2. Why Version Control?
• Siebel developers continuously make changes
to Siebel objects
• The smallest changes often introduce new
defects
• The quickest way of finding these defects is by
inspecting the recent changes
• These problems get compounded with
multiple developers that are geographical
spread out teams
3. Cost of fixing defects
The cost of fixing defects increases exponentially with every delay
4. Why Version Control? …
• When employees, consultants and system
integrators leave so does the knowledge of
the changes made by them
• Developers forget the changes made by them
as time passes
• Without proper version control business runs
the risk running into catastrophic production
defects that can not be fixed in time or not at
all
5. Siebel Version Control
State of the mart
• Siebel repository offers Check-in/Check-out but does
not keep the version history of the objects modified
– Siebel check-ins OVERWRITE the previous version of the object
– There is no way to find out a objects state in a previous version
– The developer only has to debug based on the current state
• Support for file based version control systems
– SVN, CVS, Perforce, ClearCase, SourceSafe
7. SVN/CVS challenges
• Is built for Versioning files not objects that Siebel is
made off
• Does not integrate seamlessly with Siebel
– srcctrl.bat files have to be placed in Siebel tools (Client side
integration as opposed to Siebel repository)
• Check-ins are not atomic
– Failure to check-in in SVN does not prevent check-in to Siebel
repository and vice a versa
• Hard to query
- Files are checked based on project vs object check-in
• Hard to Maintain
– SVN usernames and password have to be administered and
8. Siebel Tools UI is suited for Siebel's objects
-easy to query and navigate Siebel Objects
-comfort zone of the Siebel developer
9. In SVN/CVS developers have to comb through XML tags and diff XML files
No easy to navigate or query like Siebel tools
Hard to find right sif (object vs project changes)
10. As a result SVN/CVS is rarely implemented
and even
rarely adopted
vs.
12. Object Hive
State of the Art
• Uses a db repository and structure similar to
Siebel Repository instead of a file based
repository makes it easy to query
• Server based solution
– Changes are captured directly from Siebel server repository
– Versioned objects are created in Object Hive repository
– No deployment necessary on individual developers tools
– Query and compare objects using UI similar to Siebel Diff
– Powerful analytics
– Does not require additional user admin, users are authenticated
against Siebel db
• UI similar to Siebel Tools
16. Powerful Repository Analytics provides insight into Siebel development
A feature not possible with Siebel tools or SVN/CVS
Sample report for user GKING shows all Integration Objects customized by the user
And how many times each object was modified by the user
and new features!
18. User Experience Hard to
use and
query
Rich UI, and
easy to query
Integration
with Siebel
repository
None Seamless
Reporting None Bar/Pie/Time
line Charts
Authentication None Same as
Siebel
Comparing
objects
XML diff GUI Tree Diff
SVN/CVS Object Hive
Comparison of SVN/CVS and Object Hive
19. ROI
• Developers can quickly pin point changes that
introduce bugs. Save hours and days
– Tools pays for itself in reduced debugging time
• Lead developers can track changes introduced by
developers. Pre-empt bugs.
• Managers powerful analytics, track the development
accurately. Compliance.
Admin keep track of changes reduce migration
headaches and manage parallel development
reduces risk of untraceable changes
20. System Requirements
• SQL Server – Object Hive db needs to be on
the same database server as the Siebel server
• Oracle – Object Hive needs a db link to the
repository db.
• Runs on any platform (Windows or Linux) and
most browsers (IE, Firefox, Chrome and Safari)