Re: [gutvol-d] interesting stat from the news letter
DP produces about 200+ new books a month. Unfortunately, the proofers at DP, finish about 250 books a month. Which means we have an ungodly backlog of texts that need to be post-processed (over 450 books right now). Our proofing output has scaled up from last year, but our post-processing has not been able to keep up. There are plans for ways to improve the bottleneck. Unfortunately, developers to implement those ideas are another bottleneck. As big_bill at DP always says, though ... Those books aren't going anywhere and we will get to them eventually. :) JHutch ----- Original Message ----- From: Aaron Cannon <cannona@fireantproductions.com> Date: Wed, 15 Sep 2004 12:13:47 -0500 To: gutvol-d@lists.pglaf.org Subject: [gutvol-d] interesting stat from the news letter
I usually skim the news letter to see if there is anything new, and I noticed this little stat: 352 Average Per Month in 2004 355 Average Per Month in 2003.
Anyone have any theories as to why we are doing fewer books per month on average this year? Not complaining, just a little curious.
Sorry if this is old news.
Sincerely Aaron Cannon
-- E-mail: cannona@fireantproductions.com Skype: cannona MSN Messenger: cannona@hotmail.com (Do not send E-mail to the hotmail address.)
_______________________________________________ gutvol-d mailing list gutvol-d@lists.pglaf.org http://lists.pglaf.org/listinfo.cgi/gutvol-d
I'm one of those people guilty of big backlogs, with 13 books in various stages of completion in Post Processing after going through DP. I always do SGML, XML, and HTML, have heavily illustrated works with lots of tables, and these take a lot of work, no matter what the proofreaders do. A few times I've just left out the most horrendous tables, to type them personally, or even drop them from the work altogether. Then, adding ASCII versions also adds to the delay, as I can generate HTML from the SGML automatically, but ASCII simply is too hard to automate, which means redoing the tables again. An average novel can be PP-ed in a few hours, but those scientific works take many hours. My longest running project (not through DP) is Alberuni's India, with long citations in Greek, Persian, Sanskrit, etc., often in the original scripts (five different scripts in this book). Spread out over five years, hundreds of hours have gone into it, and it is not yet done. Jeroen. Joshua Hutchinson wrote:
DP produces about 200+ new books a month. Unfortunately, the proofers at DP, finish about 250 books a month. Which means we have an ungodly backlog of texts that need to be post-processed (over 450 books right now). Our proofing output has scaled up from last year, but our post-processing has not been able to keep up. There are plans for ways to improve the bottleneck. Unfortunately, developers to implement those ideas are another bottleneck.
As big_bill at DP always says, though ... Those books aren't going anywhere and we will get to them eventually. :)
JHutch
Jeroen Hellingman <jeroen@bohol.ph> writes:
I'm one of those people guilty of big backlogs, with 13 books in various stages of completion in Post Processing after going through DP.
Thanks for your commitment!
I always do SGML, XML, and HTML, have heavily illustrated works with lots of tables, and these take a lot of work, no matter what the proofreaders do. A few times I've just left out the most horrendous tables, to type them personally, or even drop them from the work altogether.
Things like these take many resources ;-( Nevertheless, droping contents alltogetther is not appropriate - better include them asis, even if they would look very wrong. Thus the reader can see this material and lend a helping hand...
Then, adding ASCII versions also adds to the delay, as I can generate HTML from the SGML automatically, but ASCII simply is too hard to automate, which means redoing the tables again.
Heresy alert: As long as the book is readable with lynx, w3m, or another text browser, I wouldn't spend time on ASCII versions. Better consider to create PDFs from the SGML/XML source files. BTW, pdftotext from the xpdf suite is worth a try!
An average novel can be PP-ed in a few hours, but those scientific works take many hours. My longest running project (not through DP) is Alberuni's India, with long citations in Greek, Persian, Sanskrit, etc., often in the original scripts (five different scripts in this book). Spread out over five years, hundreds of hours have gone into it, and it is not yet done.
Yes, those books are not suitable for DP. Also other books could be proofread better and/or faster if one would loudly read the original text (plus punctions and stuff like that) and a partner proofreader would listen to the reader while checking the OCR results. (Of course, using a "Diktiergerät" (dictating machine and an appropriate player with pedals for control) you can go this way without a partner. -- | ,__o | _-\_<, http://www.gnu.franken.de/ke/ | (*)/'(*)
On Sat, 18 Sep 2004, Karl Eichwalder wrote:
Yes, those books are not suitable for DP. Also other books could be proofread better and/or faster if one would loudly read the original text (plus punctions and stuff like that) and a partner proofreader would listen to the reader while checking the OCR results. (Of course, using a "Diktiergerät" (dictating machine and an appropriate player with pedals for control) you can go this way without a partner.
Not being able to find a willing partner, I have occasionally before simulated this effect by reading out loud sections of a text (including my own shorthand conventions for punctuation), recording it on a small handheld voice recorder, and then playing it back while following along with the digitised version. I like the idea of the machine you mention... Andrew
Good to know. I didn't realize that there was such a backlog. Good luck with that, especially with regards to automating the process. I would love to help, but I really have my hands full with the CD/DVD project. :) Sincerely Aaron Cannon At 01:40 PM 9/15/2004, you wrote:
DP produces about 200+ new books a month. Unfortunately, the proofers at DP, finish about 250 books a month. Which means we have an ungodly backlog of texts that need to be post-processed (over 450 books right now). Our proofing output has scaled up from last year, but our post-processing has not been able to keep up. There are plans for ways to improve the bottleneck. Unfortunately, developers to implement those ideas are another bottleneck.
As big_bill at DP always says, though ... Those books aren't going anywhere and we will get to them eventually. :)
JHutch
----- Original Message ----- From: Aaron Cannon <cannona@fireantproductions.com> Date: Wed, 15 Sep 2004 12:13:47 -0500 To: gutvol-d@lists.pglaf.org Subject: [gutvol-d] interesting stat from the news letter
I usually skim the news letter to see if there is anything new, and I noticed this little stat: 352 Average Per Month in 2004 355 Average Per Month in 2003.
Anyone have any theories as to why we are doing fewer books per month on average this year? Not complaining, just a little curious.
Sorry if this is old news.
Sincerely Aaron Cannon
-- E-mail: cannona@fireantproductions.com Skype: cannona MSN Messenger: cannona@hotmail.com (Do not send E-mail to the hotmail address.)
_______________________________________________ gutvol-d mailing list gutvol-d@lists.pglaf.org http://lists.pglaf.org/listinfo.cgi/gutvol-d
_______________________________________________ gutvol-d mailing list gutvol-d@lists.pglaf.org http://lists.pglaf.org/listinfo.cgi/gutvol-d
-- E-mail: cannona@fireantproductions.com Skype: cannona MSN Messenger: cannona@hotmail.com (Do not send E-mail to the hotmail address.)
On Wed, 15 Sep 2004, Joshua Hutchinson wrote:
DP produces about 200+ new books a month. Unfortunately, the proofers at DP, finish about 250 books a month. Which means we have an ungodly backlog of texts that need to be post-processed (over 450 books right now).
Interesting. I thought it was more than that. Right now, the list of "silver star" etexts (those which have finished first and second round, but have not yet finished post-proofing) has 1,890 titles. (With the oldest apparently going back to late 2002.) Andrew
participants (5)
-
Aaron Cannon -
Andrew Sly -
Jeroen Hellingman -
Joshua Hutchinson -
Karl Eichwalder