Re: [WWdevel] webwork page caching

[P G. L. <gl...@um...>]

Hi Sam,

The biggest reasons I was thinking this could be a problem were:
  - On gateway tests I maintain a timer on the page indicating how much
    time is left.  At the moment I'm doing this as a countdown from the
    amount of time left when the page is first loaded---going back to this
    page therefore results in the student thinking that s/he has more time
    than is the case.  I can work around this if need be, but I think it's
    really part and parcel of the second comment, viz.,

  - In general, when going from page to page working a problem the previous
    page's information won't reflect the current state of a problem.  This
    is particularly relevant when there are a fixed number of attempts at a
    problem, so that going 'back' makes it appear that one has more
    attempts than is actually the case.

I agree that being able to go back through the browser's history is very 
useful, however.  One option would be to no-cache student pages, or for 
students only.

I'm not entirely familiar with where WeBWorK is using GET vs POST, so I 
don't know where that becomes an issue.

Gavin

On Wed, 2 Nov 2005 at 23:35 Sam Hathaway wrote:

> On Nov 2, 2005, at 11:34, P Gavin LaRose wrote:
>
>> Hi all,
>> 
>> Watching some of our gateway testing, the following occurred to me: it 
>> would be good to avoid students being able to use the browser's "back" 
>> button to navigate.  And then I thought that this should be a desirable 
>> state of affairs in general.  Because WeBWorK pages contain state data, we 
>> don't really want people going back to previous pages.  I use this 
>> occasionally when I'm working as a course administrator, but on student 
>> pages I think it's something we don't want happening.
>> 
>> Would it make sense to add a 'no-cache' command to the header of all 
>> WeBWorK pages, so that it would force the browser to go back to the server 
>> to get each new page?
>
> Is the problem just that there could be out-of-date information on cached 
> pages? (Like the case where a user attempts a problem for the first time and 
> then hits back to return to the problem list and it tells her that she hasn't 
> attempted it?) Or are there more dangerous cases? (If there are those should 
> be removed regardless of whether we add no-cache.)
>
> I actually like being able to quickly pop back to the previous page and then 
> forward to the current one one without reloading.
>
> If we added no-cache we would also want to be more careful about when we use 
> POST versus GET. Good browsers warn you when a POST request needs to be 
> resubmitted (like in the case of no-cache) and this is annoying when the 
> re-POST isn't actually dangerous.
> -sam

-- 
P. Gavin LaRose, Ph.D.  Program Manager (Instructional Tech.)
Math Dept., University of Michigan
gl...@um...                        "There's no use in trying," [Alice]
734.764.6454                             said. "One Can't believe impossible
http://www.math.lsa.umich.edu/~glarose/  things." "I daresay you haven't had
                                          much practice," said the Queen.
                                                               - Lewis Carrol