Store data on paper with Twibright Optar [LWN.net]

Twibright Optar is a new and unique software project by Karel 'Clock' Kulhavý, developer of the Ronja optical network link project. Here's the project description:

Optar stands for OPTical ARchiver. It's a codec for encoding data on paper. Optar fits 200kB on an A4 page, then you print it with a laser printer. If you want to read the recording, scan it with a scanner and feed into the decoder program. A practical level of reliability is ensured using forward error correction code (FEC). Automated processing of page batches facilitates storage of files larger than 200kB.

One may wonder why, in this high tech world, would you want to use paper as a data archive medium. Paper tape and 80 column punch cards went out of style in the early 1980s. Optar is probably not for those who are intent on running a paperless office. Here are some unique benefits and features of Optar:

It can be used for storing images, sound, and any other type of data.
It requires a 600dpi laser printer and scanner.
Data is printed on 3x3 pixel dots and encoded with Golay code spread across multiple strips.
It prints synchronization imagery to handle media nonlinearities.
It is useful for sending data through the postal system.
It can be used for publishing data in magazines and other print media.
It can be used to archive data on microfiche.
Data on paper can be notarized and used for legal archives.
It has a much greater data density than printed characters, reducing the size of mandated paper archives.
It is useful for data distribution to those without network access.
Error correction is able to handle folding of the paper media and other noise.
Optar images can be quickly duplicated with a scanner or a digital camera.
Paper media is less expensive than floppy disks and USB memory sticks.

Usage of Optar is fairly straightforward, the optar command encodes data into a series of .pgm files. Those can easily be converted to PostScript with the convert command from the ImageMagick suite, then printed to most laser printers. Conversion from paper back to data involves scanning the pages with SANE or other scanner software, saving as .png files, then feeding those to unoptar, which outputs the original data.

While functional, Optar is still in an early stage of development. Some desirable options would be the ability to select output paper sizes such as US letter and legal on the command line, and choose the encoding density. The documentation is currently limited to a README file, there are plans to make man pages for the two Optar commands. The code is without a version number at this point, presumably because there is only one version that has been released.

Optar has been released under the Gnu GPL, the source code is available for download here. The code is written in C and builds with the standard make and make install commands.

to post comments

Store data on paper with Twibright Optar

Posted Jul 26, 2007 11:10 UTC (Thu) by jengelh (subscriber, #33263) [Link] (2 responses)

>* It is useful for sending data through the postal system.

In fact, you can even send it over RFC1149 links!

Store data on paper with Twibright Optar

Posted Jul 26, 2007 19:28 UTC (Thu) by flewellyn (subscriber, #5047) [Link] (1 responses)

It's a bit higher-level than an RFC 1149-compliant AAN (Avian Area Network), since you're sending data in an application-specific encoding, rather than as hexdumped IP datagrams.

But I suppose the principle is the same.

Store data on paper with Twibright Optar

Posted Jul 26, 2007 19:31 UTC (Thu) by jengelh (subscriber, #33263) [Link]

Well, print it, and stick the paper on an Avian Carrier (most people will wrap it around the carrier's feet).

Store data on paper with Twibright Optar

Posted Jul 26, 2007 13:04 UTC (Thu) by climent (guest, #7232) [Link] (2 responses)

Remember those times when some code was taken outside US using printed form, because the code could only be exported that way?

Optar could have been of help automatizing the process :)

Store data on paper with Twibright Optar

Posted Jul 27, 2007 16:30 UTC (Fri) by scarabaeus (guest, #7142) [Link]

Yeah, that was the first version of PGP - it was published as a book by MIT Press. :-)

Store data on paper with Twibright Optar

Posted Jul 28, 2007 19:24 UTC (Sat) by ekj (guest, #1524) [Link]

That was PGP. I know 'cos I participated by proofreading the OCR of 25 pages of it. Got it in the mail, big brown envelope. The OCR-data by email. Compare and correct. Great fun :-)

Store data on paper with Twibright Optar

Posted Jul 26, 2007 19:03 UTC (Thu) by eli (guest, #11265) [Link] (6 responses)

This could be interesting as an element in posters, handouts, art, etc.
to embed a 'hidden' message. It would be particularly interesting to see
if one could embed the source materials of an image in the image as an
opttar fragment.

Hmmm.... Should we keep a printout of the optar source with the data?

Store data on paper with Twibright Optar

Posted Jul 26, 2007 20:26 UTC (Thu) by hein.zelle (guest, #33324) [Link] (1 responses)

Printing the source on the back of the paper might not be a bad idea :-) (in ascii form, notably)

For archiving and notary purposes it'd sure be handy to know afterward how to decode the data.

Store data on paper with Twibright Optar

Posted Jul 29, 2007 9:05 UTC (Sun) by AndyBurns (guest, #27521) [Link]

> Printing the source on the back of the paper might not be a bad idea :-) (in ascii form, notably)
>
> For archiving and notary purposes it'd sure be handy to know afterward how to decode the data.

Print the source of a C compiler too, oh I spot a flaw ...

Store data on paper with Twibright Optar

Posted Jul 26, 2007 22:51 UTC (Thu) by elicriffield (guest, #33738) [Link] (3 responses)

Hey i thought i was eli.....

I don't know about the color requirements but seems like you could make pictures that don't look to abnormal, maybe more like a painting that would have 100k of data in them.

It'd be a neat trick for an album cover.

Eli Criffield

Store data on paper with Twibright Optar

Posted Jul 26, 2007 23:24 UTC (Thu) by eli (guest, #11265) [Link]

Hey i thought i was eli.....

Heh, so did I. ;)

Eli Carter

Store data on paper with Twibright Optar

Posted Jul 29, 2007 9:07 UTC (Sun) by man_ls (guest, #15091) [Link]

Heh! You could even encode an image in an otherwise normal album cover, and in the encoded image (e.g. a PNG) you could also encode another image. Ad nauseam...

I don't think you could create an auto-encoding image: one which when decoded yields the same image, but it's an interesting problem to be sure.

Store data on paper with Twibright Optar

Posted Aug 4, 2007 13:29 UTC (Sat) by pimlottc (guest, #44833) [Link]

PARC has a (non-free) 2D barcode system called Dataglyphs, composed of patterns of simple fore- and back-slashes, which can be used as halftone dots, changing their thickness, lengths and colors as needed to render an image. It's very nice and unobtrusive, especially at high resolution. There's a bunch of example images on their webpage and even an interactive demo page where you can encode and decode your own test images.

The density is not as good - 1 KB/inch^2 @ 600 dpi (comes to 96KB for an A4 page) - but the simplicity of the patterns makes it fade easily into the background and allows for a good amount of alteration for aesthetic purposes which still being machine-recognizable.

Store data on paper with Twibright Optar

Posted Jul 27, 2007 12:46 UTC (Fri) by pkmorrison (subscriber, #6111) [Link] (1 responses)

Does anyone remember "paperbytes" from Byte magazine?

Store data on paper with Twibright Optar

Posted Jul 31, 2007 0:33 UTC (Tue) by ssavitzky (subscriber, #2855) [Link]

I remember paperbytes. Very clever, though it required a special reader and the density was pretty poor. This looks considerably denser, but it's still not clear how useful.

Store data on paper with Twibright Optar

Posted Jul 27, 2007 21:52 UTC (Fri) by riddochc (guest, #43) [Link]

This is a cool project!

It reminds me of those 2d barcodes you see on printed shipping labels -- seems like exactly the same idea.

Store data on paper with Twibright Optar

Posted Jul 30, 2007 22:21 UTC (Mon) by cook (subscriber, #4) [Link] (1 responses)

The author of Optar has responded to my comments about support for
US letter, the source code now contains the necessary constants
for that paper size. One wonders if this technology could be stretched
to higher densities by using a limited number of easily recognized
colors for the dark pixels. This might mess with the error correction
concepts a bit, but it seems like it should be possible to get at least
a gigabyte of information on a page of paper.

Store data on paper with Twibright Optar

Posted Jul 31, 2007 21:19 UTC (Tue) by roelofs (guest, #2599) [Link]

One wonders if this technology could be stretched to higher densities by using a limited number of easily recognized colors for the dark pixels. This might mess with the error correction concepts a bit, but it seems like it should be possible to get at least a gigabyte of information on a page of paper.

A gigabyte?? That's almost four orders of magnitude beyond its current capabilities... Even in principle, using every single dot as a bit, you couldn't exceed 4MB or so (8*11*600*600/8). If you assume 8 colors, then maybe 32MB is vaguely imaginable, but that has to be close to an absolute upper limit. (Yes, I realize "dot" is ill-defined and that printers have some ability to vary the size or ink-density of individual ones, but I don't believe that capability is anywhere near precise enough to depend on for something like this, particularly given wear on print-heads/drums.)

Also, you'd use a bunch of ink/toner on each page. And assuming we stick to the current claims, how much do 10,000 pages of paper cost? My gut feeling is that it's way beyond the $20 or so you'd need for a 2GB thumb drive (but since I never buy paper, I have no idea).

Greg

Store data on paper with Twibright Optar

Posted Aug 3, 2007 6:45 UTC (Fri) by muwlgr (guest, #35359) [Link]

In USSR since about 1985 you could see hexdumps published in popular science-technical magazines. Often on multiple pages of each issue. These were firmwares and programs for several models of DIY and industrially-produced i8080-compatible computers. Sometimes there were even octdumps for PDP11-compatibles :>

In exUSSR, since about 1992, there was a real invention called ArVid - http://en.wikipedia.org/wiki/ArVid . This ISA board allowed to use VCR/VHS tape recorder as a streamer and store up to 2GB per 180-min tape.

This OPTAR is great project. It reminded me of ones above. And being on its site, be sure to go to its root and look at Ronja as well. A great and creative radioamateurs' excercise in survival :>

Store data on paper with Twibright Optar

Posted Aug 5, 2007 10:51 UTC (Sun) by johnny (guest, #10110) [Link]

This reminds me of a system used for URLs in Japan. You'll often see a little graph together with web page URLs, and you can take a picture of the graph with your mobile phone instead of taking a note of the address. (More info: http://en.wikipedia.org/wiki/QR_Code)