In 9.4 Must ODS LAYOUT be ABSOLUTE or GRIDDED? Which is the Default?

[AnotherTim]

I am updating some ODS code which runs as desired under 9.3 in preparation for our move to 9.4 next month.  The code has stuff like

ODS REGION X=0 PCT Y=0 PCT HEIGHT=50 PCT WIDTH=100 PCT;

I think that means it uses absolute layout.  The 9.4 ODS manual describes ODS LAYOUT ABSOLUTE <options>; and ODS LAYOUT GRIDDED <options>; and does not seem to document my ODS LAYOUT START; statement.

Does this mean 9.4 (TS1M4) is really running 9.3 ODS (because my code runs) and I should use the older references, or should either the ABSOLUTE or GRIDDED keyword have been shown as optional (as the default), or am I just mis-reading the documentation?

I am having some trouble with a TITLE1 getting lost when I code a TITLE2, and it would be nice to know which version of layout I am using as I try to work through it.  You don't want to see the code--over 500 lines of ODS and macro and GPLOT, begun a long time ago and updated by many programmers as ODS evolved.

[ChrisNZ]

I *think* your interpretation about ODS LAYOUT START is correct.

START appeared with version 9.1 and the GRIDDED capability came later.

TITLE1 resets TITLE2, but TITLE2 should not affect TITLE1.

[AnotherTim]

More of an update than a reply.  I suspect that rather like PROC DELETE, ODS LAYOUT START was too valuable to kill when it was dropped from the documentation.  Dropping it from the documentation made it impossible to document the change in behavior from generating a STARTPAGE=NOW to not generating a STARTPAGE=NOW.  One might even suspect the numerous references to STARTPAGE=NOW only being useful after coding STARTPAGE=NO were written by a technical writer unaware that ODS LAYOUT START is still valid.

Chris is correct about TITLE1 and TITLE2.  My current theory is that ODS LAYOUT START or STARTPAGE=NOW initiates a scope for title statements and I lost my TITLE1 when I emerged from that scope.  There is some discussion of ODS RTF having an issue with TITLE statements, but it did not exactly match my experience.  Anyway, I finally have code which generates similar results in 9.3 and 9.4, so I can put it in production before the conversion instead of being tied to the exact day of the conversion.

If you fight a similar beast, the suggestions to use background color styles during debugging are helpful.  Seeing the sizes of my regions change from page to page gave me valuable clues.