Rays of Life…

Reviewing Liferay for a project this week.   It’s powerful stuff.  The integration with Web 2.0 is tight.  Solid Twitter, Facebook and Tumblr support.

To my surprise, it contains Lucene for searching Web 2.0 items as well.   It also does some simple content management.  It’s not very robust (as say Alfresco).   It is still a very impressive tool for collaboration.

It seems the JBoss bundled version doesn’t integrate with source builds very well.  There are no targets for deploying to JBoss in the Ant file.  You must use Tomcat to deploy, after building locally.

Searching for the limits….

I spoke with a product company in DC this week.  They are leveraging SOLR and using custom algorithms to allow clients to intuitively drill into client datastores for SPECIFIC content.  It’s a beautiful thing.

A long time user of FAST search product, it’s great to see someone taking the search game to the next level.  Applying learning techniques to the analyzers and truly learning what it is that you want to see is a largely untapped portion of the ECM/Search/WCM solution.

Documentum who?

I drank the Kool-Aid.  I wrote and implemented Documentum modules for the last four years.  I won’t bore you with details.  I’ll just say it’s a welcome thing for many of us to have a robust solution enter the market that can easily work in its capacity.  It also supports CMIS.  So you’re content solution can be replaced with no affect on dependent systems.  Very nice when you consider EMC’s licensing fees.

Documentum had it’s day.  In post-9/11 projects, something was needed to integrate many government agencies and make their content available.  Thing shave changed since 2001.

Today you have a few solutions that can handle your information on various levels.  You’re also not tied to a vendor search solution to drill into the data either.  Lucene and SOLR are solving that problem very nicely.

Let’s see where this goes in 2011.

Search, flagging content and the metadata dilemma

I spoke with Amazon.com a few weeks ago.  They have an internal system which needs to flag content based on pure keywords.  It’s vital for security.  I did this for the patent trademark office in DC.  They want to flag national security based keywords and move that content for further review, based on full text searching.

The issue Amazon (and many other vendors I’ve talked to this month) is testing.  How do you place content in the system and validly run automated tests to make sure your indexing picks up content as predicted?

Moving from Alfresco to Liferay for WCM….

I’ve implemented Alfresco Share for 2-3 clients in the last couple years.  Most of the systems were search and retrieval based.  Alfresco works nicely for this.  The Share interface competes directly with Microsoft SharePoint.  It has a great interface.  Very “Facebook like”.

The real power in using Alfresco is that any comments, ratings and so on are all searchable and able to have terms and meta data applied to them.  If you have users that wish to mine into content and find not just formal documents based on keywords but also on comments and social network ratings this is available out of the box.

This week I’m researching Liferay’s WCM capabilities.  We’ll see how this measures up.

Portals, WCM and ingestion. Oh my….

I’m currently migrating an existing project from BEA’s product line to an open source solution.   The challenge is moving the existing crawled information from a custom coded process to a robust open source solution.

We decided on the Alfresco CM system.   Liferay is being brought is as a portal.  The issue was finding a migration product that will bring in crawled web content into Alfresco.  Some products consider for that were Talend and OpenMigrate.

If you have any experiences you’d like to share about, feel free.