Hermitian and rotations

NBC’s latest ridicule is all based on “trust me it’s trivial”.  Well re-writing the PDFReferenceXX.pdf specifications is not trivial dude.  You see — you don’t get to just make up new rules when your Xerox guy doesn’t deliver the goods.

This is just plain weird, I have shown how the PDF created by the Xerox workflow saved by preview matches all that is found in the WH LFBC PDF and somehow Hermitian believes, or insists that the Xerox workflow does not deliver the goods.

I’ve told you time and time again that all PDF rotations are between two coordinate systems — not between one coordinate system and an object.  Thus if your six-vectors don’t include a rotation matrix then you don’t have any rotations.

That is correct. However, the cm matrices are not the only way to rotate. The page object contains a /Rotate 270 command. A to PDF rotations being between coordinate systems, it basically takes a set of points that make an object in one coordinate system and applies a set of transformations which rotate, scale, skew and translate these points, forming a new object

So let me make this perfectly clear.  Your Xerox guy doesn’t deliver the goods.  Xerox doesn’t have the right stuff !  All that’s left for you to do is to release the Xerox and Preview PDF files from your best Xerox/Prevew combo special and then we can prove that your dumb machines could not have possibly produced the WH LFCOLB PDF image.

ROTFL.. So I have shown that it delivers the goods but now our friend insists that I should deliver the PDF files so that he can support his claims. Hilarious… He should instead show us how I was wrong, not claim that I am wrong, if he only had the evidence. Why is it so hard for Hermitian to repeat my experiment of printing a document to PDF using Preview…

You are confusing the rotations that your human operator of the XEROX machine applies to a piece of paper with the software rotations applied to the PDF user coordinate system within a computer grahpics program which are written to the lines of PDF code.

Still not understanding what I am telling you based on the data I have presented…

You don’t seem to understand that the PDF rotations between the two coordinate frames must be applied within a software program that can calculate the correct rotation matrix and write the matrix to the PDF file.  Your human Xerox operator cannot

That is of course self evident. My ‘xerox operator’ merely applied two rotations to correct for the upside down image. But even on a right side up image, the document still contains rotations because, as I have shown, the images are stored internally in a landscape format. This is why you see rotate operations in the WH PDF, this is why you see rotate operations in the Xerox document printed with Preview, that’s why you see rotate operations in the original Xerox PDF. Of course, in the latter case, the rotation is applied at the page level.

Your Xerox operator cannot do that because he/she is not applying software rotations within a computer graphics software program.

Any physical rotations applied by the Xerox operator to the sheet of the paper original before the scan to PDF is initiated can only control the orientation of the scanned image.  However, because the Xerox scanning software does not apply any PDF rotations between coordinate systems, then these physical rotations applied by the Xerox operator to the paper original must show up in the orientation of the scanned image.

So if you scan the paper original right-side-up then the scanned image must be right-side-up when the PDF is opened in any PDF Viewer or editor.  And if the operator places the original up-side-down on the glass, then the scanned image must be up-side-down when the PDF file is opened in any PDF Viewer or editor.  This is basic physics stuff.  By the way did you ever take college level physics?  How about college level drafting, CAD/CAM or Descriptive Geometry?

You are now confusing various issues. 1) How the document was scanned 2) how the images are stored internally 3) how the images are transformed into place 4) how the final PDF is oriented.

Let’s walk through this in small steps:

I scan a document on a Xerox WorkCenter

The document is scanned in landscape orientation because the long edge of the paper is perpendicular to the scanning direction which is from left to right. The images are stored internally in a landscape direction (this is why you see them rotated in place in the WH PDF). In the Xerox PDF, instead of rotating them individually into portrait orientation, the software places them ‘as is’  (landscape) into a landscape oriented canvas and then rotates the whole page before it renders it. This is why Illustrator shows them still in landscape as it fails to obey Rotate 270 instruction (which is a know issue).

When you print the document to PDF using preview, the operations are as follows.

1. The images remain stored in their landscape internal position 2. The cm matrices now add rotation to rotate th images from the landscape into a portrait oriented canvas 3. No more page wide rotations are found at the /Page level.

If you scan the document upside down, the same happens, only now the page opens up upside down in readers. When the ‘human operator’ who had just scanned the document noticed this, she/he rotated the picture using Preview and printed it to PDF.

This results in a correctly oriented PDF and the images are still stored in landscape form however rotated 180 degrees (so instead of clockwise, they are now stored counter clockwise).

It’s trivial really.

And you still haven’t told us how Preview can change the form of the Xerox six-vectors ???
To demonstrate that you must multiply any rotation matrix applied by Preview and the concatenated matrix cm from the Xerox scan to PDF and get the final Preview six-vector.  Of course any of the readers could do that check as soon as you release your final Xerox and Preview PDF files.

Preview simply does what it does. Instead of apply the Rotate and page level, as the Xerox PDF instructs the reader to do, it adds the rotation to the individual objects and rotates them into a portrait oriented canvas directly. This is why the preview PDF shows a similar cm matrix when compared to the WH PDF. I have released all the relevant data, such as the cm matrices and the object orientations as stored within the Xerox PDF created by Preview. If you doubt me, then repeat my experiment and show me wrong.

So the bottom line is your fleet of dumb machines don’t have the right stuff.


So you keep saying and yet my data show otherwise. How embarrassing isn’t it?… Now, if you disagree, do the simple experiment and print your PDF using preview and analyse, just like I have done, the rotations.

Looking forward to your detailed “rebuttal”…