I received some complaints about the fact that some links in my books were dead by the time they were printed.
For the next book I am working on, I have proposed to my editor to set up a website to keep track of all references mentioned.
By using unique short ascii references throughout the book, it’s easy to provide a simple redirect service to the target URL, and to fix it when it changes (just by setting up a mail alert if your redirect reaches a 404).
For example, if I am referring to mod_wsgi in my book, I can write this reference: #mod_wsgi, and provide a redirection to http://code.google.com/p/modwsgi into my website, through a unique, mnemotechnic permanent URL : http://ziade.org/urls/mod_wsgi.
This small service, à la Tiny URL is not a burden for the reader imho : he is using his computer anyway when he visits an URL mentioned in a book.
It’s a simple idea I am sure a lot of people have thaught about before, but I fail to see it applied in the books I am buying these days. Is there any good reason I fail to see ?
There was a good blog post on the great impending linkrot from url shorteners:
http://joshua.schachter.org/2009/04/on-url-shorteners.html
I guess what you are proposing makes sense as long as the publisher (or author) keep the site up and running for as long as the book is about.
-Matt
Comment by Matt Hamilton — April 19, 2009 @ 6:29 pm |
Sounds like a very good idea to me – wish I’d thought of it for IronPython in Action!
Comment by Michael Foord — April 19, 2009 @ 6:49 pm |
Great idea! A generic service, provided by the publisher, could have URLs like http://publisher/bookname/reference
Comment by Pekka Klärck — April 19, 2009 @ 9:32 pm |
But then you have a single point of failure… I’d prefer it if there was also a table printed at the back of the book mapping the references to the full URLs at time of publishing.
Comment by Floris Bruynooghe — April 19, 2009 @ 9:45 pm |
For mod_wsgi site you should use:
http://www.modwsgi.org
It is already a redirect to where ever is the current place and given it is a dedicated DNS name, not likely to be lost unless the bill isn’t paid.
Comment by Graham Dumpleton — April 19, 2009 @ 9:51 pm |
As long as you include a paper bibliography with the links in full.
Comment by Peter — April 19, 2009 @ 11:03 pm |
If I was to do it I’d do it from the site that *I* set up for the book – and not rely on the publisher.
Comment by Michael Foord — April 19, 2009 @ 11:16 pm |
You might want to consider using github or something similar instead, that way people could fork the list of links and publish corrections long after you and your publisher has moved on to greener pastures (or even the afterlife). Although the best thing you could do is link to internet archive as ordinary webpages especially personal ones go down, it has become a serious issue in Academia as things that get referenced tend to only survive a few years.
Comment by Celc — April 19, 2009 @ 11:35 pm |
This is a very good idea, and I don’t dislike the concept of publishers setting up such a service on behalf of the readers of their books. It really makes me wish I had thought of this for my ‘Ruby on Rails for Microsoft Developers’ book. Oh well, I’ll see what I can do for the next one.
Comment by Antonio Cangiano — April 20, 2009 @ 2:39 am |
Very good idea.
Comment by gsempe — April 20, 2009 @ 4:31 am |
I would definitely second the recommendation that your bibliography or an appendix at the end of the book must contain a list of the canonical urls at the time of printing. Alternatively, you could even use the canonical urls in the text and have a page or two at the front suggesting that people can get up to date errata and the an updated list of referenced sites from your book-specific web site.
If I were to page through a book that contained nothing but tinyurls and which was not printed in the last year I would put the book down and find another one. As a potential purchaser I don’t have the time to whip out my phone to make sure that the book’s web site is still active and available. I have been burned once or twice in the past by book web sites that either disappear completely or suffer from slow bit rot, with the canonical urls and the name of the package or page I can at least dig through google or the wayback machine to find the right data…
Comment by evgen — April 20, 2009 @ 4:41 am |
Real links with corrections included in the Errata published on the book publishers site. Thats the purpose of errata already.
You wouldn’t replace postal addresses with your own and promise to forward letters to the right location. You’d simply publish errata with correct postal addresses.
Perhaps you shouldn’t be citing such transitory sources to begin with?
Just my $0.02, in practice your solution may be friendlier, but it just rubs me the wrong way.
Comment by schmichael — April 20, 2009 @ 6:35 am |
Thanks a lot
Comment by muziqcantiq — April 20, 2009 @ 7:05 am |
You could use whatever redirection scheme you like. If I were trying to chase a reference from a paper book, I would type “mod_wsgi” into Google search anyway. It would involve the same number of keystrokes or fewer than any other scheme and it is be more stable in the face of disappearance or migration of the target site or the redirection site.
Comment by Alastair — April 20, 2009 @ 7:50 am |
[...] the great it would be if links in texts could be updated – someone of course is already doing this! I am intrigued! Possibly related posts: (automatically generated)Perception of Memories?Dutch National identity [...]
Pingback by Update to SL Identity Idea « Theories of Texts and Technology — April 20, 2009 @ 9:45 am |
I won’t reply to each one, but many thanks for all your comments !
More comments here : http://news.ycombinator.com/item?id=569887
I am now convinced that :
- it’s good to have short ids instead of urls in a book (ala LaTeX references)
- it’s good to have an index at the end of the book with the id pointing to a short URL *and*
the target URL.
- the website that manages the URL has to be maintained for a long period (wich is not a problem imho,
I have been running the same websites for my books since the first one was out 4 years ago)
Comment by Tarek Ziadé — April 20, 2009 @ 10:12 am |
@schmichael
Aren’t errata for corrections and not for canonical references? Having an index in the book with the *real* URLs seems like a good idea, but the errata seems like the wrong place to put it.
Comment by Michael Foord — April 20, 2009 @ 2:43 pm |
Marty Alchin’s Apress book “Pro Django” uses footnotes for the URL’s, and each URL points to the domain name set up just for the book at prodjango.com. He doesn’t use “tiny” URL’s but rather slugs which makes them reasonably short and, more importantly, easily typed.
Comment by Van Gale — April 20, 2009 @ 6:36 pm |
[...] URLs in books I received some complaints about the fact that some links in my books were dead by the time they were printed. For the [...] [...]
Pingback by Top Posts « WordPress.com — April 22, 2009 @ 12:12 am |
I was surprised that some russian guy just got a patent on this stuff:
http://linkod.com/en.html
http://www.ideaconnection.com/technology-for-sale/8973-METHOD-OF-SIMPLIFICATION-OF-ACCESS-TO-INTERNET-RESOUR.html?ref=nl042809
Comment by Ronan Amicel — April 28, 2009 @ 12:52 pm |
This guy – I am.
It not absolutely that that is offered by the author.
Comment by Vladimir — May 3, 2009 @ 11:32 am |
LaTeX Web Companion did this in 1999:
http://www.amazon.com/LaTeX-Web-Companion-Integrating-Typesetting/dp/0201433117
http://www.tug.org/docs/books/lwc/lwc-url.html
Comment by Jouni — April 28, 2009 @ 7:01 pm |
[...] través de un tweet de José Frechín llegaba a un artículo muy interesante, URLs in books « Fetchez le Python, en el que el autor del libro, se plantea opciones para mejorar la referencia a URLs en libros. [...]
Pingback by URLs en libros. Soluciones y la solución | factor Kippel — April 30, 2009 @ 4:50 pm |