[DVIPDFMx] extractbb man page

[Karl Berry]

> It may be helpful to read the comment in pdfximage.c around line 800

Very helpful.  How about if we insert this info in the man page
for extractbb?  See diff (and new file) attached.

Thanks,
Karl

--- extractbb.1	(revision 22321)
+++ extractbb.1	(working copy)
@@ -1,4 +1,4 @@
-.TH "EXTRACTBB" "1" "28 April 2011" "20110311" "DVIPDFMx"
+.TH "EXTRACTBB" "1" "6 May 2011" "20110311" "DVIPDFMx"
 .PP 
 .SH "NAME" 
 extractbb, ebb \- extract bounding box information from graphics files
@@ -60,9 +60,21 @@
 .TP
 .B -x
 Write output in dvipdfmx's ``xbb'' format (default), which is the same
-as that used by the PDF format.
+as that used by pdfTeX.
 .PD
 .PP
+.SH "OUTPUT FORMATS"
+.PP
+Here are more details about the bb and xbb formats:
+.PP
+The original ebb (from dvipdfm) ignored density information in bitmap
+images, and generated bounding boxes with 100px = 72bp = 1in.
+Unfortunately, screenshots (especially) tend to look bad with this
+approach.
+.PP
+So, extractbb (from dvipdfmx) uses density information if present.
+Otherwise, it generates bounding box with 100px = 100bp.  This is what
+pdfTeX does.
 .SH "BUGS"
 .PP
 There is currently no way to specify which bounding box is taken

-------------- next part --------------
A non-text attachment was scrubbed...
Name: extractbb.1
Type: application/octet-stream
Size: 2356 bytes
Desc: not available
URL: <http://project.ktug.or.kr/pipermail/dvipdfmx/attachments/20110506/7b777d67/attachment.obj>