In article , RJH
wrote:
The main problem for future historians and academics is how to *know*
the results are always perfectly accurate when they may have no access
to an original or a plain scan. If they need to refer to a more
'reliable' version then they may as well use that!


Yes, I see what you mean now. Can't say it would have occurred to me to
be *that* important! But of course libraries and so on would need some
form of standard.

Is there a preferred method of archiving documents to pdf? I'd assume
it'd be a page-to-image type arrangement? A quick search shows PDF/A:

https://www.pdfa.org/pdfa-faq/

And from there, Adobe Acrobat appears to offer the option.

The problem being that this may not be something someone 100 years from now
may have to hand. And the more layers of tweaking there are, the more ways
to get mistakes or problems there will be.

So PDF isn't 'preferred' at all. Although, as will millions of Word docs,
future historians may have to struggle with them when MicroSoft have long
gone.

Wonder why Keith saved them in that way?

I've exchanged a couple of emails with him about this. I think he felt this
was the best way to minimise file size whilst keeping visual quality.
Certainly OCR is useful for that. Problem is that this makes a number of
implicity assumptions about what rendering software people are using, etc.
Which may not be so for the far future.

Indeed, he said to me he scans at 600 dpi and doesn't use jpegs. But the
patchwork images in the file I looked at *are* jpegs, and not all 600 dpi.
I'm not sure yet if he knows that his PDF software may be doing that
without him realising.

Simple example of potential causes of problems. When rendered on-screen the
tendency for many OS is to assume a base of 72dpi. Yet the main OS I use a
lot of the time is based on 90 dpi. So the same 'size' of image on-screen
calculated in terms of a paper size specified by a PDF in inches (or mm)
may look better on one machine than the other. Then consider rips or
renderers for paper. There you'd have to rely on an lpi that is large
enough to ensure this won't matter.

What dpi will screens on devices 100 years from now be based upon? Seems
doubtful it will be as low as 72 or 90. But who knows (yet)?

Hence even if a PDF 'optimally' scales things for reading, that
'optimisation' includes some assumptions that may be false. And some PDFs I
see have no dpi values for some images.

Hence the simplest approach is to keep to plain bitmaps with a specificed
dpi. One bitmap per page/scan. This removes all the added 'clever'
processes that give more ways to slip up.

Jim

--
Please use the address on the audiomisc page if you wish to email me.
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Armstrong Audio http://www.audiomisc.co.uk/Armstrong/armstrong.html
Audio Misc http://www.audiomisc.co.uk/index.html