Clarification: "Replace Endevor with existing open source tools"
This is required because generally if you want to replace an existing Endevor install you replace it with another COTS tool or with an inhouse bespoke application.
Let's go with the 'replace Endevor with existing open source tools' as see how far we can get.
It may be possible to replace Endevor with existing open source tools. Let's get a few points covered first though. The CA Endevor SCM software runs on the mainframe while open source tools generally run on midrange. This is much like saying that the software is running on Linux and you have Windows. This is a good simile as the mainframe has linux services and users in enterprise environments generally run Windows.
Endevor controls source code and the objects produced by source code generation. So we are talking about controlling source and various outputs which includes binary executable objects that are related to the code. These objects are generally stored on the mainframe file system. Each type of object is stored in an appropriate file container. Each environment stage has a set of these folders plus folders for the base and delta objects.
Endevor functions include retrieve, check out, check in, add, update, delete, transfer, sign out, sign in, package, move. Endevor allows for stages within an environment to be security controlled. Endevor configuration allows for multiple systems to be defined per stage; to handle multiple file types and allow for multiple options per type for processing (named processor groups within Endevor). An audit trail is cut to logs for any action.
The key functions here are the standard source control functions, element generation which includes program compiling, packaging, and release functions. The Endevor system calls mainframe programs to perform element generation - which includes the cobol compiler IGYCRCTL and the CICS translator DFHECP1$ - to produce code that can be executed on the mainframe. In several cases, such as REXX and SAS programs, no modification may be required and the output is just a text file - although in some cases environment variable substitution occurs.
In order to run these programs you still need to store the result of element generation to mainframe libraries.
While your open source solution can store the source and store the binaries and other objects produced the code runs on the mainframe.
There is an assumption here that the end result of this is that the binary executables and other code objects arrive in the destination environments in the same controlled manner for which Endevor provides.
It is possible to access mainframe datasets from midrange.
It is possible to secure mainframe datasets so that only specific system userids can write to them and anyone can read them. See RACF or ACF2.
It is possible to compile cobol and mimic the inheritance of mainframe libraries from midrange.
So, yes, It is possible to replicate Endevor functions using existing source control tools.
In addition, the mainframe has a linux side, omvs, for which can run possibly run the open source tools.
However, there are shortcuts. Serena software offers this functionality via its SCM suite, same for IBM RDz. If you really wanted 'free' software there is always IBM's SCLM (you get what you pay for).
Most of the time when a company who has or needs Endevor does not get it then they either get a competing product or code their own on the mainframe using mainframe tools - ISPF and REXX.
Endevor has both source control (SCM) and release management (DSL) for which covers multiple envrironments (DEV TEST and PROD) so you would need to be able to cater for the scenario where source objects and binary objects are delivered to a target environment.
If you do come up with a coherent answer please post it back here. You never know... someone may try to implement it one day.
