In a recent post on the Microsoft Enterprise Search Blog, Bjørn Olstad (CTO of FAST) published a post titled, Innovation on Linux and UNIX, which explains the longer term technology strategy for FAST with respect to supported operating systems. The post describes an increased focus of innovation for the product, for Windows operating system installations, while gradually deprecating support for Linux and Unix installations.
I can sort of understand the negative reactions to this announcement and the notion that customers may turn to other search platforms to maintain their Linux/Unix operating standards. However, I believe that this is a truly good decision by Microsoft, which will result in a higher quality product.
The reality is; in order to push innovation of the search platform forward beyond its current boundaries of functionality and performance, it is an inherihent requirement that the low level levels of the architecture be tightly bonded with the operating system. By not making this long term decision, a forced common denomintor is then imposed on the platform at higher levels of the architecture, ultimately inhibiting the product. Multiple OS support for installations actually sets the platform back. We've seen this as recent as 5.3 when we are required to install FAST ESP as a 32 bit application along with Java runtime 32 bit. This is a 64 bit world now, get with the program.
My sense is that the product roadmap will offer great improvements in the document processor and query processor pipelines, linguistics, user experience, etc. These improvements are required to allow search to become the leading functional component for indexing, transforming, and presenting not only content but navigation, functional components, and structures on Web sites. The relationship that search platforms will have with tomorrow's Web applciations, is comparable to the relationship that relational databases have with yesterday's Web sites. This innovation simply cannot afford to be spread thin.
No comments:
Post a Comment