we're looking at rfrank's "roundless" experiment at fadedpage.com... *** ok, it's been a while since i went through this little drill, so i'm gonna give you a little refresher-course on where i stand on a workflow that distribution proofreaders could be using, particularly from the viewpoint of rfrank's roundless research. *** first off, i would substantially re-ramp the "preprocessing"... this is the first step, where you're doing the scanning and o.c.r., or fetching the results from one of the major scanning projects. we'll start at the very beginning... *** i've advocated -- strongly -- a filenaming system that indicates, for each file, the contents of the file, so the name is meaningful. specifically, the _page-number_ of the file should be included in a consistent way in the name of a file. thus, for instance:

http://z-m-l.com/go/myant/myantp123.jpg

the image-file for page #123 in the book "my antonia" is named myantp123.jpg. the "myant" prefix is common to files for the book, and the "p123" part refers to page 123. this is so straightforward that you couldn't be faulted for assuming that it's taken for granted. but it's not. many of the content-providers over at d.p. name their files differently, so the filename is _not_ unambiguously tied to the page-numbered contents inside; and a price is paid for this unclarity. yes, rfrank is one of these ill-advised content-providers, and he has carried over this bad habit to this experiment on roundless proofing. this shortcoming is even worse in a roundless system, because there is a need to refer to specific pages in such a system, and the absence of a sensible naming-scheme therefore becomes a bigger problem... (in a round-based system, where all the pages are treated as a "batch", the problem isn't as bad, but it's still something that should be fixed.) *** next, there are a number of things that are done in "preprocessing", at distributed proofreaders and by rfrank in his roundless experiment. some of these things should _not_ be done, in my considered opinion. on the other hand, there are other things that _should_ be being done. these are some things that are done which should _not_ be done: 1. run-heads and page-numbers are eliminated. 2. end-line hyphenates are being joined. 3. end-line em-dashes are being "clothed". these are some things that _should_ be done, which are not: 1. obvious and easily-located problems should be fixed. 2. spacey-quotes should be fixed. 3. ellipses should be standardized on a 3-dot ellipse. we can engage in debate on all of these suggestions, but it will be instructive for us to see some of the o.c.r. errors i'm talking about. appended to this post is a list of bad words or lines that were pulled from a book currently being proofed at rfrank's roundless experiment. these are almost all errors that will certainly need to be fixed. aggressive preprocessing can find these errors without looking for them. that's very important, in a roundless system, because if you can find and fix all the errors _before_ a page is subjected to a word-by-word review, then that word-by-word review can become the first "no-diff" in a chain of "no-diff" reviews, and the shorter that chain is, the more efficient you are. in the other scenario, if the page is dirty, you might have to have 1 or 2 (or even 3) proofings before the page is clean enough to receive a no-diff, meaning your efficiency has plummeted. there's no reason to make a human _search_ for errors when those errors can be located quickly and efficiently by a computerized search routine... -bowerbird p.s. sorry i've been sluggish with this series. i thought i'd been away from this stuff long enough that it wouldn't bore me to do it all again, but so far it has been a drag... i can only repeat this so many times... i'll try to get the motivation back again, but no promises if i cannot... p.p.s. here's that list of "probable errors" pulled out of rfrank's book: of'more Fd say pver t'other Enew a'slipped pn curtiss-robih somethin'to ght asighin' buncft whick mother'Ship ground^igood hefteaw iny wonderf ^Devolutions punkins outen tpreviously bustl gun'ls thaft ag^ny sud-v denly ij J>ack jumpin'his blame'em Jiack apture twise Weil numba chorteled pvounded Wowl oaly ^{w}Oh! haulded uae vre knovt apeak gink givin* stretch.'* J'ack wheen clost you'l althought eyefull weuns I"<cr> oa pHot etchin' jest'magine iframe ha.nd you'fe valk a'been outen morc'n MCGrath tc Unc' wuss'n pizen fresfi hirn orr hinsel| But you never can tell just what may hap]7^{en sion of weird noises springing up from the g<9^{a}^ 74-* /' 161 17S "Working over a bird with red feathers,'^{1 fall for such a decent game 93 taxidecentry or 18? was the work of a few second3. Hardly had the THE COMEBACK 227 HUNT OF THE S-18 belonged to the Hun pHot, Oscar Gleeb. must be pesU,@Jn' you like all get-out, so I made cruel to keep me a'guessing any longCi. so anxious to get started oA their way <i> Porter Press </i> disappears from an airpnQp .Watch my smoke, that's all." .until finally it died away completely. This gave .keep your eyes fastened on him. Whatever . A fortune hangs in the balance when young Dan Tierney, press , "Gripes! that was worth somthin'to glimpse, ; for Perk valued a few words of praise ;whooped the delighted Perk as he squatted pn you to hold my own. That's j^t how it should say,^{r} she bust out o' that little fog cloud right such as would tell ^ business being put through rendezvous and^ it's our game to chase after them, light, whether <i>^{r</i>}'and or gulf, the chances were holding ground^igood--a heavily laden sailing path, and going through the most wonderf ^Devolutions. and be ready to pounce down on their inten<J^{e}d prey after the fashion of a hungry eagle striki <i>^{n} a fat</i> duck that had been selected out of the flc^{c}k One thing he did do was to cut his intend^ wide circle short and again head toward ^{ne }scene of action, a move that certainly afford^ the eager Perk more or less satisfaction, he bel^{n} thrilled with the expectation of breaking into <i>th^{e }game</i> without much more loss of time. But you never can tell just what may hap]7^{en }when rival forces are striving against one ^^{n}' other. The best laid plans often go wrong a^{n}^ there was always a chance of the unexpected happening. Hardly had the airship whipped around ag^^{n }so as to head into the north than Perk beca<"^{e }aware of the fact that there was a sudden acc^{es}' sion of weird noises springing up from the g<9^{a}^ toward which they were now aiming. Jack, t^ must have caught the increased volume, for ^{ne }sheered off as if to hold back a bit so as to gr^^{s}P the meaning of the new racket. Men were no longer simply talking or laug^h" ing as they so cheerfully labored in transferring some of the contraband from the sloop to ^{ne }deck of the speedboat--their voices were rai^^{e}^ to shouts in which surprise, even the element ^ the frenzied sufferers in their ag^ny had been ^{w}Oh! That can be put through without muck said that name exactly three tin^s, like it meant operator as Oswald Kearns pick oui^an ordinary would rather have Jack praise him than ^ny one ^{r}fully five feet long and as thick through the body him, no matter where he goes-^-sorter dude, I'd will y^{u}> boy--two--three fellers jest swarmed "Working over a bird with red feathers,'^{1?? M^ans our gent has a raft o' ships comin' an' through fire or some similar means of destruc^ our man ditto. Mebbe now I'll soon^{x}get a chance "Gosh, amighty, we're flyin'\Mgh, buddy!"