Why Yahoo Maps chose AJAX

Ted Patrick has posted a very frank and thoughtful post discussing the reasons why Yahoo has moved from Flex to AJAX for the latest version of Yahoo Maps. To summarize he identifies 3 main reasons for the shift :

• Shortage of Flex developers
• Problems integrating Flash advertising into Flex applications
• Complexity of migrating from Flex 1.5 to Flex 2.0

Despite these points it seems an odd decision in a year when so many great Flex applications have emerged.

Silverlight vs Flex : the hype and the search trends

There has been no shortage of hype about Microsoft's Silverlight during 2007. But I was playing around with Google Trends the other day and started wondering how well the hype translated into user interest. The graph shows the relative number of Google searches on the terms Flex and Silverlight. The most interesting things was that although Silverlight wasn't far behind in terms of news reference it was a very different story when you looked at actual searches, even when there were news spikes for Silverlight. Obviously search trends aren't the ultimate indicator for market interest but they are still help to give some idea of whats really happening among developers. These results can also be skewed when a search is done on terms that may have alternative contexts. For example, my original search included AJAX and it seemed that The Netherlands were obsessed with AJAX. This seemed odd until I looked at the news items and realised there was a soccer team named AJAX.

Data Loading Benchmark

James Ward has created a Flex application for benchmarking data loading for RIA's. Thinking about the results from this tool there was something I thought worth mentioning. Firstly, what is clear from the results is how much faster Flex is for parsing and rendering data sets. It's ability to parse XML using E4X is extraordinary compared to AJAX. Therefore Flex is ideally suited to anything that requires repeated and regular data requests. But one thing that isn't immediately evident from the benchmark is that the applications file size also needs to be taken into consideration. Flex applications can quickly become quite large. While most AJAX frameworks are comparatively small. Consequently an application with fewer data requests might be better suited to AJAX. You also need to consider how often users will use the application because caching will reduce the impact of file size if the application is accessed frequently. So the more often an application is used the more appropriate Flex will be.

I will be very interested to see Silverlight included in these tests so that we can start to filter out the FUD from the facts.