• Home
• About
• Writers
• Landscape
• Wizards Index
• Fraud
• Wanna Be Like Larry?

Microsoft Fast for Portals

Author’s Note: The images in this Web log post are the property of Microsoft Corp. I am capturing my opinion based on a client’s request to provide feedback about “going with Fast for SharePoint” versus a third party solution from a Microsoft Certified Partner. If you want happy thoughts about Microsoft, Fast ESP, and search in SharePoint environments, look elsewhere. If you want my opinions, read on. Your mileage may vary. If you have questions about how the addled goose approaches these write ups, check out the editorial policy here.

Introduction

Portals are back. The idea is that a browser provides a “door” to information and applications is hot again. I think. You can view a video called “FAST: Building Search Driven Portals with Microsoft Office SharePoint Server 2007 and Microsoft Silverlight” to get the full story. I went back through my SharePoint search links. I focused on a presentation given in 2008 by Two Microsoft Fast engineers–Jan Helge Sageflåt and Stein Danielsen.

After watching the presentation for a second time, I formed several impressions of what seems to be the general thrust of the Microsoft Fast ESP search system. I have heard reports that Microsoft is doing a full court press to get Microsoft-centric organizations to use Fast ESP as the industrial strength search system.

Let me make several observations about the presentation by the Microsoft Fast engineers and then conclude with a suggestion that caution and prudence may be fine dinner companions before one feasts on Fast ESP. Portals are not a substitute for making it easy for employees to locate the item of information needed to answer a run-of-the-mill business information need.

Observations about the 2008 Demo

First, the presentation focuses on building interfaces and making connections to content in SharePoint. Most organizations want to connect to the content scattered on servers, file systems, and enterprise application software data stores. That is job one or it was until the financial meltdown. Now organizations want to acquire, merge, search, and tap into social content. Much of that information has a short shelf life. The 2008 presentation did not provide me with evidence that the Microsoft Fast ESP system could:

• Acquire large flows of non-SharePoint content
• Process that information without significant latency
• Identify the plumbing needed to handle flows of real time content from RSS feeds and the new / updated content from a SharePoint system.

Second, on the second viewing I noticed a number of different programming interfaces. On the first watch of this one hour lecture, the detail of the methods caught my attention. On the second pass, I noticed these interfaces:

Third, the presentation included a system diagram for Fast ESP. I did not pay much attention to this block diagram on the first viewing, but on the second run through my thought was, “Am I looking at a simple idea that has become incredibly complex over time or was I looking at a very complicated approach to search that is now going to be presented as a simple search upgrade?” Here’s the diagram:

Beyond Search Concerns

Most organizations struggle with search. What my thoughts are this morning are gated by my first hand experience with organizations awash in information and unable to locate what is needed to handle routine business decisions. I am also aware of the need in some organizations to find a way to tap into the flows of real time information that can provide important clues to the world outside the organization.

First, the Fast ESP system that was unable to generate sustainable revenues to remain viable as an independent company appears to be unchanged. The misstatement of revenues is a clear indication that the plumbing itself may have some issues. An investigation into the Fast Search & Transfer operation continues. Microsoft, based on the information in this video, has chosen to rely on the Fast ESP largely as it was at the time of the acquisition.

Second, the video provides useful examples of presenting SharePoint information in a way that makes it easy for the user to narrow a result set. The problem with key word systems and lists of results by metadata categories such as Author or Time may not help the employee who does not know much more than “Smith invoice”. In short, the Fast ESP demonstrations in the video struck me on my second viewing as somewhat confusing to me. If I can’t track what I am looking at, how will the average employee know how to use Navigators to narrow a result set. Many organizations have one document that provides the answer to a particular query. In that situation, why not display what the employee needs, stripping away the unnecessary elements for that query.

Third, the inclusion of Silverlight controls will be useful in organizations where there is a wide range of rich media. Most organizations have poor control over existing rich media. Those with rich media are struggling to manage the rich information in a cohesive manner. Once these problems are addressed, the bandwidth demands of pulling videos and images into the system and then pushing result sets to users seems to invite bottlenecks. My experience suggests that theses bottlenecks will require considerable investment in infrastructure to resolve. Organizations want to reduce costs and increase efficiency. I saw some hints that this presentation assumes that the bandwidth and the plumbing are in place or at least not an issue for Fast ESP users.

Finally, the complexity of the interfaces and the system itself concern me. You can criticize Coveo, Exalead, and Google for simplifying SharePoint search. The reality is that these third party systems offer organizations a way to provide meaningful information retrieval without the complexity presented in this video. Echoing in my mind as I write this are two statements from the presenters (Jan Helge Sageflat and Stein Danielsen): ““Now I am copying the wrong one. There we go.” and  “I am just selecting from the 60 or 70 options.”

I want to close with the road map presented in the video:

Notice that for the 2008 period and I think for the release in 2009 “no change in packaging or pricing.” Fast ESP will run on Windows, Linux, and Unix. I think this is interesting because the Fast ESP system consists of original code, licensed components, and acquired code. Frankly I don’t see the Fast ESP system being converted to a Microsoft code base in the near future. So the old Fast ESP is the new Fast ESP.

And for the future, Microsoft will create a Fast ESP for SharePoint. If I understand this point, it means that the demo in the video is a work around. The organization using Fast ESP will have to perform the work needed to make the nuts and bolts fit together without leaks.

My recommendation, based on a second watching of this video, is to turn to a third party solution that “snaps in” to a SharePoint environment. There are Microsoft Certified Partners with products that have not created a financial hassle such as the one that entangled the “old” Fast Search & Transfer, that work out of the box, and that deliver meaningful faceted navigation without cutting, pasting, and juggling multiple coding interfaces.

The addled goose is old and I suppose somewhat jaded. With lots of SharePoint installations, my thought is that if I were 25 years old, one could make a fine living dealing with the intersection of SharePoint and Fast ESP. No more. I will just pick up the phone and call a third party vendor. I would load the software and call it a day. Clean, Easy. Simple. And in the long run a more practical, economically sound approach. Honk.

Stephen Arnold, August 17, 20009

Comments

Copyright © 2008 Beyond Search · ArnoldIT.com · Disclaimer · Revolution News theme by Brian Gardner · Log in