March 2010 - gutvol-d - lists.pglaf.org

Re: New Tool "pgdiff"
by Bowerbird＠aol.com 18 Mar '10

18 Mar '10

jim said: > Post your tools, including source code, and then let’s talk about it. maybe you're too new here to know this, but i don't post my source code. even if i did, it wouldn't do you any good, unless you can use realbasic... likewise, your code doesn't do me any good, because i don't deal with whatever language you've posted it in, so you have done me no favors, and thus i don't feel any need to "reciprocate" for your posted code... but what _might_ do you some good is for me to talk in pseudo-code, if you're interested in hearing that, which i am more than happy to do. but none of this is difficult. especially from a line-based perspective. you read one file into one array, and the second file into a second array, and then compare the two arrays, item-by-item. it ain't rocket-science. the main difficulty in any comparison routine is the re-sync process; but if you work in a line-based way, your lines don't get out-of-sync. (in the rare cases where they do, you can make a manual adjustment.) so with this, the interface is more important than the underlying code. but i'm even willing to post compiled versions of a comparison tool, so that you can get a very good idea about the interface i am using, provided you can get a handful of people -- i.e., 5 people -- to say publicly, right on this listserve, that they would like to see my tool... but i haven't gotten the impression that anyone here can code a g.u.i. for an offline app. i'd love to be wrong about that, so please please do correct me, anyone listening out there, if you can indeed do that... -bowerbird p.s. if you can't get 5 people to say "please", then you should read the design description that keith wrote up last night, as it's decent.

3 2

Popular Science back issues
by Dan Weber 18 Mar '10

18 Mar '10

Sorry. The address is http://www.popsci.com/announcements/article/2010-03/new-browse-137-years-pop sci-archive-free I had to grab the jpg files from my temp internet files folder (Win Vista) Hopefully someone else will know a better way to get them They say they are partnered with Google danweber(a)mindspring.com

1 0

Re: New Tool "pgdiff"
by Bowerbird＠aol.com 18 Mar '10

18 Mar '10

keith said: > BB you wanted some clarification so I will try. um, ok, i guess. > First, I do not have the time to look closely at either > tool nor overs, as it would take to long for me to > properly analyze them, and due the analysis merit. um, ok, i guess. :+) > So I will proppose a possible design. > Take the good, leave the bad. um, ok, i guess. ;+) > In my opinion would be a tool with four windows/frames: > 1) the scan > 2) version 1 > 3) version 2 > 4) the proofed version ... > The scan (1) would be optimally syncrhonised > with the passage being checked. ... > Now, we have three cases to consider: > a) version 1 is correct > b) version 2 is correct > c) neither 1 or 2 is correct > d) the case where 1 and 2 are correct is > actually not possible in our context > unless version 1 and 2 are comming from different edition. > Still we can handle this in the same manner as c. ... > Hope this helps. If not hit delete. ok. now i'm curious... :+) keith, how do you think i pull off all the comparisons i do? how do you think i can sling around lists of diffs like i do? > http://z-m-l.com/go/gardn/gardn-hybrid6.html how do you think i can mount entire books with diffs? > http://z-m-l.com/go/gardn/gardn-hybrid7.html how do you think i resolve the diffs in all the books i do? i can tell you how i do it! i do it with tools i've programmed that do all the things that you talk about, and more. that's how i do it, keith. so you don't have to do hypothetical writeups, keith, especially if you're short on time, because i have a big batch of post-hypothetical reality sitting in my toolbox. it's not that your writeup doesn't "help". it's that we are past that point in time... and we have been, for a while. -bowerbird

2 1

a rose smells just as sweet
by Bowerbird＠aol.com 17 Mar '10

17 Mar '10

rfrank has informed people that he doesn't consider his site to be "competition" to the d.p. site... so from now on i'll characterize it instead as "an alternative" to d.p. it's all good. and a rose smells just as sweet... -bowerbird

1 0

Re: New Tool "pgdiff"
by Bowerbird＠aol.com 17 Mar '10

17 Mar '10

jim said: > You are attacking my reply re attaching licensing terms > by attaching it to unrelated discussions. jim, i'm not "attacking" anything. let's steer clear of antagonistic language. and if you had quoted what you were replying to, i would have _known_ the scope of your reply. don't blame me because i'm not a mind-reader. > Again, you are assuming the problem presupposes a solution > which is one of "Choose A" or "Choose B". well, since my tool lets a person _edit_ either choice a _or_ choice b, i think i've given the user plenty of leeway. > If you use the tool on other than trivial problems > you will find out that life is not that simple oh please, jim, wake up. i've actually _done_ the kind of comparisons and diff-resolutions that you're just _talking_about_. moreover, i have done them _many_times_. so i think i have much better _experience_ with the actual situation than you do. so don't try to "school me", ok? > frequently both A and B have some degree of errors that need > to be corrected and/or merged to get you where you want to go. now you're just repeating your "frequently" term, which i have already demonstrated to be false, using the most recent example i have shared. in all the resolutions i've done, the vast majority of diffs involved cases where _one_ of the versions was correct. very rarely were both wrong... > If one wanted to make a graphical tool to do this i not only _wanted_ to make such a tool, i actually _programmed_ it... > If one wanted to make a graphical tool to do this > you would not only need the "Choose A" and "Choose B" options > but "Edit in Context while displaying a copy of the original scanned page" you need to read what i wrote, jim. my tool gives people the option to edit either choice a or choice b, before selecting that particular choice. as for displaying the scan, does your pgdiff tool show that information? because if your tool can show that info, my tool can display the scan... but as far as i can see, from your limited example, you don't show that. (and when i say "my tool" in this post, i mean the tool that i wrote that _supports_ your tool. in the tools that i've programmed for _my_use_, i _always_ retain the page-scan information, so i can _display_ the scan.) > and if one wants to make that kind of tool one would be > better off to put the time and effort into figuring out a tool and here we wander off into the land of unnecessary complexity... > This is not my experience i've shared many of my experiences in doing the comparison method. if your experience doesn't match mine, you should share yours as well. > Post a portable version or one compiled for windows > and I will tell you how it works for me in practice. i need to have some questions resolved before i put it out in public. that's why i furnished you some sample texts to get pgdiff output on. -bowerbird

1 0

Re: New Tool "pgdiff"
by Bowerbird＠aol.com 17 Mar '10

17 Mar '10

jim said: > I decline to attach any verbiage at all. > I tell you I wrote it and you can use it any way you like > -- at your own risk and amusement, obviously. except that some of that "verbiage" was people asking just how exactly your program differs from one that they've been using all along. don't you wanna tell 'em? and as for me, perhaps you noticed i congratulated you for programming a tool. was that just "verbiage" to you? in addition, i will analyze any new tool, to check how well it performs the job for which it is intended. it's fine if you don't want to discuss it, but such a review is not "verbiage". it's necessary to take an objective look at our tools to see if they do the job, how they can do it better, and so on... you specifically said your tool helps in 3 areas: 1. line-break recovery 2. error-flagging 3. versioning you even said: > my new tool has several tricks > that haven’t been seen before (if anything has been "verbiage" in this thread, it's that!) so, at the end of this post, i'll begin to look at those 3 areas. > If you need to get more serious than that > contact me by email and we can talk about it. imagine the d.p. people had told you to make your complaints "via e-mail". i'd venture a guess that you would laugh at that... *** > On Vim I type: > :/[{|}]/ > Which highlights the edits and takes me to the next set of edits but that selects both the options, and the surrounding characters. that's not really what you want -- what _most_people_ would want. and it involves typing. either typing or a lot of delicate deleting. both of which increase the probability that errors are introduced. > When you are versioning it is frequently not as simple as > “choose A” or “choose B” but often a mix of both that > you have to edit. i'm sure i know the reality much better than you do, jim, because i've actually _done_ this resolution job, for lots and lots of books. but maybe rather than schooling me personally, you've said this for the benefit of the lurkers who might not have thought about it very much, if at all. (and that's an entirely appropriate thing to do.) but if we're going to enlighten them, let's do it properly, ok? your word "frequently" is simply (but completely) out of place. in the vast majority of cases (96%) where there is a difference between the two versions, _one_ of the versions is _correct_... there _are_ some cases where both are incorrect, meaning that you need to do some editing, but such cases are relatively rare. in the last book for which i did a comparison, gardner's text, there were 159 differences. there were only _3_ cases where _both_ versions were incorrect. so yes, it happens, but rarely. > And I like seeing each next to each other in context > to help figure out what the “correct” editing moves are. oh yeah, the context is _crucial_. but i'm not sure that your _display_ is the optimal one... it takes a lot of visual parsing to figure out a diff like this: > no way. There { warn't | wam't } a window to it big enough personally, i find this display _much_ easier to understand: > no way. There warn't a window to it big enough > no way. There wam't a window to it big enough > ================^^============================ (i hope the monospaced font came through. if so, you'll see the "^^" markers line up with the diff.) and i believe most users would agree that this display is better. but, you know, if some users like _your_ display better, _fine!_ :+) oh, and one more note on "context". sometimes it can fool you. the choice that looks right might not be what was in the book... that's why it's vitally important that your tool show you the scan. otherwise, you're doing your edits blind... > PS: You criticize me for doing that which the creator of wdiff > said he would do if only he had the gumption. you'll need to provide a little more information to be understood. > How do you use wdiff to recover lost linebreaks? i don't use wdiff for that. i wrote my own program. i asked carlo to explain how _he_ does it, but he never answered. i found it humorous he was willing to come out to challenge you, but isn't willing to come out when he is challenged... *** anyway, in order to "kick the tires" on your pgdiff program, jim, i'll set up some files that we can compare. (real books, real files, and not of my own choosing, either, but from rfrank's test-site.) i'll run the files when i next find myself around a p.c. machine... or, if you feel like it, jim, you can run them and post your output. once i have some real output to look at, i'll be able to do a much more thorough review of this new tool. while you're waiting for that, though, here's a screenshot of a tool that i wrote that makes it easier to work with jim's output. > http://z-m-l.com/misc/jim-tool-addon-screenshot.png basically, when it finds a line with a diff in it, it presents the options to the user, who can then click a button to choose one, or enter a number -- 1 or 2 -- to activate the appropriate button. in the case where editing is needed, either option can be edited before the button is clicked to select it. the "stop loop" button will stop the loop that presents the next diff display; otherwise, the app loops through the entire file, jumping to the next diff. so, you see jim, i'm really trying to _help_ you in your quest here. -bowerbird

3 2

Co-operative proofreading
by Lee Passey 17 Mar '10

17 Mar '10

Inspired by Mr. Frank, Mr. Adcock, Ms. Miske and Mr. Morasch, I decided to try to implement my own vision of a co-operative proofreading process. Anyone wanting to watch me flail about can follow my work at www.ebookcooperative.com. Login as guest, no password. Apparently the engineers at Microsoft have not yet figured out how to implement CSS percentages, and I haven't had the time (or inclination) to build an Internet Explorer-aware implementation yet, so visitors would be advised to use a different browser.

4 4

Re: Co-operative proofreading
by Bowerbird＠aol.com 17 Mar '10

17 Mar '10

i said: > unless i'm missing something, it's just a mockup? while i don't think a mockup is very hard to do, not when compared to the programming per se, it's not like i'm mocking mockups... so if you're interested in mockups, there are more. here's one from dkretz: > http://www.pgdp.org/~dkretz/c/editpagedemo.php?projectid=projectID45c572a14… here's a page giving several from cpeel: > http://www.pgdp.org/~cpeel/prototypes/ -bowerbird

1 0

Re: New Tool "pgdiff"
by Bowerbird＠aol.com 17 Mar '10

17 Mar '10

keith said: > True enough. Yet, the arguement stands. perhaps you didn't catch my entire gist. _of_course_ one needs to allow for the possibility of editing either version, since both might be incorrect. as i pointed out, my tool (which supports jim's tool) does exactly that. > At least in my opinion. The trivial cases are > easy to handle, yet it is always the RARE cases > where tools can shine and set themselves > apart from the rest. but jim's methodology -- where his tool simply _marks_ the differences after which a person uses a reg-ex editor to actually _make_the_changes_ -- handles neither the rare cases nor the trivial well. whereas my tool-in-support-of-his handles both, equally well. a reg-ex editor, by requiring manual editing even in the "trivial" cases, handles neither trivial nor rare very well, in my opinion. my tool-in-support-of-his makes the "trivial" cases, which are by far the most common, trivial to handle, with a mere button-click or keypress. and the user only has to do manual editing in the rare case, where it simply cannot be avoided. > Actually, both methods are kind of primitive > from a Human Interface standpoint. i always appreciate it when someone analyzes my tools. so let's see what you have to say here, keith. > a better way would be having > two windows containing two or more lines > above and below the diff and marking each. a little bit of context can help elucidate the difference. too much context can bury it, depending on the display. i'd have to see exactly what you mean in order to decide. in my-tool-in-support-of-jim's tool, the change-window is a movable modal, so people can simply look back at the main window if they need to see more than 1 line of context. (i could also put multiple content lines in the top box of the change-window, if feedback indicated people wanted them.) > If you ever work with critical editions > you will understand the cavet of this method. if it impossible for you to explain in words? > The changes can then be made in a third. again, not sure what you really mean here... > All can be enhanced with colors and other neat features. you can always "enhance" anything with "other neat features". the hard part is _coming_up_ with those "other neat features". *** we should remember that my tool-in-support-of-jim's tool isn't how _i_ would do the job. i was just trying to show how to make his tool work better. i've shown how i do the job... here's how i showed diffs with gardner's book, on 23 february: > http://z-m-l.com/go/gardn/gardn-hybrid6.html that laid out the entire book, with diffs in different colors... here's a simple reworking of that file, which i just posted: > http://z-m-l.com/go/gardn/gardn-hybrid7.html this version of the file lets you click a link to see each scan, and gives you radio-buttons where you can select the correct alternative for each diff. (or choose neither if both are wrong.) this is how i would approach this task with an _online_ thrust, working in a collaborative manner. but i'd probably prefer to do it with an _offline_ app instead, since that's more efficient. -bowerbird

2 1

Re: Co-operative proofreading
by Bowerbird＠aol.com 16 Mar '10

16 Mar '10

jon said: > Interesting, and it's good to see someone do people really think lee's system is "interesting"? unless i'm missing something, it's just a mockup? it doesn't actually save the text, not that i can see. and the reg-ex cleanup doesn't really work, does it? > Interesting, and it's good to see someone > using a rich text editor for the text, rather than > expecting proofers to mess around with <i>, etc. except there is a tremendous conundrum at work... because we don't want proofers to go "presentational", do we? we want them to make structural distinctions... but with all those presentational w.y.s.i.w.y.g. buttons littering the interface, how would proofers ignore them? > I'm using a widescreen 1680x1050 monitor, and > there was still material off the bottom of the page. > This is using Google Chrome. i had similar problems in safari. camino worked fine. > It would be nice to have (the option of) a horizontal > rather than vertical layout. I used to really like > the vertical layout, but found I was more accurate > at proofing with a horizontal one. i don't understand the appeal of a horizontal interface -- way too much scrolling for me! -- but lots of people seem to prefer it. > I really prefer block paragraphs rather than > indented ones for computer-based text. but you want the display to look like the p-book, not? -bowerbird

1 0