Re: Problem in file retrieval
David Widger wrote:
A new quirk in access to our PG files and not a happy one. A main advantage of the new directory system is direct access to a file when its url is entered and until this morning that has been the case. I now find that when a direct link is entered one is _not_ taken directly to the file but rather to the PG catalog. I can only imagine this an unintended side effect of some program change.
For example:
http://www.gutenberg.org/dirs/3/1/7/3176/3176-h/3176-h.htm
has for the past year taken users directly to Twain's "Innocents Abroad", now it takes them to the bibrec where they have to choose again from a long confusing list of files.
I cannot reproduce this. Clicking the url in the mail or copy and pasting the url into a browser window gives me the file and not the bibrec. This is related to a recent change in the site programming which I am testing. I redirect all deep links to files from external pages to the bibrec page. This has some advantages: - the link won't go dead when we REPost the file - the user gets a choice of formats - the user doesn't get an outdated edition - our books get a better google ranking - the user gets to see our site. On the downside, sometimes you have to click some more to get the file you want. How does this work? _If_ the browser provides a referrer _and_ the referrer is not from our site, the user gets redirected. When opening the file from a bookmark or entering the url in the location bar, the browser should send _no_ referrer, and the user will not be redirected. When clicking on a link on a page, the browser should send a referrer (can be turned off). The list of "good" referrers, which will never be redirected, contains www.gutenberg.org only, but can be expanded to contain any "Independent Gutenberg Search" site. How did you access the file ? I'd like to try this approach for a few weeks and see if I hit any "hard" problems with it. -- Marcello Perathoner webmaster@gutenberg.org
Marcello Perathoner wrote:
When opening the file from a bookmark or entering the url in the location bar, the browser should send _no_ referrer, and the user will not be redirected.
Maybe you could add known on-line mail readers and things with "webmail" or "neomail" in the referrer string to the exception list.... I sometimes read my email from a website. I use the same technique to stop hotlinking images of my website -- did you also achieve that effect? Jeroen
Jeroen Hellingman (Mailing List Account) wrote:
Maybe you could add known on-line mail readers and things with "webmail" or "neomail" in the referrer string to the exception list.... I sometimes read my email from a website. I use the same technique to stop hotlinking images of my website -- did you also achieve that effect?
I turned off image inlining a while ago. Nobody noticed so far. -- Marcello Perathoner webmaster@gutenberg.org
Marcello Perathoner <marcello@perathoner.de> wrote:
David Widger wrote:
A new quirk in access to our PG files and not a happy one. A main advantage of the new directory system is direct access to a file when its url is entered and until this morning that has been the case. I now find that when a direct link is entered one is _not_ taken directly to the file but rather to the PG catalog. I can only imagine this an unintended side effect of some program change.
For example:
http://www.gutenberg.org/dirs/3/1/7/3176/3176-h/3176-h.htm
has for the past year taken users directly to Twain's "Innocents Abroad", now it takes them to the bibrec where they have to choose again from a long confusing list of files.
I cannot reproduce this. Clicking the url in the mail or copy and pasting the url into a browser window gives me the file and not the bibrec.
This is related to a recent change in the site programming which I am testing. I redirect all deep links to files from external pages to the bibrec page. This has some advantages:
- the link won't go dead when we REPost the file - the user gets a choice of formats - the user doesn't get an outdated edition - our books get a better google ranking - the user gets to see our site.
On the downside, sometimes you have to click some more to get the file you want.
How does this work? _If_ the browser provides a referrer _and_ the referrer is not from our site, the user gets redirected.
When opening the file from a bookmark or entering the url in the location bar, the browser should send _no_ referrer, and the user will not be redirected.
When clicking on a link on a page, the browser should send a referrer (can be turned off). The list of "good" referrers, which will never be redirected, contains www.gutenberg.org only, but can be expanded to contain any "Independent Gutenberg Search" site.
Search engine bots don't usually give a referrer and so they will not see any change. -- Philip Baker
Philip Baker wrote:
Search engine bots don't usually give a referrer and so they will not see any change.
Thats exactly what I want. Otherwise they would index the bibrec page instead of the book. The user clicking on the search results will get to the bibrec page, where she can select a file format and compression. -- Marcello Perathoner webmaster@gutenberg.org
In article <41E379E4.5090603@perathoner.de>, Marcello Perathoner <marcello@perathoner.de> writes
Philip Baker wrote:
Search engine bots don't usually give a referrer and so they will not see any change.
Thats exactly what I want. Otherwise they would index the bibrec page instead of the book.
The user clicking on the search results will get to the bibrec page, where she can select a file format and compression.
You listed getting a better Google ranking as one of the aims of the change. -- Philip Baker
participants (3)
-
Jeroen Hellingman (Mailing List Account) -
Marcello Perathoner -
Philip Baker