This post is more a memo for my self than a tutorial.
Download Solr (http://lucene.apache.org/solr/) to c:/km/solr/
Dowload Jetty (http://www.eclipse.org/jetty/) to c:/km/jetty/
Download Curl (http://curl.haxx.se/dlwiz/?type=bin&os=Win32&flav=-&ver=2000%2FXP) to c:/km/curl
Follow this tutorial to run Solr : http://lucene.apache.org/solr/4_3_0/tutorial.html
Starting Solr : ~/solr/example$ java -jar start.jar
Index All XML documents from the exampledocs folder : ~/solr/example/exampledocs$ java -jar post.jar *.xml
Assume you have a doc file called doc2.doc in the folder, index it using : C:\km\solr\example\exampledocs> c:\km\curl\curl.exe "http://localhost:8983/solr/update/extract?literal.id=doc2&commit=true" -F "firstname.lastname@example.org"
Don’t forget to change the literal.id or it will replace the last document.
To verify that it has been properly indexed go to : http://localhost:8983/solr/collection1/browse
And type in something contained inside the document.
You can do the same with pdf.
To add the filename to the indexed file open : C:\km\solr\example\solr\collection1\conf\schema.xml
and add :
Then index a file using : C:\km\solr\example\exampledocs> c:\km\curl\curl.exe "http://localhost:8983/solr/update/extract?literal.id=doc2&literal.fullfilename=C:\km\solr\example\docexamples\doc2.doc&commit=true" -F "email@example.com"
Even simpler :
Index a file using : C:\km\solr\example\exampledocs> c:\km\curl\curl.exe "http://localhost:8983/solr/update/extract?literal.id=doc2&literal.resourcename=C:\km\solr\example\docexamples\doc2.doc&commit=true" -F "firstname.lastname@example.org". It will appear directly in velocity !
The objectives of KM man in an organization are to promote knowledge creation, knowledge growth, knowledge communication and knowledge preservation in the organization (Steels, 1993). It entails managing knowledge resources in order to facilitate access and reuse of knowledge (O’Leary, 1998)
How to efficiently introduce KM ?
KM starts by Managing the Change
Focus on the people, their habits, their way of working
Accompanying the tool introduction with training
Spread best practices
don’t focus on the tool but maximize the way people use the tool (your tool can be great but if no one use it…)
A good KM strategy requires
support from the hierarchy
money and time
resource (a knowledge manager)
continuous evaluation of the tools and the practices
Make the difference between DMS (document management system) & Information Portal
Don’t neglect social media
Extend the DE/PV system used in one site to all test labs. Possible Improvements: additional tags, link to the file, notifications send to the people when a new file is added in their area of interest.
Create a spider (or wikimedia or quickR) for persistent information written in an encyclopedically way or How-Tos, processes. Permalink to files may be included in the pages.
Store persistent knowledge (Test reports, former projects) in a plain text searchable system (DMS). The system should allow tagging. This will replace the Windows folder architecture and the duplication of files.
Use the shared Windows folder growing knowledge (current project). Even better use a file sharing platform that allows versioning (to track changes) and/or sametime editing of files. Prefer emailing the file link than the file itself.
plain text search means that the search engine can search inside document (not only in the filename!)
Tags allow to create dynamical virtual folder called bins without duplicating files.
Some details on the criteria :
Simplicity of retrieving knowledge : Getting access to the information, to the files
Simplicity of adding knowledge : Adding files to the knowledge
Search function: Searching for a specific subject
Tagging function: Allow the use of meta data to organize projects based on tags
Versionning (history): Keep track of all previous versions
Security (confidentiality): Are the data safe ?
Right controls: Manage Admin User, block users, create group of user
Ease of getting support: Community support, paid consultant support
Compatibility: In case of change of system, is it easy to switch ?
Mobility and multi-site: Access to the information from anywhere
Installation: Easy to install ?
Group solution: Does it fit into Company Group policies ?
Interactions (notifications,…): Neswletters, email notification when new knowledge arrives.
Price (and price of support): Price of acquisition, annual fee, price of support
A point of view on the excel sheet subject
Convert xls to matlab :
Matlab can do more things
Matlab can protect the software against editing
Matlab is a lot more complicated for doing simple things
Matlab is more expensive
Excel let the people make the software evolve
If you want to empower the people, let them use and customize their excel sheets. They might have great ideas. But collect these improvements and needs to validate them and distribute to all.
If you want to go into a large software that can do a lot of things in a controlled way hire a matlab consultant. If you do not add all function into the matlab soft, the people will continue to use excel sheets…
Alternative : distribute an official version of the sheet where everyone should verify its data produced by its own sheet.
The task of the knowledge manager
Collect permanently the needs and habits of workers
Identify unreachable knowledge (brain located knowledge or C:\ located knowledge or printed knowledge) and make them accessible
Facilitate the adoption of new tools and best practices
Keep the knowledge bank always up-to-date
Benchmark new solutions
Manage access and rights
Measure and report the ROI of the Knowledge Management strategy