Michael wrote: "I did not do much comparisons between jump versions but identified several bottlenecks in zoombar code and just committed an optimized version : 1 - I prevent the renderer to display invisible layers as wireframes during mouse dragging 2 - I used Larry's coordinate's decimator to render features as gray wireframes (I made the decimator's resolution a Java2DConverter property to use a special 2 pixel wide resolution for the wireframe display) 3 - Instead of displaying 100 random geometries as wireframe, I display 200 geometries choosen for their bigger size and so I get a better feedback (the way geometries were choosen, with a 10000 points dataset and another 100 polygons dataset, I get only one or two polygons on the screen during mouse dragging)."
This sounds like awesome work Michael! Michael wrote: "May be it would have been cleaner to create a special renderer for the gray wireframe renderering..." This is an interesting idea. I had set up my pluggable renderer to select a custom renderer based on the type of object being rendered, not on the "mode" that OpenJUMP was in. I'm not sure how I would select a custom renderer based on the "mode" of OpenJUMP, but I'll think some more about it. (By "mode" I mean something like "OpenJUMP is in mouse wheel or scale bar zoom mode.") Michael wrote: "Hope those optimizations will be useful for a mousewheel zoom implementation." I think you are correct about this. I was hoping to enable mouse wheel zoom with my work on the cursor tool framework, so this will be of interest to me. I think Larry Becker has already done some work like this. I wonder if he implemented the types of optimizations we are discussing. The Sunburned Surveyor On 5/27/07, Michaël Michaud <[EMAIL PROTECTED]> wrote: > Hi, > > I did not do much comparisons between jump versions but identified > several bottlenecks in zoombar code and just committed an optimized > version : > 1 - I prevent the renderer to display invisible layers as wireframes > during mouse dragging > 2 - I used Larry's coordinate's decimator to render features as gray > wireframes (I made the decimator's resolution a Java2DConverter property > to use a special 2 pixel wide resolution for the wireframe display) > 3 - Instead of displaying 100 random geometries as wireframe, I display > 200 geometries choosen for their bigger size and so I get a better > feedback (the way geometries were choosen, with a 10000 points dataset > and another 100 polygons dataset, I get only one or two polygons on the > screen during mouse dragging). > > May be it would have been cleaner to create a special renderer for the > gray wireframe renderering... > > Hope those optimizations will be useful for a mousewheel zoom > implementation. > > Michaël > > Larry Becker a écrit : > > > Hi Michaël, > > > > Here is Stefan's post about the problem. The polygons must be very > > large to prevent the simplification logic of zoombar from working. > > > > Stefan: > > >I loaded a large shp file with 16 very large polygons (> 5000 points) > > >and zoom to full extent. > > >When i moved the slider of the zoom bar (zooming out) the systems does > > >nothing (is blocked) for more than 30 seconds (or even more). > > >If make the same thing with an older version of jump it takes one 2 sec > > >after seeing the outlines and one sec more for filling. > > > > Hei Larry, > > > > Larry Becker schrieb: > >> Hi Stefan, > >> > >> 1. Is there somewhere I can get a copy of the shape file to test with? > > > > i upload it here: > > ftp://ftp.geo.unizh.ch/pub/sstein/openjump/brdlaender.zip > > <ftp://ftp.geo.unizh.ch/pub/sstein/openjump/brdlaender.zip> > > > >> 2. Is the speed up working for other large shape files? > > i have not tested > >> 3. Does it perform better if you zoom to full extents instead of using > >> the Zoom bar. > > yes - (or as usual) > >> 4. What is the Committed Memory showing after you load the file? > > mhm.. not that much: 11MB > >> 5. After the blocked behavior, does the Committed Memory go down? > > 14 MB > > > > btw: panning is fine, and if i use Zoom to scale it is fine as well. > > > > thanx for taking care > > stefan :) > > > > > > regards, > > Larry > > > > > ------------------------------------------------------------------------- > This SF.net email is sponsored by DB2 Express > Download DB2 Express C - the FREE version of DB2 express and take > control of your XML. No limits. Just data. Click to get it now. > http://sourceforge.net/powerbar/db2/ > _______________________________________________ > Jump-pilot-devel mailing list > Jump-pilot-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/jump-pilot-devel > ------------------------------------------------------------------------- This SF.net email is sponsored by DB2 Express Download DB2 Express C - the FREE version of DB2 express and take control of your XML. No limits. Just data. Click to get it now. http://sourceforge.net/powerbar/db2/ _______________________________________________ Jump-pilot-devel mailing list Jump-pilot-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/jump-pilot-devel
