FreeHEP%202000

1
FreeHEP 2000

• CHEP February 2000
• Tony Johnson SLAC
• Saul Youssef - SCRI

2
Contents

• What is FreeHEP (2000)
• Goals
• Implementation
• Status

3
What is FreeHEP?

• FreeHEP is a database of HEP software
• Conceived in 1991 at SSC HEPLib workshop
• Desire to break free from CERNLib
• Implemented in 1992 using
• Newly invented WWW
• SPIRES database at SLAC
• Anonymous FTP server at FSU
• One of the first databases interfaced to the web
• The first to allow incremental searching

4
FreeHEP circa 1993

• FreeHEP grew fast
• Contained almost 200 programs
• About 1/3 programs are MC generators.
• System of Editors for each subject area
• Good for getting things started
• Editors actively solicited entries in their
  subject area

5
FreeHEP circa 1998

• First WWW database becomes
• Most obsolete WWW database
• Moved from VM to Unix but maintained original
  architecture (CGIRexxSPIRES)
• Entries have not remained up-to-date
• System of Editors proved to be too inertial

6
Why FreeHEP 2000?

• New era of small reusable open-source software
  projects makes motivation for FreeHEP stronger
  than ever.
• Need to be able to find out what others have done
• New strategy to make database self-service
• Distributed management of database entries
• Anyone in HEP community can create and maintain
  entries
• Update the database to use latest web
  technologies
• Interest areas for collaboration/discussion

7
Goals

• New entries can be created using web form
• Each entry has one or more maintainers
• Maintainer can update information via web form
• Each entry must be updated or checked each year
• Automatic e-mail notification
• Automatic checking of URL links
• Stale entries marked as inactive
• Import/Export to other databases
• Allow mass import from other databases
• e.g. Fermitools imported entries marked
  read-only
• Allow mirroring of entire database

8
Implementation

• Implemented as a Web Application
• Use Model-View-Controller architecture
• Pure Java, portable to many different
  platforms/architectures
• Object database for data storage
• XML for import/export

9
Model-View-Controller Architecture
Database
View Templates
JSP
FreeHEP Servlet
Browser
Response
JSP
JSP
10
Database Architecture

• Wanted to use SQL database (Access)
• Planned to use JDBC 2 to access/update db
• With Updateable RowSets
• Allows db rows to be updated/added/deleted
  without needing to generate explicit SQL.
• Reference implementations
• Dont exist in JDK 1.2
• Broken in JDK 1.3
• Tried several commercial implementation
• All had bugs which made them hard to use

11
Simple Object Database

• Eventually abandoned JDBC
• Simple homegrown object database
• Uses Java serialization soft pointers
• Optimized for Web Use
• Many readers multi threaded
• Single write lock
• Isolated transaction commit/rollback

12
Database Layout
13
Program Database
14
XML Import/Export

• Simple format for input/output
• Can be set up for automatic (nightly)
  import/export
• Imported records marked as read-only
• Ideally would use common format with other
  commercial/shareware databases

15
XML Example
16
Java Interest Area

• First interest area is set up
• Web page with links to HEP Java Projects
• Mailing list
• CVS repository
• Anyone can get write access
• Sandbox where anyone can put code and get
  feedback
• Hopefully will help to combine projects to form
  basic HEP-wide Java utility libraries

17
Status

• Loaded new database with contents of old FreeHEP
  database
• Testing editing features now
• Debugging authorization scheme
• Will send e-mail to all maintainers of original
  programs
• Should be ready for new entries by mid-March
• First interest area ready now
• http//java.freehep.org

18
Conclusion

• Remember this address
• www.freehep.org
• Check back in April