Forum OpenACS Improvement Proposals (TIPs): Tip #103: Drop Oracle requirement for core releases starting with 6.0 (rejected)

The need for Oracle support and the lack of publicly available Oracle testing resources has resulted in delays of nearly every OpenACS release in the last year(s). Additionally it is becoming ever more difficult for core team members to find the time and resources to test on Oracle before the release and make sure that all commits to core are working in Oracle as well.

I hereby propose to drop the Oracle release criteria in the release following 5.4 and name this release 6.0, indicating by the version number that something large has changed.

This TIP does *NOT* propose to drop Oracle or MultiDB support. The infrastructure of OpenACS to handle multiple databases should stay the same and no existing code shall be removed.

Additionally in the commit guidelines we are going to stress that DB specific code is not allowed to be committed in the .tcl or the .xql files and developers are encourage to provide an -oracle.xql file if the -postgres.xql file has changed.

Some additional rationale which is not part of the TIP:

To make it easier for Oracle users we could provide a mailing-list or RSS feed where we publish all changes to acs-core "-postgres.xql" files, so they can figure out if somthing needs to be done.

If the .LRN consortium wants to continue providing Oracle releases than the responsibility is just lifted from the OpenACS community onto the .LRN community, no change will happen for the Oracle Users at large.

Well, we discussed the future of Oracle not long ago, and the OCT was convinced that we would continue oracle support for core for the forseeable future.

I don't see that anything has changed that would justify revoking that decision, myself.

Forseeable future at the time we discussed it was 5.3 and 5.4 release.

What has changed in my opinion is that the support needed and assumed when we talked about continuing Oracle has not manifested itself.

1. UNED is porting LORS to Oracle, check the CVS logs

2. UNED tested .LRN upgrade 2.2.0->2.2.1 last weekend and will be doing so in the future ...

There's not the level of help for oracle I'd like to see but it's not true that no one has stepped forward when the problem was raised.

My vote is NO BTW, just in case there's any doubt.