I think this is a browser/webserver issue, not something RH really has any control over.
I believe you can turn off caching on the web server, so perhaps talk to your server guys about that.
There are also meta nocache tags you could try, but there seems some doubt about how reliable they are.
There's a bit of a discussion here, but I admit I get somewhat lost.
I was redirected here as we have a question regarding how to set up our online help in the cloud so that user's don't have to log in. How did you guys set up your help icons to redirect users? Did you use a token or cookie to verify they were logged in? How do you manage the user names/passwords? Here's our post:
While reading your post, I noted you have a similar browser cache issue we encountered. We created a master page and put the meta cache tags in the HTML (we also assigned the CSS to it to spare ourselves from attaching it to each topic). We then assign the master page to each page in the project when we compile. Seems to work well.
Huge thanks for your time and help!
Is this online help for a software application? Is the help published to a cloud server then accessed from the application?
Yes, this is help for three different applications. One is a PowerBuilder application and the other two are web applications. The applications will authenticate the users. We don't want to get into maintaining usernames/passwords esp since our software is deployed in campus environments where turnover can be high. We are setting the applications' help icons up so that when the user clicks them, they will access the cloud server where our help files will live. Our developers have been looking at using link redirects that include a cookie or toke, certificate-based authentication, or encrypted URL query string values that include our own key. We were looking to see how others were handling it and if anyone had any recommendations!
Unless I have misunderstood, only someone who has previously accessed the
help will be able to access it again as I believe the cache is per user. On
the basis they were authenticated once, is there an issue here?
Do you publish directly to the cloud server?
No issue. Just looking for insights or guidance based on our plan.
The current plan is to publish the help to an IIS cloud server bringing up static RoboHelp pages. We'll provide authentication tokens that will be submitted along with any help page request. Those in turn will provide users with a cookie (if the token validates). If the user browses to a bookmarked page and the session associated with the cookie has expired, we will redirect them back to one of the two web applications so they can log back in. Once logged back in, they'll be redirected back to the original help page they requested.
Our devs are curious if this is the similar to the process being described by amandada, the original poster? We were also looking for additional details they could provide that might give us insights into how to make this a secure and robust approach.