VMs: Re: Mickey Mouse in the VMS

[Jorge Stolfi <stolfi@xxxxxxxxxxxxx>]

    > [Stolfi:] Believe it or not, these drawings actually had a
    > serious technical purpose...
    > 
    > [Petr Kazil:] Let me guess. You tried to discover how long it
    > took to make one VMs page and to make an estimate of the time
    > that was necessary for writing the whole manuscript? What were
    > your results?

That was a secondary goal. I didn't actually measure, but as I recall it
took me between 0.5 to 1.0 hour per page. Considering that most VMS
pages are more complicated than my parodies, the entire VMS (~250
pages) must have taken at least 30 days of full-time work (8 hours per
day).

That, of course, is only the bare drawing and writing time.
That is how long it would take for a copist to reproduce the book,
or (as in the Chinese theory) to write it under dictation and then
copy the illustrations.

If the book is an original composition, the time must have been much
longer. The author must have spent far more time researching the
contents of each page, sorting his/her notes, planning the pages's
layout, etc.. Surely the author must have made drafts and preliminary
sketches, even though he/she obviously cared more about contents than
about looks. If those preparatives are included, I would guess at
least six months of full-time work --- probably spread out over
several years.

    > [Gabriel:] Come on Jorge, tell us!
    
Well, since you ask...

Recall that in 2001 we were still hoping that Beinecke would produce a
good quality digital edition of the VMS with their in-house equipment.
So the main purpose of those drawings was to find the best scanner
resolution and compression parameters that would allow us to see all
important details (such as the order and direction of pen strokes),
but still fit in a reasonable number of CDs.

I also intended to use those fake VMS pages as tests for cleanup tasks
that we may have to do on the real scans. One such task was the
re-alignment of the red, green, and blue channels, which in some
scanners are produced by three separate mechanical scans. (As I
recall, you did apply such correction to some Beinecke-provided scans.)

Finally, those drawings were meant to test an idea that I had for
removing the bleedthrough from the other side of the paper. That
explains the margarine. (And I *did* have to explain it to my puzzled
colleagues, who apparently were thinking of some Marlon Brando movie
which I haven't yet managed to see. 8-) The idea sort of worked once,
but then I "improved" the code and it is presently broken. I must get
back to it some day...

All the best,

--stolfi