The OCR Mystery

[captlychee]

Back in the day, BYE magazine had an d. It was a picture of a pair of verylovley green eyes, with the caption 'the only character recognition system that beats ours'.

Apropos of that, I have been going through the old files and correcting the OCR errors in a scanned story I wrote in 1977. This has proven fun in the 'decrypting' stage where I try to make sense of what is on the screen, and cringeworthy when I actual read what I wrote. /But I fondly remember the time sitting in a hot bedroom typing out stuff, and occasionally taking a break to dictate to Michael Kennett, who typed much more slowly than I did, as we worked on stuff that he would think up and I would write.

So, I pulled out another of these OCR'ed files and tried the same thing. I discovered immediately that it was going to be a lot harder with the second story. What can you make from this?

bout çhrt t ‘in’ briept ye” h’ ‘ ‘n ie ve°nn’t-e - inn” njn betntn rn ri “he ‘vlver’e ‘ovkln ar U ‘n o -‘r be” ‘n’i -rnn’ntnd m” s’dvsnai ‘ien4’ to g jntn rraptl’qe, ir ‘t datjp of the prier a? en,’oorine npoductinn laid ‘inun my most— 1” ‘ner.cs’ful lifent”rle. “hc’re ‘c so ‘iuch cdre’wlin t1ot ‘hrfl%,nh r’ bln’dntaep, t felt “e if’ I aoulR eve Ii n”t” feet in the ci-’.

Copyright © D J Rout 1977, 2017

Doubtless, mount_oregano has experienced worse writing in her medieval manuscripts, but that is IRIS software's attempt at deciphering my typing.

The only differences between the typescripts of the stories is that the first one is written with my old Imperial on cream-coloured paper that was, forty yeas ago when I wrote it, white and that the sample above was written on blue paper using an Olivetti Lettera 32. I can't even attempt to decipher what's up there, and that's a relatively good bit, with real letters and not just symbols and commas. I stll have the originals somewhere so I could rescan them with better OCR now than in 2004, but that will take some time. I have the original TIFFs to read, but that's proven difficult enough for the first story. I don't know how best to view TIFFs in Windows 10, but if I found a way I could put the TIFF alongside the Word document and transcribe that way. It would be like re-writing the thing and, because of its cringeworthiness, not much fun at all.

We shall (blurrily) see.