Al, The problem with family trees is that they can get extremely wide. By going left-right I can make the tree fit into 70 columns. I want to have something that will be readable on a Kindle ultimately. My first inclination was to do the tables as images. Unfortunately, the text refers to these trees quite extensively, so the plain text version would be difficult to use. Also, when Amazon grabs a MOBI to distribute for free they tend to get the one with no images. The left-right examples in the email are as bad as any of them get. I will do top-down whenever possible. On the macron issue, the author clearly wanted to use macrons but his publisher in India could only give him circumflexes. I am using circumflexes in my text. The only change I'm making is to make the use of circumflexes more consistent. The proofreading of the original text was not very good. This is my third transcription of a book in English published in India and lousy proofreading seems to go with the territory, or it did back when the books were published. If the consensus is that left-right trees are objectionable I'll go with images throughout. Otherwise I'll stay the course and try to make the plain text version as readable as possible. James Simmlons On Sun, Dec 18, 2011 at 3:33 PM, Al Haines <ajhaines@shaw.ca> wrote:
I've heard of, but never used, the double-underscore trick. Whatever works. Just be careful it doesn't mess up alignments when you replace the two with one.
You can copy your finished UTF8 text file to a second one, do whatever search/replaces are needed to convert the Unicode characters to the DP format, then save the second file as Latin1, or convert it to Latin1 with Unitame. In cases like this, DP routinely submits two text files (UTF8, Latin1), sometimes three (UTF8, Latin1, ASCII). I've done the same myself, several times.
Not sure what you mean by top-down. If you have to rotate the physical book to read the table/tree, then render the table/tree as it appears with that rotation, just as you'd rotate an illustration.
Al
-----Original Message----- From: James Simmons [mailto:nicestep@gmail.com] Sent: Sunday, December 18, 2011 1:05 PM To: Al Haines Cc: Gutenberg Volunteers Subject: Re: [gutvol-d] Recommendations on handling family tree tables - critique wanted
Al,
The two trailing underscores are actually recommended by one of the PG guides. The idea is that when you convert to HTML you replace the two underscores with </i> and then the remaining single underscores with <i>. Of course replacing two underscores with just one in the final text file is trivial.
This book has a LOT of macrons, plus an occasional accented S. I'm going to stick with UTF-8 for this.
If I can combine the <small> tag with <pre> I will. That might give better results for the Kindle conversion.
I take it that you don't mind family trees that read from left to right instead of top down. That seems to me to be the only way to get some of these to work.
James Simmons
On Sun, Dec 18, 2011 at 2:25 PM, Al Haines <ajhaines@shaw.ca> wrote:
They look reasonable to me, but you might want to experiment with the font size. Maybe a bit smaller? I use a bit of in-line CSS for this, e.g. <pre STYLE="font-size: 10pt">, adjusting the point size as needed.
Table A - wny the double underscores after some items, e.g. "_m.__"? If nothing else, Gutcheck will probably complain mightily.
Table B - you could reduce the amount of vertical space, e.g. between Pratosha/Santosha, to a single vertical bar, instead of two bars.
A couple of things I forgot to mention initially:
Your source book has a lot of macronized characters (the short horizontal bar over them). You should consider creating two text files, one UTF8 and one Latin1. For the latter, this article (http://www.pgdp.net/wiki/Proofing_Interface_tips_and_tricks) describes DP's methods of rendering Unicode characters in a Latin1 text file. (In your HTML file, you can use either the actual Unicode characters or HTML entities to display the characters.)
Be sure to left-pad your tables/trees with a couple of blank spaces. This will (hopefully) prevent them from being errantly reformatted. See http://www.gutenberg.org/wiki/Gutenberg:Volunteers%27_FAQ#V.89._Are_th ere_any_places_where_I_should_indent_text.3F.
A trick I've used when doing trees or complex tables in a text file is to have a bunch of lines, 60-70 characters long, that contain only blank spaces. This lets me navigate around that area, inserting text and generally moving around as necessary, without suddenly finding myself back at the left margin. Trailing blanks can be easily trimmed off afterwards.
Al
-----Original Message----- From: James Simmons [mailto:nicestep@gmail.com] Sent: Sunday, December 18, 2011 8:24 AM To: Al Haines Cc: Gutenberg Volunteers Subject: Re: [gutvol-d] Recommendations on handling family tree tables - critique wanted
Al,
I've started on the pages with family trees and I have some samples that I'd like to get some feedback on before I do too many more of them. There are a LOT of family tree tables in this book.
TABLE A.
(Transcriber's Note: In the original book some tables were turned sideways. For reading as an e-book, these tables have been modified to read left to right rather than top-down).
SVÂYAMBHUVA MANU _m.__ SATARÛPA -----+---Priyavrata | +---Uttânpâda | +---_Akûti__ | _m.__ Ruchi +--- KAPILA | | +---_Devahûti__ | | _m.__ Kardama ----+--- _Kalâ__ | | _m.__ Marichi +---_Prasûti__ | _m.__ Daksha +--- _Anasuyâ__ | _m.__ Atri | +--- _Sraddhâ | _m.__ Angirasa | +--- _Havirbhu__ | _m.__ Pulastya | +--- _Gati__ | _m.__ Pulaha | +--- _Kriyâ__ | _m.__ Kratu | +--- _Khyâti__ | _m.__ Bhrigu | +--- _Arundhati__ | _m.__ Vasistha | +--- _Sauti__ _m.__ Atharvan
TABLE B.
RUCHI _m.__ ÂKÛTI --- YAJNA --------------+ (married his |--+--- Tosha sister) Dakshinâ. --+ | | +--- Pratosha | | +--- Santosha | | +--- Bhadra | | +--- Sâuti | | +--- Idâmpati | | +--- Idhma | | +--- Kavi | | +--- Vibhu | | +--- Svâhra | | +--- Sudiva | | +--- Rochana
N.B. The sons of Yajna are the Sushita Devas of the 1st. Manvantarâ.
TABLE C
Marichi _m. Kalâ__ | | --+--+-----------------+----- | | Kasyapa Pûrnimâ | +--------------+------------+-------- | | | Viraja Visvaga _Devakulyâ__ (River Ganges in subsequent incarnation).
TABLE D.
Atri _m. Anasûyâ__ | +----------------+--+------------------------+ | | | Datta Durvasas Soma (Rudra) (Brahmâ)
TABLE E.
Angirasa _m. Sraddhâ__ | ---+----------+-------+-+--------+----------+--------+----- | | | | | |
_Sinivali__ _Kuhû__ _Râkâ__ _Anumati__ Utathya Vrihaspati
TABLE F.
Pulastya _m. Havirbhu__ | -------------+------------+----------------+ | | Agastya Visvaras _m. (1) Ilavila m. (2) Kesinî__ | | Kuvera +------+------+-------------+ | | | Ravana Kumbhakarna Vibhisana
TABLE G.
Pulaha _m. Gati__ | -----+---------------+-+---------------+--- | | | Karma Sreshtha Bariyas Sahishnu
I hope the samples above are acceptable. The only other option I can think of is to treat every diagram as an illustration. Since this book is a translation of a Hindu scripture the diagrams are something we might live without in the plain text version.
The samples above should use the Courier font, and there are no tabs in my file.
Thanks,
James Simmons
On Fri, Dec 9, 2011 at 5:26 PM, Al Haines <ajhaines@shaw.ca> wrote:
James, to say you've bitten off a mouthful would be an
understatement <g>! To answer your question...
Highly structured material (trees, tables, etc) (such as that on
page 13 and elsewhere) can exceed 70 characters. For reproducing trees in HTML, the simplest method (and the one I've used), is to take the text version of the tree, copy/paste it into your HTML file, and wrap it in <pre></pre> tags.
For trees with preceding diagrams (page 13 again) insert something
like "[Illustration: Purusha]" in your text file, and the actual diagram in your HTML file.
For extremely wide trees/tables, if any (more than 100 characters,
say), split the tree into left and right halves for the text file. At the end of the file, include a transcriber's note mentioning which tables have been treated this way. For HTML, you can rejoin the halves, or leave them as in the text file, again using the <pre></pre> trick.
I notice that some tables (pages 58, 59) are split across pages.
Unless you can figure out how the two portions are joined, it's probably best to treat them as two separate tables. If you can figure how they're joined, it's acceptable to join them, but it looks as though you might end up with an extremely wide table. I'd say to keep things simple, and don't attempt a join.
From what I've seen of the book, you should be able to render all
trees/tables with alpha-numerics.
Good luck!
Al
-----Original Message----- From: gutvol-d-bounces@lists.pglaf.org
Sent: Thursday, December 08, 2011 11:56 AM To: Gutenberg Volunteers Subject: [gutvol-d] Recommendations on handling family tree tables
I am working on transcribing this book:
http://www.archive.org/details/studyofbhagavata00benaiala
This is a translation into English of an important Hindu scripture, originally in Sanskrit. Now that I have committed myself to
[mailto:gutvol-d-bounces@lists.pglaf.org] On Behalf Of James Simmons transcribing it I find that it has many, many family tree tables in it. These are not part of the scripture itself, but were added to clear up who begat who. (All mythical characters). Some of these stretch out so wide that the page is rotated in the original. Others could easily be duplicated with ASCII characters in 80 columns or less.
I've looked at some PG titles with geneologies like this one about
Bach:
http://www.gutenberg.org/files/35041/35041-h/35041-h.html#toc49
In this one the family tree tables are treated as illustrations.
That would certainly be a simple way to deal with them, but again about half of the diagrams could be done with ASCII text. It would be painful, but it could be done. So do I do half and half, or try and do all text, or just do illustrations all the way?
Thoughts?
James Simmons