OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help

cgmopen-members message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [Elist Home]

Subject: Help Needed

Hello CGM Open Members --

I need some contributions to help finish the WebCGM test suite.

I have finished work (for now) on one major part of the WebCGM suite -- the static-graphics parts.  Unfortunately, this has taken *much* more effort than I had planned (and I lost significant time in April for personal reasons).   Accordingly, I must put this part aside for now and finish an acceptable "BE subset" of the dynamic stuff. 

Per the project plan, I started with the ATA grex (2.4) suite, threw out a bunch of tests (e.g. ASFs, parabolic arcs, etc), modified all tests for the global differences between ATA & WebCGM (e.g., the legend block), fixed broken stuff, and modified for differences between ATA & WebCGM (a real pain, esp little differences like ATA allows 'text' if monospaced fonts are used, subtle differences in the color/monochrome rules, etc).

I have produced PNG reference images for the current 232 static-graphics tests in the WebCGM suite, using a batch run with the only tool available to me.  I have not figured out how to properly control the size and resolution of the rasterized picture.  As a result, especially the text rendering is pretty low-quality, on all of the reference pictures.  It might technically pass the "max{1%, 1/2 pixel}" criterion, but it should look better for the reference images of the suite's public release.

You will find what I have done at:


Unzip that file, then unzip the two resulting ZIP files (w/ the 'use folder names' option checked).  You will see all of the CGMs.  I believe that these are all now correct WebCGM test cases, and they all pass "metachek -r webcgm".

You will also see the file, webCGMframe.html.  Open this in IE5.5 (NN won't work).  It is the Javascript-based NIST harness, modified for the WebCGM suite.  You can use it to navigate and view the reference images for the test cases.

Question.  So who can contribute correct, better-quality reference PNGs?  Please let me know, either by public reply or privately.

FYI, I think all of the reference images are more-or-less correct except the following list (but this assertion could be wrong).  I said "more-or-less", because there are some marginal renderings.  For example, dotted lines are made with a short dash, that to my eye is too long (look at the construction lines on any of the circular/elliptical arc tests, for example).  The list:



(BTW, I can get correct reference images by a painful process of view [hsiview or metaweb], alt-shift-PrtScreen, paste into Adobe ImageReady, crop, SaveAs png -- but this is very time consuming compared to what several of you ought to be able to do with rasterizers.)



(I haven't verified yet whether the pattern orientations are correct in these picture -- maybe, maybe not.)


p.s.  In the Javascript harness, you will notice that new tests and some others have "TBD" for the Test Purpose and the Operator Script.  I'll sort this out later.

Lofton Henderson
1919 Fourteenth St., #604
Boulder, CO   80302

Phone:  303-449-8728
Email:  lofton@rockynet.com

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [Elist Home]

Powered by eList eXpress LLC