- gutvol-d - lists.pglaf.org

Re: [gutvol-d] book of james -- 002
by Bowerbird＠aol.com 21 Dec '11

21 Dec '11

james- just hang loose for now. we'll get it all straightened out. :+) -bowerbird p.s. roger, that screenshot from earlier happened because i had "zoom text only" selected. now it resizes much better.

1 0

book of james -- 001
by Bowerbird＠aol.com 21 Dec '11

21 Dec '11

james said: > I am interested in trying out your tool. cool. let's mount it as a demo project. > I'm currently working on a book > that has a special meaning to me > but which is hell to format. ok. > http://www.archive.org/details/studyofbhagavata00benaiala i'll check it out. > I've more or less committed myself to > doing family tree tables using ASCII text. i'd suggest doing them as text _and_ images, and putting both in the book, as a fall-back... text can be copied; images preserve the look. > There are a LOT of these in the book. > I'd need some way to indicate that these > should be surrounded by "pre" tags in the HTML. that's easy enough to do. > Second, I need to have some illustrations. > How would your tool handle those? with simplicity, grace, and elegance, of course. ;+) > My understanding of your ideas is that > I can make a file with some simple markup > and generate many different formats from that. you understand correctly. > I used the HTML generator in guiguts for > my PG submissions so far. I'm not entirely happy > with it, especially the way it uses "span" tags > to indicate indented text. I end up going through > and replacing these with "blockquotes". > The style sheet guiguts gives me is good for HTML > but needs to be tweaked to produce a good Kindle book. > And of course once I've generated my HTML > I now have two files to correct. i hear everything you're saying. > Also, if the book has multiple levels of headings, > how would you deal with that? you simply indicate the level of the heading by the number of blank lines preceding it... for instance: > h1 heading = 14 blank lines preceding > h2 heading = 12 blank lines preceding > h3 heading = 10 blank lines preceding > h4 heading = 8 blank lines preceding > h5 heading = 6 blank lines preceding > h6 heading = 4 blank lines preceding or, if you prefer: > h1 heading = 9 blank lines preceding > h2 heading = 8 blank lines preceding > h3 heading = 7 blank lines preceding > h4 heading = 6 blank lines preceding > h5 heading = 5 blank lines preceding > h6 heading = 4 blank lines preceding note it's the _relative_ number that counts. (the absolute number merely determines how much vertical space the header uses when the book is output in .pdf format... some designers like to put it way down.) the program creates a nice _outline-style_ indented list of the headings for you, and keeps track of any change that happens, so once you've determined the outline is right, you can pretty much stop thinking about it. (if you've already used the h1-h6 terms, you can just do a global replace on them. if you like, you can _leave_in_ the h1-h6 terms, in addition to the blank lines, and wait to take them out until you "go final".) *** so i'll mount the scans. send me the text, and i'll incorporate it into the public demo. however... ...here comes the bad news... it might well be the case that what you want is impossible. because the kindle can't do it. is _incapable_ of doing it. not just kindle, but _any_ screen that's "too small" (e.g., a phone). for instance, most of the family trees that you gave as examples in your december 18th post _won't_ work -- dependably -- in _any_ system. > http://lists.pglaf.org/mailman/private/gutvol-d/2011-December/008590.html not as text, anyway. you can "fool yourself" into thinking they work, by setting the text-size down to a small level... but if you want a test, do the opposite instead: set the text-size to its biggest level, and then see how many characters fit on a kindle screen. you will then be tempted to say something like "well, as long as the text is a reasonable size..." but again, that's just fooling yourself, because if you ask a 50-year-old person, you might get a totally different answer m a 20-year-old as to what text-size constitutes "a reasonable size"... and your problem is worse than the "poetry" one, because your lines, if split, _become_gibberish_. so you are going to have disaster. perhaps not all the time, on all of your trees. but sometimes. and you need to face that fact before you start, or you're only gonna be disappointed later... so, if you wanna work with me, send me the text. -bowerbird

4 5

exploratory code
by Roger Frank 21 Dec '11

21 Dec '11

Though I've run many books through DP's proofing and formatting workflow, I always thought it would be easier and more fun if the computer gave feedback about what it thought of the page as it was being proofed. To test this out, I have written a multi-pane editor that may be useful for etext production. I loaded a book into the demo version of the program at http://etext.ws/ppe.php and I'm interested to know if it works for different users with different browsers. It uses several HTML5 pieces that I think are safe. Open source, MIT license, as is all my software tools for etext production. The left pane is a text edit area for a page. Center pane shows the results of analysis for that page. Top pane is a clickable image of the page. Various status information is lower right. Still to do: spelling check, common scanno flags and many more regex checks. Right now it includes spaced or crunched punctuation, poetry detection, end-of-line hyphens (which may be resolved with the DHYP button) and a few others. I also want to use a magnifier jquery plugin on the image to dynamically enlarge the area under the mouse instead of having to open another window. I will solo post-process a few books using it. If I can make it work well enough, I'll add the page queue code (i.e. "next unproofed page"), instead of how it is now, where you select the page to proof or it gives you a random page. As you proof, the status area displays how many times the page has been proofed, as this is all tracked along with all earlier revisions in the database. There is no login required and I'd appreciate if you would take a look at it. Feedback welcomed to the email shown in the application. I'll summarize (if there is any interest) back to this thread. --Roger

2 2

gerbils expose chilling factoid
by Bowerbird＠aol.com 21 Dec '11

21 Dec '11

keith said: > just submit pandoc, and will > we will see what Lee has to say. oh, so now you want to try and send _me_ on that wild-goose chase! no way, mister. :+) but really, keith, what did you mean by that puzzling "no human intervention" comment? do you have some gerbils who'll run the test? :+) *** that $100 "challenge" can't be taken seriously. for anyone who wants to know why, seriously, i am all too happy to explain the subterfuge... lee _thought_ that he could demonstrate that light-markup languages are not fully specified. thus, by requiring different ones to produce what is essentially the same underlying code, he thought he could expose the holes in each. what he didn't know is that the schemes _are_ fully specified, and thus can indeed create code which accomplishes even a fairly complex book, in a fairly standard manner, such that that code would indeed qualify as "essentially" equivalent. in fact, the degree of full-specification obtained is complete enough that these markup systems can even be translated in a one-to-one fashion. and not only can they "be translated" that way, the conversion can be done _programmatically_, and pandoc is the existence-proof for _that_... so lee was significantly underinformed, and his "challenge" merely exposed that chilling factoid. but lee is a clever and devious person, who knows how to cover his ass, so he found a way to do that, in the off-chance that he unknowingly was wrong. his challenge specified that _he_ would prepare the various light-markup files, meaning he could _sabotage_ the full-specification of the schemes. lee made it _appear_ as if he was betting that you couldn't make a basket shooting from mid-court. (or what he believed would be a mid-court shot, even though, with pandoc, it was a mere lay-up.) but, with a little shift at the end, he changed it, such that he was betting that _he_ would miss a shot from mid-court. would you take _that_ bet? would you take a bet that he could miss a lay-up? ok, fine, lee wanted to make sure that he wouldn't have to give away his $100. who could blame him? but to think this proves anything is just ridiculous. nonetheless, a simple contradictory point remains. namely, none of the benefits of any light-markup system _ever_required_ that such a system could be converted flawlessly into _other_ light-markup. the object of light-markup is to convert to .x/html, .epub, .mobi, .pdf, or some other _output_format_, and _not_ to convert to some other _input_format._ i mean, it's _nice_ that pandoc can convert between markdown and restructured-text and asciidoc, and many other formats, but it's certainly not necessary. markdown does what it's _meant_ to do, regardless. and so does restructured-text. and so does asciidoc. lee is trying to shift your attention to a trivial thing. he wants you to focus on a meaningless "conversion" between light-markup formats, instead of the _more_ important matter that light-markup converts to .html. because lee wants you to do your authoring in .html, not light-markup. but light-markup is much simpler, so i'm not really sure why he wants to waste your time. and let's be absolutely clear about something vital... this choice of a "master format" does _not_ entail any need to give anything up. you don't have to "choose" between using light-markup and .html. because light-markup can be converted to .html... with the simple click of a button -- boom! -- html! not to mention .mobi. and .pdf. _automatically_... -bowerbird

2 1

Re: [gutvol-d] book of james -- 001
by Bowerbird＠aol.com 21 Dec '11

21 Dec '11

carlo said: > You seem to be unaware that > you can get the text for a single page > from the Internet Archive djvu files, through > the djvutxt command and the -page option. holy crap... i did not know that... i figured that there _had_to_ be a way to do it, since the .djvu knows the text on a page, but i didn't know how. how/where exactly does one issue this command? -bowerbird p.s. james, if you could still zip up your files, that would let me splice in your corrections...

4 4

Re: [gutvol-d] book of james -- 001
by Bowerbird＠aol.com 21 Dec '11

21 Dec '11

don said: > If you read the whole page > it discusses pulling djvu text sigh. i read the whole page. you can pull the djvu text for the entire book, but -- once again -- i want it for _each_page_. -bowerbird

2 1

a review of some digitization tools -- 022
by Bowerbird＠aol.com 21 Dec '11

21 Dec '11

david garcia (dpname="donovan") said: > I'm surprised that no one has yet > mentioned pandoc as a useful piece of kit. i'm surprised you weren't already in my kill-filter, donovan. but you are now... :+) *** anyway... last week: > http://lists.pglaf.org/mailman/private/gutvol-d/2011-December/008456.html last month: > http://six.pairlist.net/pipermail/markdown-discuss/2011-November/002390.html > http://six.pairlist.net/pipermail/markdown-discuss/2011-November/002392.html 6 months ago: > http://www.thebookdesigner.com/2011/06/my-monospaced-life/ 1 year ago: > http://lists.pglaf.org/pipermail/gutvol-p/2010-July/000074.html 3 years ago: > http://radar.oreilly.com/2008/09/why-you-should-care-about-xml.html 5 years ago: > http://six.pairlist.net/pipermail/markdown-discuss/2007-January/000488.html i didn't bother to mention pandoc to you idiots because you were so convinced that light-markup wouldn't work, and willing to bet (and lose) all your credibility against it. *** also, since we're on the subject of pandoc once again, someone who wanted to win that $100 prize from lee could convert my .html output for "books and culture" into markdown and restructured-text, using pandoc... he has no credibility left. might as well take his cash too. -bowerbird

3 2

Re: [gutvol-d] book of james -- 001
by Bowerbird＠aol.com 21 Dec '11

21 Dec '11

don said: > Even simpler, you can do um, no, that shows the image. (buried inside the viewer-app.) i want the _text_ for the page. -bowerbird

1 0

Re: [gutvol-d] book of james -- 001
by Bowerbird＠aol.com 20 Dec '11

20 Dec '11

james said: > Bowerbird, > If you check the link I gave you > you'll see what I'm up against. i see it... it's not an easy book, but it's not a hard one, either, not if you accept that some parts are impossible. :+) i've already set up its online skeleton: > http://z-m-l.com/go/bhaga/bhagap123p.html i can make it so you can work on it online. or offline, if you prefer, makes no difference. online would let other people work with you, if you'd be interested in having that happen... (if you want to work on it online, by yourself, just don't tell anyone where you've located it.) what i will need from you, as soon as possible, is a version of your text-file with the pagebreaks marked, so i can splice the text into the skeleton. a simple line of equal-signs will do the job fine... this is the last line of page 97... ========================= and this is the first line of page 98... again, as soon as possible, so i can do it right away, so things will be ready for you shortly after christmas. -bowerbird

3 2

Re: [gutvol-d] book of james -- 001
by Bowerbird＠aol.com 20 Dec '11

20 Dec '11

james said: > I do have the book as 1 text page per file. ok. (i think.) > I got it this way by downloading the page images, > making TIFFs out of them, then running tesseract oh dear. that was a waste of time. > I can give you the separate text files > in a Zip archive if you wish. are these the files after you made your corrections? if so, then yes, those are exactly the files that i need. zip 'em up, and put it in your dropbox. > My work method is to use guiguts to > remove page numbers and reformat paragraphs first. oh dear. more wasted time. oh well. (also, removing pagenumbers is the _last_ thing to do. they help let you be aware where you are in the book.) > The link you gave gives me a 404 error. yes, here's the correct one: > http://z-m-l.com/go/bhaga/bhagap123.html sorry about that... > I'm not sure what you mean by online. i mean you do your corrections on the web... which means that other people can help you. (at least if you give them the web-address.) but if you prefer to work offline, you can do that. > I thought you would > provide a command line utility i'm a mac person, james. we believe in a friendly interface. only a sadist seeks to saddle you with command-line crap... > that would convert ZML to the various formats. but first you have to get your text _into_ .zml format. > Were you thinking of something like DP uses? "something like" that is a fairly accurate description. my system isn't nearly as convoluted or bureaucratic. -bowerbird

2 1