Re: Customization of SQL*Forms.

From: Jeff Reynolds <reynolds_at_sun44.synercom.hounix.org>
Date: Fri, 28 Aug 1992 02:02:00 GMT
Message-ID: <1992Aug28.020200.6019_at_sun44.synercom.hounix.org>


I'm in no position to confirm the rumour about the .INP files, but last week some Oracle reps came out to our organization to give a technical overview of Oracle7, and a glimpse of future product directions. Part of the discussion centered around SQL*Forms 4.0, and according to the reps, Oracle is "considering" the elimination of the .INP files in 4.0. Apparently the matter has not been finalized yet, and it was suggested that our concerns be aired in order to help Oracle make their "decision".

  Issue #1 - Change control - Tools like RCS and SCCS (UNIX) or CMS (VMS) are

             essential in a multi-person development environment to prevent
             simultaneous (accidental) modifications.  Developers "check out"
             the .INP file to ensure exclusive access to the form definition
             during the development cycle.
  Issue #2 - Version control - Every once in a while, Murphy strikes and 
             something goes drastically wrong.  It becomes necessary to go 
             back to a previous generation of a form and reengineer 
             the changes.  With SCCS or CMS, these earlier generations of 
             .INP files can be reconstructed and reserved.  
  Issue #3 - It is also valuable to be able to determine what changed between 
             generations (and not necessarily adjacent generations).  To do 
             this requires "differenceable" code (aka, ordered text files).
             (Note that .REX files from SQL*ReportWriter are hardly
             "differenceable".)
  Issue #4 - Many development shops are willing to circumvent "supported"
             actions if the result is going to be cost effective (faster),
             as long as safety/reliability is not significantly impacted.
             The direct editing of .INP files, although not supported by
             Oracle, is relatively safe (especially with a code management
             system where the original version can be refetched) and is
             a proven time-saver (global edits, search/replace operations).
             The cost to apply the same change (such as boilerplate text) to
             many forms can be enormous compared to using a stream editor like 
             SED on .INP files.

IMNSHO, until Oracle can provide change control, version control, the ability to identify differences between different forms or generations of the same form, AND global search/edit capabilities, there will be a STRONG DEMAND for .INP files (or their equivalent) by most development shops doing serious development work.

                                        ...jeff reynolds

Disclaimer: Blah blah my opinions blah blah not necessarily my company's blah blah blah. Received on Fri Aug 28 1992 - 04:02:00 CEST

Original text of this message