10 Technology Trends You Can't Afford to Ignore – From Gartner

There have been lots of posts around Fatwire and Google Analytics integration but none with much information. Here are some of the details which might help to provide some insight around integration:

There are four different types of tracking defined with the integration tag:

• Tracking the use of a particular asset; be a PAGE or any other type.  A PAGE asset will trigger a page viewed tracking event
• Tracking when a recommendation has been selected or clicked by the user
• Tracking when a list of recommendations is requested. The object type is set to RecAsked.
• Tracking when a recommendation is listed for display purposes. The object type is set to RecListed.

Based on above tracking system, two modes of operations are defined:

• Basic which tracks physical page views using the GoogleAnalytics (GA)  _trackPageview function and the request to display Assets using the GA Event facility function _trackEvent. This basically covers first tracking type.
  • Examples:
    • <Asset Name> [<Asset Type (c)>=<Asset ID (cid)>]
    • FSII Home page [Page=1223456789]
• Advanced is extension on top of the basic tracking system to record Engage Recommendation information. This mode makes extensive use of the GA Event Tracking facility to record asset related information for recommendations and segments. The three components of a GA Event (category, action, and label fields) are used to store FatWire information. This covers the last three tracking types.
  • Examples:
    • Recommendation Listed
      • <Recommendation name>  [recId = <recommendation Identifier>]
      • FavouriteBlogList [recId = 123456789]
  • Segment information with list of assets
    • FavouriteBlogList [segId = 123424325435], shishankinfo [Blog = 987654321], apoorvinfo [Blog = 1357924680], contentprise [Blog = 2468013579]

AddAnalyticsImgTag is the element that renders tracking request to an Analytics data collection server either FatWire Analytics product and/or a 3rd party product such as Google Analytics.

Example:

   <render:callelement elementname=”Analytics/AddAnalyticsImgTag”>

         <render:argument value=’<%=ics.GetVar(”c”)%>’/>

         <render:argument value=’<%=ics.GetVar(”cid”)%>’/>                                                           

         <render:argument value=’<%=ics.GetVar(”site”)%>’/>

         <render:argument  value=’<%=ics.GetVar(”childpagename”)%>’/>

         <render:argument value=’<%=ics.GetVar(”recid”)%>’/>

</render:callelement>

Reporting on Gathered Data

Once the data has been collected, it’s time to display them in different reports.  There are different options available:

• Directly login to Google Analytics and generate customize reports; this could be displayed within the Fatwire UI accessed through inline frame
• Custom reporting by accessing analytics data through the Google Data Export Protocol. The only caveat with this is that the web page producing the reports be accessible from the Internet.

Limitations of using Google Analytics

Google limits the amount of activity between a client and Google Analytics. Here are some facts:

• Maximum of 5 million page views per month for any Web property (unlimited if the site is a Google AdWords subscriber)
• No more than 500 tracking requests per session.
• No more than 10 requests in any given 1-second period.
• No more than 10,000 requests per 24 hours.
• Data query is limited to 10,000 entries per feed.

• Versioning:
  • No control of number of versions to be maintained at content type level. It’s all or none
  • No control over which content type needs versioning and which not. Its either all or none
  • No versioning at component (Image, Menu, Navigator etc.) level  
• Syndication/ Content Publish:
  • All the syndicators will be either scheduled or none of them. You cannot choose few syndicators to be automated while leaving others as manual process
  • Once the content becomes live/published it is moved across from one to other environment. There is no way to control approval of the content for specific publisher/syndicator as with other products
  • Tough to debug the cause of failure of syndicator
  • If there are some failed items with syndicator, it is tough to determine which content items failed
  • Lack of customization
  • Syndication/Publish history view is not available 
• WCM Portlets
  • Out of the Box WCM authoring and rendering portlets are legacy IBM portlets and are not JSR standards
  • Customization is painful as it uses IBM specific code and it is not exposed 
• Customizations are
  • Tough as they are IBM legacy portlets
  • Limited set of API and taglibs available which prevents customization 
• Rendition Services
  • No rendition service available out of the box 
• Workflows
  • Basic workflows can be defined
  • Parallel workflows I do not think are possible
  • Workflows available only for Authoring templates but not for components
  • Granular level privileges not possible to implement  Example: Controlling function privileges at each stage
  • No funda of voting and group approvals and rejects 
• Caching
  • No Database caching
  • If using JSP components for a functionality through Presentation Templates, there is no Out of the Box caching available 
• Rich Text Editor
  • Out of the Box RTE adds unnecessary HTML tags while creating content. Example: Adding <p> tags
  • Out of the Box RTE doesn’t provide features like copy from word which is quite common these days 
• Context upload and Extract:
  • No Out of the Box facility to import content  to WWCM. We need to write a custom module to upload
  • No Out of the Box facility to export content in readable format. Example: XML 
• Reporting:
  • Custom build is needed for any reporting. Example: Content extract between dates to a specified format, Content changes between two dates, content status between two servers etc. 
• Secured binaries (like images, documents etc. )  are stored in database than file system
• No basic WebServices available for Content Services

Extract from one of the Press release:

“Up to now, domain names had to use the 26 Latin letters in the English alphabet as well as 10 numerals and the hyphen.Technical efforts have enabled display of parts of Internet addresses in other scripts, but the two-letter suffixes had to be made up of those 37 characters.The approval for non-Latin characters applies for now only to domain names connected with the two-letter country codes, like .ru for Russia and .cn for China.Languages that could become available in 2010 for Internet-site names include Arabic, Mandarin Chinese, Cyrillic, Hebrew, Hindi and Korean.he so-called generic top-level-domain suffixes, like .com, .net and .gov, will remain Latin-characters-only for now.”

Tha above announcement could bring in number of changes in current industry. Here are few of them:

• Web Content Managment: With new non-latin domains getting registered everyday, WCM demand will increase than ever before with the focus on multilingual support. In my experience with WCM, I have not come across many implementations that support multilingual. So, its going to be a challenge for both the product vedors as well as System Integrators. It is important to test the current implementations for non-latin languages and becoming a Must Have feature for any implementations going forward. The same will hold good for e-commerce .
• Translators: The market for translators will increase and will become more prominent. The content will start becoming more localized to regions and translators will come to rescue to retrieve any such localized information to outside world. And there is a good chance that translators becoming one of the offerings of WCM products.
• Search: Local/regional seach will start occupying the search space. This might lead to emergence of lot of localized search engines and will provide stiff challenge to some of the bigger players today in the industry
• Social Networking: With Web 2.0 and social networking the key in todays world, I can imagine local facebook, twitter etc. versions emerging in the market. Localized social search is other area to watch out.
• Migration tools: Emergence of tools which could help to migrate your existing site to localized version

The announcement will:

• Bring in new business opportunities and whole new dimension for non-latin countries
• Internet users will increase exponentially
• Internet will start becoming the preferred source of communication for non-latin countries through localized email and social networking sites
• System Integrators local market in non-latin countires will increase many folds
• Add more challenge to the big brands as they will be pressured to register their current domain names in any number of non-Latin-script languages to prevent fake sites
• Another potential issue highlighted is that some characters in non-Latin scripts appear similar to those of Latin-alphabet characters. So, if we substitutea non-Latin character for the similar Latin character, it will createa unique URL — and the potential for site-spoofing

What are your thoughts about it?

• Setup desktop implementation of WWCM and Portal (6.0.1.3), using DB2 and connecting to shared LDAP. Lets refer this setup as “WS6013”
• Make sure that JCR schema is separated as a database instance
• Setup syndication from source (Environment to pick content from) to WS6013
• In parallel, setup desktop implementation of WWCM and Portal (6.1.0.2) on SAME MACHINE, using DB2 and connecting to shared LDAP. Lets refer this setup as “WS6102” 

• Once syndication is completed on WS6013, apply the following iFixes: 
  • PK57970
  • PK60183
  • PK62044
  • PK63553
  • PK64160
  • PK64247
  • PK64833
• On WS6102, apply the following iFixes: 
  • 6.1.0.2-WCM-PK85196
  • Run iFix Post install tasks
    § Run {WAS_PROFILE_ROOT}/ConfigEngine.bat remove-wcm-migration -DPortalAdminPwd=<password> -DWasPassword=<password>
    § Run {WAS_PROFILE_ROOT}/ConfigEngine.bat configure-wcm-migration -DPortalAdminPwd=<password> -DWasPassword=<password>
    § Run {WAS_PROFILE_ROOT}/ConfigEngine.bat update-wcm -DPortalAdminPwd=<password> -DWasPassword==<password>
• Disable JCR text search indexing on WS6102 at location <wp_profile_root>/PortalServer/jcr/lib/com/ibm/icm/icm.properties
• Remove all subscribers & Syndicators, and disable searches on the WS6013
• Run portal-pre-upgrade task to backup WS6013 setup
  o Run from WS6102: WPmigrate.bat portal-pre-upgrade -DbackupDirectory=<Backup Directory> -DcurrentPortalDirectory=<WS6013 Portal Directory> -DcurrentPortalAdminId=<WS6013 Portal Admin Id> -DcurrentPortalAdminPwd=<WS6013 Portal Admin Password> -DbPassword==<WS6102 Database password> -DSkipAiMigration=true
• Take backup of JCR database on WS6013
• Copy WS6013 JCR database to WS6102
  o db2 “restore database JCRDB from <backup-location> taken at <timestamp>”
• Connect WS6102 to new database by running the following commands:
  o Run {WAS_PROFILE_ROOT}/ConfigEngine.bat validate-database-driver -DTransferDomainList=jcr -DPortalAdminPwd=<WS6102 Portal password> -DWasPassword=<WS6102 WAS password>
  o Run {WAS_PROFILE_ROOT}/Run {WAS_PROFILE_ROOT}/ConfigEngine.bat validate-database-connection -DTransferDomainList=jcr -DPortalAdminPwd==<WS6102 Portal password> -DWasPassword=<WS6102 WAS password>
  o Run {WAS_PROFILE_ROOT}/ConfigEngine.bat connect-database-jcr-migration -DPortalAdminPwd==<WS6102 Portal password> -DWasPassword=<WS6102 WAS password>
  o Run {WAS_PROFILE_ROOT}/ConfigEngine.bat create-wcm-persistence-tables -DPortalAdminPwd==<WS6102 Portal password> -DWasPassword=<WS6102 WAS password>
• Run task to update Node Type
  o WPmigrate.bat update-nodetypes -DPreviousPortalVersion=6013 -Dinitial.wcm.nodetype.version=2 -DbackupDirectory=<WS6013 backup directory>
• Run database tasks to re-org information of JCR database
• Run a script on WS6102 to bring it up to 6.1 level
  o Windows: WPmigrate.bat update-wcm-data -DWasUserId=username -DWasPassword=password -DbackupDirectory=directory
• Recreate subscribers and syndicators on the 6.1 server, and rebuild search indexes, enable search

All the best !!

• Retrieve user’s workspace
• Get Authoring Template Id
• Retrieve Site Area Id
• Content Search based on Authoring Template, Site Area
• GetIds of all the content list retrieved from Content Search
• Retrieve Content based on specific requirement from the Content Search result
• Return back the content item

The average amount of time taken is between 0.4 to 0.6 seconds. Out of which % of time taken for retrieving Document Ids for Authoring Template and Site Area forms 50 % of the time to retrieve information. So, if a portlet/page has 6 calls for content, the total time for just WCM is between 2.4 to 3.6 seconds which doesn’t meet SLAs for a page load. This basically lead to defining WWCM Site Structure caching design pattern.

Key Aspects:

• WCM works on the principle of Document Ids. Whether it is a category or a Site Area or Authoring template or content itself, each individual has a unique id within the system
• Utilizing the above concept and initial analysis of the APIs, calls like getById(), findByName() are taking bulk of the response time which are required to extract WCM structure information before actually searching for content keys
• Caching the WCM site structure ids within dyna cache for ever so that retrieving the WCM structure could be fast to search content within them
• Caching can be proactive or reactive

Unknowns:

• There could be potential unknowns that are still to be explored

Futuristic:

• As WCM base works on Document Id, this pattern holds good for future WCM versions till the underlying technology changes

WCM Site Structure Caching Pattern

Advantages:

• Generic WCM structure caching without any parameter specific – User Role, Product Type etc.
• WCM Site structure and design artifact ids doesn’t change
• Content changes will be reflected without any cache flush
• Dyna cache size will be in control
• Drastic % improvement in retrieving site structure (almost 50%)

My thoughts are that this caching strategy will work in any scenario and it is not an overhead. What are your thoughts about it?

1. Installed Wordpress v 2.1 which I was using on pcmspace.com
2. Migrated pcmspace Database making minor changes to the domain/user details
3. Update install path to map domain path
4. Check point 1: Unit test the migration
5. Upgrade Wordpress to V 2.8.2
6. Run migration tasks
7. Check point 2: Unit test the migration again

Its important to come up with the right migration strategy else you can invest lots of time. For example: in migrating my above blog, I started with installing the latest version of Wordpress and tried all sorts of options – just moving content and not database schema, running sqls etc. But in the end, it was too much of work and it doesn’t provide the feeling of doing right.

I have not fully finished the migration, as I need to point my old domain to new. Also changes to my install path means, I need to play around with some of the urls that are not working on the website. I hope I can complete that fast and can be back in blogging  world

This search widget can be configured in your blog’s Template | Page Elements
tab, in the “Add a Page Element” popup.

• “Usability Improvements: These include portal and user administration as well as content management. Portlets may be managed overall or for individual instances including default definition. User administration simplifies user creation, provides a list-based view, and includes user search. Basic content management provided out-of-the-box now includes action-based management within a familiar directory view.

• Advanced Personalization: Now users can personalize individual portlets, including themes, layouts, and portlet content, to increase the productivity of specific roles and people within a business process or collaboration effort. Further enhancements include user created user interfaces, drag-and-drop portlets, personalized dashboards and more granular controlled access at the portlet level.

• Content Management System (CMS) Workflow: JBoss jBPM provides content management approval workflow in a configurable process that enables or disables this role-based approval capability.

• Additional Web Services for Remote Portlets (WSRP) Support: WSRP support offers expanded functionality beyond the basic producer and consumer profiles. Version 2.6 adds implicit cloning capability to both the producer and consumer and supports advanced WSRP profiles.

• Identity: With JBoss 2.6, developers have pre-built LDAP integration with LDAP server. Supported servers include Red Hat Directory Server, OpenDS, and OpenLDAP.

• Google Gadget Integration: Now, developers have a simplified way to drop any Google Gadget as portlets. Google Gadgets are mini-applications that work with the Google homepage, Google Desktop, or any page on the web and can range from simple HTML to complex applications. Examples include a calendar,a weather globe, or a media player.”

 Here is my take on new release: There has been a significant change in the product usability but I still feel it lags in front of its competitor open source product, Liferay. Though the drag and drop feature is implemented but its no where near what Liferay provides. Google Gadget integration has added much life to the portal which lags the number portlets available out-of-the-box. Pre-built LDAP integration is nice and JBoss jBPM is really capable to handle pretty complex workflows. Above all one thing that remains same from many versions is that it works only on JBoss Application server and this totally defeat Liferay’s SOA point of view (Service Oriented Liferay).