[gs-cvs] rev 7448 - trunk/gs/doc

[ray at ghostscript.com]

Author: ray
Date: 2006-12-04 14:25:49 -0800 (Mon, 04 Dec 2006)
New Revision: 7448

Modified:
   trunk/gs/doc/pscet_status.txt
Log:
Close some 'setflat' issues as AOK where GS is close, but better at each
resolution. Now that CPSI implements setflat, the differences in Adobe
implementations justifies GS differing (slightly) as well. Cases where
GS is worse (less smooth) than CPSI and Tek still need analysis.

Reassign 09-24-2 group of problems to Ray.


Modified: trunk/gs/doc/pscet_status.txt
===================================================================
--- trunk/gs/doc/pscet_status.txt	2006-12-04 21:41:08 UTC (rev 7447)
+++ trunk/gs/doc/pscet_status.txt	2006-12-04 22:25:49 UTC (rev 7448)
@@ -551,11 +551,19 @@
 
 09-36-7  DIFF	LOOK 14 has minor differences.  Assign: Ray
 
-09-37-1  DIFF	CPSI has setflat always 1.0. GS similar to Tek, but the value
-		at which polygon edges changes from 16 to 8 differs (GS at 53,
-		Tek at 19). Also Tek changes from 'smooth' to 16 at 5, GS at 13.
-		May be resolution dependent (spec is 'output device pixels').
-		assign: Ray
+09-37-1  AOK	GS similar to Tek, but the value at which polygon edges changes from
+		16 to 8 differs (GS at 53, Tek at 19). Also Tek changes from 'smooth'
+		to 16 at 5, GS at 13. CPSI fixed to implement setflat changes from
+		smooth to 16 sides at 9 (GS at 13), CPSI never has less than 16 sides.
+		Resolution dependent (spec is 'output device pixels') and PLRM 3rd ed
+		states "If the flatness parameter is large enough to cause visible
+		straight line segments to appear, the result is unpredictable."
+		When GS is run at 300 dpi (same resolution as Tek), the transition
+		from smooth to 16 sides is at 7 (vs. 5 on the Tek) and from 16 to 8
+		sides at 27 (vs. 19 on the Tek).
+		-
+		Since Tek and CPSI differ, and GS generally uses more line segments,
+		the GS output conforms and is superior to the Tek output.
 
 09-37-10  OK	GS matches Tek and CPSI output now matches Tek (12/04/06).
 		Imagemask turkey fixed by r7123.
@@ -566,11 +574,15 @@
 
 09-37-3  OK	Matches Tek hardcopy output. RJJ
 
-09-37-4  DIFF	CPSI has setflat always 1.0. GS similar to Tek, but the value
-		at which polygon edges changes to very coarse differs (GS at 73,
-		Tek at 23). Also Tek changes from 'smooth' to bumpy at 5, GS at 19.
-		May be resolution dependent (spec is 'output device pixels').
-		assign: Ray
+09-37-4  AOK	GS similar to Tek, but the value at which polygon edges changes to
+		very coarse differs (GS at 73, Tek at 23). Also Tek changes from 'smooth'
+		to bumpy at 5, GS at 19. When GS is run at 300 dpi, the transitions
+		from smooth to bumpy is at 11, and from bumpy to very coarse at 39.
+		Resolution dependent (spec is 'output device pixels'). CPSI fixed to
+		implement setflat transitions to bumpy at 13 and to coarse at 33.
+		-
+		Since Tek and CPSI differ, and GS generally uses more line segments,
+		the GS output conforms and is superior to the Tek output.
 
 09-37-5  OK	Matches Tek hardcopy output. RJJ
 
@@ -579,18 +591,18 @@
 09-37-7  OK	Very small fill differences. RJJ (CPSI ignores setflat, GS
 		similar to Tek hardcopy)
 
-09-37-8  DIFF	CPSI has setflat always 1.0. GS similar to Tek, but GS differs
-		in the exact polygon shapes.
-		May be resolution dependent (spec is 'output device pixels').
+09-37-8  DIFF	GS similar to Tek, but GS differs in the exact polygon shapes.
+		Resolution dependent (spec is 'output device pixels'). GS at 300
+		dpi has less edges on large outline in top two images. CPSI (at
+		600 dpi) also smoother than GS on top two outlines.
 		assign: Ray
 
-09-37-9  DIFF	CPSI has setflat always 1.0. Tek hardcopy shows setflat does
-		NOT apply to line cap/join (GS has polygonal round cap/join).
-		"Stroked_Objects" and "Filled_Objects" differ slightly with
+09-37-9  DIFF	"Stroked_Objects" and "Filled_Objects" differ slightly with
 		number of polygonal edges. Tek shows strokepath of "Technical
-		Paths" IS subject to setflat -- GS has smooth round cap/join.
-		May be resolution dependent (spec is 'output device pixels').
-		assign: Ray
+		Paths" IS subject to setflat -- GS has smoother round cap/join.
+		Resolution dependent (spec is 'output device pixels'). CPSI fixed
+		to implement setflat matches Tek ouptut for "Technical paths".
+		Also has same issue as 09-24-2.  assign: Ray
 
 09-38-1  OK	Minor differences visually reviewed by RJJ