More on Google Print
Lachlan Hunt had the best suggestion yet for defeating Google Print DRM with Firefox: put the line
img[src="images/cleardot.gif"] { visibility: hidden; }
into chrome/userContent.css, and then (following somebody else's suggestion) Edit/Preferences/Web features/Advanced JavaScript Options/Allow Scripts To: Disable or replace context menus (no). That is the end of Google Print's control over your right mouse click feature, at least until Google changes the name of cleardot.gif.
![[image of Advanced JavaScript options]](/images/java.png)
A lot of people read my earlier piece on Google Print through links from slashdot and Boing Boing, and a few sent me comments on it. They weren't necessarily as convenient in their suggestions as Lachlan Hunt's solution. One reader suggested that a proxy called Proxomitron can perform regular expression substitutions in retrieved pages, and even provided a plausible regular expression to "provide thumbnails of background images" in all retrieved pages, thus allowing browsers which provide limited access to background images to make use of direct links. It sounds like this technique was earlier developed by people who were frustrated with recent tricks used by IMDB to prevent people from saving images.
The peculiar tragedy of Proxomitron, as I understand it, is that it was Windows freeware maintained as a public service by a man who abandoned the project (last year) and then died (this year). The result of all this is that nobody else will readily be able to maintain or enhance the project. (They can't even readily fix bugs.) I don't know any reason to believe that the author meant for his project to die with him. I would be happy to hear from a Proxomitron user that the story has a more optimistic conclusion...
Google Print seems to return a consistent, stable URL in response to a given book search result (even if you search some time later with a different browser). It also seems to return a consistent, stable "sig" parameter for a given page of a given book, regardless of which particular search was used to obtain it. That means that URLs to book images can be bookmarked and can also be given to other people. Browser cookies are not relevant; I was able to delete my Google cookie and even access URLs with lynx.
Contrast this URL from an actual search with an edited version of the same.
Gervase Markham has some analysis of the parameters in a Google Print URL, including a similar conclusion about "sig".
You can download Google Print precious images with Firefox or lynx, but not with wget, as someone noted -- the user agent makes it give a 403. The same is true of curl, which implies that Google Print has a list of banned user agents. However, wget --agent="Crystal Geyser Natural Alpine Spring Water" worked just fine, as did wget --agent="Opening Brief of Petitioners". (Perhaps the user-agent discrimination is aimed at preventing people from using wget -r, although I don't know if wget makes any attempt to parse CSS. It doesn't seem to; even --page-requisites doesn't appear to try to parse CSS, merely to download external CSS files.)
Google's URL structure seems to make the virtual host name merely cosmetic; all URLs at print.google.com start with "http://print.google.com/print?" -- something I noticed after reading Google's robots.txt and noticing that it was the same as
