Proxy redirect (officially a VPN) and Google Books page images

reddykj

I often work logged in to my university's VPN (UC Davis) for which I use Zotero's proxy redirect. However, I run into trouble when trying to load Google Books. I can only view the first page image that results from my search; any scrolling results in blank pages with the gray "Loading..." message in upper and lower left corners. I set an exception for books.google.com but because Google's search runs through google.com the exception doesn't solve the issue; the actual page books.google.com comes through without the VPN but search results, even with terms entered on the Books home page, come back as a google.com page through the VPN.

Anyone else have this issue? Any ideas about how to craft a workaround or how to tweak the exception to maintain the redirect for all Google searches except Books?

Google Books were unable to provide a solution.

(I use Zotero 3.03 in Firefox 11.0 on a Windows 7 Professional computer.)

adamsmith

why do you need google.com to run through the proxy? Wouldn't it be enough to have that work for scholar.google.com?

aurimas

I'm also a little confused as to why you need to use Zotero's proxy redirection when you're already connected to your university's VPN.

Once you're connected to the VPN all traffic is going through your school's routers (that's not necessarily true, but it is by default), so Zotero has no option but to use the school as a proxy.

adamsmith

Some Universities call their off-site proxying systems VPN - I don't understand why*, but it's not unusual. As I understand the OP, that's the case for UC Davis.

* I actually think it's a rather sad statement on part of the IT dept. of an academic institution to just assume that they needn't/shouldn't bother their users with the difference between the two.

reddykj

Ah, perhaps that's it. The university also offers what they call a proxy, but they strongly recommend the VPN (or "VPN"?!) for speed and security.

In practice, our VPN desn't send all traffic through the university's routers, only traffic initiated directly from the library's website after VPN login.

For example, with the redirect off, I may have two tabs open, one open to Google and the other open to Google as linked on the VPN list of resources. In the first tab, results I obtain through the browser's search bar or through google.com link me to any resources as a restricted user. In the second tab, results I obtain through a search on the VPN Google page link me as a university user. With the redirect on, I don't have the issue -- any search considers me a university user, whether done via the browser's search bar (which I use very often), google.com, or my bookmarked VPN version of Google.

Perhaps rather than my initial question I should ask whether there are better alternatives for this mode of using the redirect?

adamsmith

an actual VPN runs outside of your browser - it would automatically affect all network traffic - including not just http/https, but also chat, skype, pop/imap etc. and note affect the URLs as displayed in the browser - as a consequence it wouldn't interact with Zotero's proxy redirection at all. Universities often use Cisco's VPN client for that, but by now most operating systems natively support VPN for internet connection.
It doesn't sound like you're using something like that, but apparently some hybrid version between proxy and VPN.

My question about google remains - is there any reason you need google.com to run through the proxy? Couldn't you set this just for google scholar?
Also, are you certain that the proxy/vpn is the actual cause for your problems with google books?

aurimas

adamsmith beat me to it, but I was in the middle of this post, so I'll post it anyway.

What you are describing is a proxy and not a VPN.

When you connect to the VPN (and you have not changed the default
Windows settings) a connection is established to your school that acts like your internet connection. All the internet traffic is redirected transparently through the school's computers.

Example (requesting google.com page):

Without a VPN connection
Your computer <--> Your ISP <--> Google.com

With a VPN connection
Your computer <--> Your ISP <--> Your School <--> Google.com

In either case, the url of the page is google.com, so you can't tell from the URL if you're using VPN or not. Google.com, however is able to tell that your connection is coming from Your School rather than from Your computer.

Note that by default ALL traffic is redirected this way through a VPN.

And adamsmith beat me to the punch line.

@adamsmith: google books are served from books.google.com I thought. That was kind of an issue when I was tweaking the google scholar translator.

reddykj

I don't know technically what the service is, but it's browser- and page-specific (described in the FAQ here). Regardless, the issue remains -- and the inability to view more than the original book page occurs only when viewing the page through the VPN. While the books themselves are served from books.google.com, the search results are not, they're served from google.com (hence the exception doesn't help).

For example, I get different levels of success if I use the same browser to navigate to books.google.com twice and search for "heffalump," once with the redirect (and the books.google.com exception) and once without the redirect:

With =

Starting Page:
books.google.com

Search Results:
vpn.lib.ucdavis.edu/,DanaInfo=www.google.com+search?q=heffalump...

Preview (of which only the initial page image will load):
vpn.lib.ucdavis.edu/,DanaInfo=books.google.com+books?id=...

Without =

Starting Page:
books.google.com

Search Results:
https://www.google.com/search?q=heffalump...

Preview (of which subsequent pages will load with scrolling):
http://books.google.com/books?id=...

It's not a speed issue, either, as far as I can tell. I can leave the VPN Preview open interminably and the subsequent page images don't load.

So again, the question is -- any other way to establish an exception that can handle books.google.com search results that come through google.com? Perhaps it's not possible, but thanks to folks for trying to help!

adamsmith

Probably not, no (I hate that ,DanaInfo implementation...) -
But again:
"My question about google remains - is there any reason you need google.com to run through the proxy? Couldn't you set this just for google scholar?"

reddykj

Sorry, tried to explain that already. It's because run through the VPN I can access a lot of resources (especially academic journals) that I cannot fully access via Google Scholar, even if they turn up in Scholar's search results. Linked through Scholar I might be able to see abstracts or first pages; through the VPN I can see and download entire articles.

adamsmith

right, but google scholar is scholar.google.com, not google.com - so if you only need the proxy for google scholar, just set it for google scholar, not for google.com generally.

reddykj

In another conversation I'm having about this, with my connection at Google Books, I realized that part of what surprises me is that the exception seems to apply only if I directly enter an address (eg, books.google.com). If the exception applied no matter whether a page were reached via entering an address or following a link, then this wouldn't be an issue at all -- results could come through the VPN, and the link to the preview via books.google.com would fall under the exception again. Alas, the exception doesn't seem to work this way.

adamsmith

That's likely a setting of your institution's proxy, not of the Zotero redirect feature.
where are you setting the exception btw.? The way to go is to fine-tune the sites where the proxy kicks in.

reddykj

I'm just using Zotero's setting for "Enable proxy redirection." I'd fine tune if I could but can't see a way to add anything besides a domain name. It'd be awesome if I could figure how to tinker with it to the level of a domain plus a wildcard in the full address, like google.com + *bks* would catch the book search results that come through google.com.

adamsmith

you can only specify hostnames, but you configure those for where the proxy kicks in, not for exceptions - where are you specifying that exception? So the way this should look imho is that you just have scholar.google.com in your set of hostnames to be proxied. I don't understand why that doesn't work.

reddykj

The disadvantage of setting the redirection just for scholar.google.com is that then I have to conduct two Google searches for the same topic, following certain types of links (eg articles) from Scholar and certain types of links (eg books) from Google. At this point, it's easier for me to just conduct one overall search and then open any specific book in a separate browser once I know it's going to be helpful.

adamsmith

I'm pretty sure this has nothing to do with Zotero.
Try this: Disable the Zotero proxy-redirect feature.
Log-in to google scholar with your institutional proxy - I don't know if you can do that through the library or if you need to do manually go to
vpn.lib.ucdavis.edu/,DanaInfo=scholar.google.com
Note that all or most of your outgoing URLs are going to be proxied/VPNd even without Zotero proxy re-direction turned on --> i.e. this has nothing to do with Zotero.

In a well set-up proxy/google scholar implementation, only URLs where the proxy is useful are actually proxied - my institution e.g. won't proxy google book links. If UC Davis does, get in touch with your IT department - since proxies are usually run by the library and not general IT, you actually have a decent chance of finding someone who'll want to help you.

reddykj

Have also contacted my university, as you are right that even without the redirect, if I navigate to Google Scholar via the VPN and then follow a Google Books link, it leads to a ,DanaInfo address -- completely unnecessary. I envy you!

I've realized all along it's not necessarily a Zotero issue, was just hoping Zotero could provide a solution or workaround. Again, thanks for trying to help.

adamsmith

yeah, sorry, this is your U. entirely - Zotero can't help here - it can't prevent your proxy from kicking in.
As a suggestion - do track down the librarian in charge of this. Librarians are notoriously helpful, even the IT-ones (and yes, I have a pretty low opinion of University IT departments...).