Search Done Right

What was unveiled during the SharePoint 2010 conference has been dissected, commented upon, interpreted, etc. all over the web. So to avoid repeating what has already been said and delve deeper into the SharePoint psyche, after exploring the Freudian slips in my last post, I propose to explore the taboos: what was not talked about during the SharePoint conference?

1. Why isn’t Fast the de-facto solution?

Microsoft’s pitch is that it made Fast a cookie cutter product. All the goodies and advanced features but none of the pain. These claims lead to the obvious question: If Fast is so powerful, so easy to install and integrate, why didn’t you make it the de-facto solution?

Having a single solution (possibly with some tiers) would make total sense:

• Simplify the product offering
• Enable the engineering team to focus on a single solution instead of dividing their effort in two
• Avoid the high cost of customers needing to switch from one solution to the other, or possibly to support both solutions simultaneously (the People search seems to be always based on the native search).

2. Where did all the Fast connectors go?

The current connector offering reads as follow:

• SharePoint
• File Systems
• Exchange public folders (i.e., no security)
• Lotus Notes
• Documentum (upcoming)
• Databases
• JDBC Database (Fast only)
• Web Crawler (Fast only)

That is, the native SharePoint connectors haven’t changed since 2007 and the only Fast connectors kept are actually redundant. What happened to the dozens of existing Fast connector? What’s Microsoft connector strategy?

3. What about all the Fast scripting and advanced features?

Many of the sessions talked about some integrated administration through the standard SharePoint UI. The Fast installation session on the contrary suggests that most Fast administration is done through PowerShell—that is good old shell Unix-like commands without any UI. (At some point in that session the presenter got really excited. “Look at this, it’s really cool”, he says, then types a man-like command to display some help. The audience was unimpressed and the irony of a Microsoft engineer being impressed by some 30 year old Unix-type user experience totally escaped the speaker.)

So it’s unclear what is and is not supported. The PowerShell are just some wrappers to some fairly basic administrative task. But what about all the Java APIs, what about modifying the query pipeline, the document conversion pipeline, using other non-Microsoft centric programming languages?

4. What will be the de-facto solution in the future?

At some point Microsoft will have to choose between the two solutions. It does not make sense from a marketing, engineering or customer standpoint to maintain two separate solutions in parallel. The big question is: which one? The common assumption in the blogosphere is that Fast will be the choice. I am ready to bet the opposite for several reasons:

• Many of the key Fast engineering staff left, a few of them to create a competing offering (Attivio).

• Microsoft is gaining some expertise and confidence in search and NIHS (non-invented here syndrome). NIHS is as prevalent in Redmond as in many software companies. Historically, that’s what usually happens: new acquired competing technologies are stripped from their most reusable components and left to die.
• Key components of Fast have been dropped (e.g., security and connector framework)  while the native SharePoint has gained some features (e.g., language handling) taken from the Fast platform
• Is Microsoft really going to pursue a UNIX like administrative interface for one of their product?

As demonstrated by the weak connector strategy, Microsoft is more concerned about SharePoint itself than external systems. As such, it makes much less sense to integrate a multi-purpose search engine rather than optimize their own engine for their own repository.

What do you think? Has the Fast acquisition lived up to its initial promises? Will the next version of SharePoint still have two separate search solutions? Will Microsoft favor usability—by limiting the set of features or flexibility—by exposing non-Microsoft technologies under the Fast hood?

This entry was posted on Tuesday, December 8th, 2009 at 9:00 am and is filed under Uncategorized. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.