most of the data on Google maps is road traffic and administrative boundary. Other poi (point of interests) information is not fixed on the map, in this way, people feel that the map is a little monotonous, but the high quality of the Wei film image makes up for this deficiency. Google uses the client script technology to display users' search results (that is, various poi data) in the form of a "pin" image on the client browser, such as the famous official pizza store example. On the other hand, I really admire Google's design: poi data is often updated, while basic data (Framework dataset, such as high-level roads, bridges, cities, lakes, etc) there are few changes. Even if some small roads change, it is very convenient to generate map patches in this range. This workload is relatively small.
As mentioned above, poi data is actually dynamically generated (or static? For example, Google's Web Crawler finds a company's portal website, analyzes the content structure of the page, and finds the contact information: #, ** Road, ** city, geocoding technology allows you to find its location on a map.
Due to the pre-generation of map images, the "Layer Control" function that was often provided by previous WebGIS sites cannot be achieved. In fact, this does not affect the user experience-Google map is so clear and clear, what is it about?
Some people may think that they are not used to abandoning the "eagleeye" window, that is, the thumbnail, because you do not know where to view it. However, it makes no sense to see how small the display box on the eagleeye is when you place the map all over the world in a very detailed place.
The info window pops up on the map to display detailed information. xml and XSLT are used in the underlying implementation, which is indeed an impact on the interface and vision, this is indeed a good idea-it saves page space and facilitates user operations.
"Graph ". This feature is painstaking. You can see that people are doing well. In the pop-up info window, there is a small map to show the surrounding conditions. You can also scale the map, it's a bit similar to placing a magnifier there. It's absolutely amazing.
The crazy use of JavaScript scripts is just like the current web design is back to Javascript, especially since the popularity of Ajax, it seems that everyone has paid attention to the script on the browser. The Google Maps script is not a two-segment script, but contains more than 10 million lines. Take the path display function on the client as an example, it must involve XML parsing, path data transcoding (Google performs encoding when the location coordinate of the Transfer Path), VML rendering, and access the server if it is not Internet ExplorerProgramDraw the path and return the image ...... There is no limit.
Google's local search service in China has also launched a beta public trial version, and the diagram in Beijing is also very detailed. I think the biggest difficulty may be domestic place name resolution, we need to know that China's street numbers, block numbers, and postal codes are not clearly divided. This is a problem of urban planning and construction ~
Since the development of GIS, it has gone through several stages of system-science-service transformation, and the boundaries of these three processes have become increasingly clear. As now, someone is working on system, while others are working on science, some people are working on the service. In the final analysis, it is still an IT technology. I personally think that whether it is engaged in scientific research, software development, or you are engaged in service money, they are all inseparable from computers and networks. Without computer technology, we cannot figure out any new things ~
