• Please review our updated Terms and Rules here

Digital Equipment Corporation - MicroFiche Underground

RSX11M+

Veteran Member
Joined
Feb 14, 2011
Messages
1,075
I'm starting this thread in an attempt to bring to the community, access to DEC's documentation, published on MicroFiche.

SUMMARY:

This Underground is being started in the hopes that it will become a collecting point for others out there, like myself, who have original issues of DEC Microfiche sets.

These sets contain incomparable engineering and service details on DEC products. Much of it, unavailable in any other venue, and in danger of being LOST forever.​

INTENT:

We are banding together, to transcribe to modern media and methods, the entire scope of knowledge contained in the various DEC MicroFiche sets.​

To accomplish this, we will do the following:
  • Register the fact of our possesion of Dec Microfiche sets with the Underground
  • Exchange Details of the contents or existence of various sets [whether or not we have them]
  • Obtain assistance [donations of equipment, software and funds] to accomplish the transcription
  • Support the output by distributing and hosting it on the internet
  • Accept requests for documents in the fiche library - to prioritize the transcription process.

APPEAL:
Please - If you are aware of any of these sets, let us know. If you possess one, consider donating or lending it to the Underground for the purposes of keeping it from being lost.​

This will be a non-profit endeavor, for the posterity of the computing industry and mankind in general. Let there be no more "Ancient Libraries at Alexandria" - Lost.

P.S. - Feel free to PM me, if you have privacy concerns with your information or contribution.
 
I've tried to access the Fiche set I have using ad-hoc equipment I've cobbled together. [an overhead projector with altered focus and lenses] It hasn't worked because the optics were not intended for the kind of resolution required, and display chromatic aberrations that interfere with images of individual pages on the Fiches.

I've taken the set to the county library, and verified they're in great shape, so at least this much is known.

I either need a real Fiche Scanner, or a standard Fiche reader I can take digital photos of with my camera.

The process will look something like this:

  • Convert Fiches "en mass" [sheet at a time] and parse the images -or- Digitally Photograph individual pages on the fiches
  • Images will be recorded and catalogued in a database, accumulated, and included in a preliminary PDF file. This will not include any OCR of the content.
  • Next, using either images or the Image-Only-PDF files as the source, each page will be put though OCR and edited to be sure photographic and diagrammatic content is excluded from the OCR process, and that text and captions correctly frame other content.
  • Once an entire PDF has been processed, a suitable TOC and INDEX will be added.
  • The finished Document will be published [along with the original image-only based one] for peer review.
  • Corrections and revisions will be applied as required and the document will be "finalized" after suffient feedback.
  • At this point, the image-only based version will be "archived" and removed from the public repository, still available upon request.

Individuals wishing to contribute time, or resources will be able to check out image pages or documents to assist in the OCR conversion process once a set of software to accomplish the work is decided upon.
 
It has been suggested that a minimum resolution for a direct scan of Fiches is 7500dpi.

I have found some photo scanners that aren't too costly that claim 9600dpi optical, 19200 interpolated.

I'll try to check out one of these with a real fiche, at the next opportunity.

It would still be good to have an old style microfiche scanner, but this may have to do for now.
 
More on methodology...

Since helpers are not jumping out of the woodwork to participate in the project, the following will probably not be required.

The original hope was that a hi-res scan of each fiche be taken and cataloged. At 7500dpi no compression, this is appx 1.2GB per fiche. I'm guessing I have 6" of fiches at 10 mils each. That's 600 Fiches or 675GB of storage. Not too bad really.

So, I can get a ~1TB drive, do the scans and then hand out individual images of Fiches to an army of participants to parse out individual document pages, and check them into a database.

These same people could then PDF them on a per document basis and enter completed image-only PDFs into another database.

A collection of these should have storage requirements no larger than the original fiche scans. Smaller in fact since there is considerable space between pages on a Fiche.

Next step is the "army" then checks out a page at a time to OCR them. These are again entered into the database and ultimately to another PDF, but with mixed text, images, TOC and Index.
Note: Usually DEC documents have a TOC and Index of their own. In this instance, it may be feasible to make these "active" using native PDF techniques.
At this point they'd be checked back into the database and available for peer review.

The entire archive [this one anyway...] ought to fit in 2TB or less.


Once the final products are completed, is there a need to keep the intermediate images?

If scanners of this resolution will always be available, then probably not. However, film is pretty obsolete, and it is possible that scanners are themselves achieving obsolescence in which case keeping the images will be cheap insurance.
 
Test of 1200dpi scanning of Fiches directly

This is an example of a test image from one of the DEC Fiches. The typeface is actually larger than that on most documents in the archive. You can see the resolution is insufficient, but not as bad as one might expect 1200dpi optical resolution to yield. [the photo album process of VCF has reduced it's size, but not the clarity]

attachment.php


This gives me some faith that 4800 or 9600dpi images of Fiche document pages will give acceptable results.

I'll try to test a Canon LiDE 700F flatbed film scanner in a local shop here some time soon.
 
Slight detour since my previous post.

We've received the donation [almost... we had to pick it up] of a real Microfiche reader. Optical quality is sufficient to make the fiches legible and as a result, we now have access to the contents. Photographs of pages should be possible, although a little laborious.

The Canon scanner is still to be evaluated, as this reader is not a scanner.

Note: I've seen a Canon MS-400 Reader / Scanner that's very nice [SCSI]. I'm in negotiations, as this will not be a "donation".

One final note, the fiche set we have does have DEC Rainbow, Pro, and DecMate contents.
 
MicroFiche Transfers

Thought I'd update the blog [in case anyone is watching] with some early MicroFiche transfer results. Frankly these are disappointing for a number of reasons.

They are photographic transfers. This means they're subject to all kinds of lens aberrations, and to the camera's idea of where to focus, which seems a little off. These don't stand up to the sense of visual clarity one gets when looking at the reader.

The processing required to get this far is more than I'd hoped too. They start out as color images.

  • I begin by cutting an area out that looks like a document page.
  • Then, it's changed to Greyscale, adjusted for contrast and brightness, and "inverted" to get the look of paper rather than "Blue print".
  • At this point some compromises in field flatness are already apparent.
  • A SHARPEN filter is tested on them, which usually makes an improvement.
  • Then it's saved under some meaningful name along with other attempts.
  • A final review selects the winner and the other "working versions" are deleted.

These are from the Rainbow 100 Technical Manual Addendum.

1-4.jpg


3-4.jpg


3-5_c.jpg


3-5d.jpg
 
Original Camera Image for Comparison

Here's an example of what the final image [page 3-5d] in my previous blog post looked like directly from the camera.


DSCN4893.JPG
 
A volunteer helper (at least one)

RSX11M+;bt403 said:
Since helpers are not jumping out of the woodwork to participate in the project ...

Hi,

As I have some experience yet in scanning and processing (tif to pdf) DEC paper manuals I'd like to participate in working on scanned microfiche data.

If I'd been given the opportunity to choose, I'd prefer microfiches on old VAXen (VAX-11) hardware and old VMS versions (anything before VMS V5.x), but I will work on any other fiches as well of course.

Is there already an inventory of the microfiche sets available to the UNDERGROUND?

Regards,

Ulli
 
Back
Top