Welcome to WebmasterWorld Guest from 126.96.36.199
Someone told me that when they go to "search the site" and type in a search, they get error messages that kick them out.
When I use IE, it takes me to the search results page, but there is no information listed. This makes three different computers in three different locations that are having trouble with the custom search engine application using Internet Explorer. The problem does not occur when using Firefox.
[edited by: tedster at 7:53 pm (utc) on May 14, 2009]
[edit reason] de-link the example [/edit]
The # after a "page name" (the default "home" page in this case) means "go to bookmark named..." where the bookmark is defined by <a name=...>. Nothing after the # is used except for this purpose. This is a browser process, not a server one. So if you are already on the google home page this would try to take you to the (non-existent) bookmark on that page instead of submit the request to google, surely?
There used to be problems with some browsers interpreting bookmarks that included both # and querystrings - you had to get them in the right order. If that is still a (partial?) problem then it may account for the IE/FF discrepancy. Or possibly FF has added an over-ride to the # if it's google? You never know! :)
We saw this happening a couple of months back, but now looks like they've expanded this quite a bit starting yesterday/today.
Looks like the client side stuff on your IE is either blocking it or it's not working on your IE browser because of some sort of error.
I don't know the purpose of this hash mark technique, but I can speculate: Maybe to stop automated query generators that check rankings? Maybe some sort of client side browser identification to do this also?
Problem is when we saw it last, our logs just say www.google.com as a referrer so we don't know how we're found when users go to our sites from these search results. Maybe analytics is now getting these queries, I won't know the answer to this until I have a couple of days of data and see if the google analytics will only get the keyword phrase from their queries as opposed to other analytics programs.
That would push more sites to use google's analytics also.
Thoughts on my diagnosis theory and reason theories anyone?
The # after a "page name" (the default "home" page in this case) means "go to bookmark named
I don't know the purpose of this hash mark technique
It's also an Ajax technique today, and not just a page fragment identifier. For some background, see this Microsoft page [msdn.microsoft.com]
One thing that did ring a warning bell is the statement "The potential problem is that some Web sites manually manipulate the hash of the URL for their own purposes." That suggests that the original "bookmark" anchor technique is a naughty aberration rather than something that millions of web sites have been using for years. Plus, as I said before, it's built into the browser to obey "bookmarking".
In particular, though, if scripting is turned off in a browser the # WILL, I would have expected, behave as it (once upon a time?) should have done and never leave the page. If that is so then google will become totally useless to me.
That's if we take "hash" to mean the # symbol. Hashing, to me, is an entirely different concept related to encoding and encrypting.
Quite apart from which, as I understand it, it's MSIE which is failing, not FF.