Still a way to go (no, not a happy camper)

Discussion in 'Internet Explorer' started by Wrothchild, Jun 30, 2006.

  1. Wrothchild

    Wrothchild Guest

    Here's what I think of the new (and old) version(s) of I.E.:

    When I installed IE7 beta 2, I had hoped (and crossed my fingers) that
    the long-standing problem of Internet Explorer disobeying its own cache
    settings might have been solved, but I see it's no better, and now I've
    installed beta 3 and still there has been no change in this problem. This is
    not a bug; it was designed this way. I am sure of that; it's been that way
    since at least I.E.5 and probably even further back. My OS, BTW, is Windows
    XP SP2.

    What I mean by this is, that when the setting "Check for newer versions
    of stored pages" is set to NEVER, it should do just that, NEVER CHECK for
    them REGARDLESS of any meta-tags or other commands present in the web page
    telling it to do otherwise. When I click the Back button, it should reload
    the page from the cache (preferably memory) instantly. Many sites that I
    visit (particularly a local newspaper site that I view daily) have ads and
    ad photos (and other content) that they change with every return to the home
    page, and I should be able to control whether the browser loads them each
    time or disregards any changes (bypass the changed content) BUT I CAN'T,
    because I.E. allows the web pages to make that decision, overriding its own
    settings. The only time the setting works, is when a web page doesn't
    contain commands to override it. When on a slow dialup connection as I am (I
    know, I know, EVERYONE is on broadband now, but not me), it's extremely
    inconvenient (and sometimes utterly useless) to try to view sites like that.
    And BTW, broadband isn't available where I am, except for satellite which is
    prohibitively expensive.

    The Opera browser (which I use mainly because of this problem that
    exists in I.E.) doesn't have this problem. When I hit the back button in
    Opera, it immediately snaps back to the previous page, EXACTLY as it is
    stored in the cache, EXACTLY as the settings are set for it to do, without
    hesitation, on ANY web page. As it should be. I have posted about this
    before, and have been told by some of you (MS-MVPs) that 'something must be
    wrong with my settings'. I know better, so don't try to tell me that again.
    I also know, that this subject has been brought up before, by people other
    than myself. This needs to be CHANGED, no matter what the software designers
    may think. Give me one good reason why it is allowed to work this way
    (notice I said a GOOD reason). If people want to be sure they don't miss
    changed content on certain sites (eBay comes to mind as an example) they can
    set it to check for newer versions of stored pages (and let it come set that
    way by default). Better yet, have a feature where certain sites can be set
    up (in a list form) to do this automatically without affecting other sites.

    You can add tabs (nice) and all the goodies you want; make it sing and
    dance if you want to, but until you fix THIS problem, I am staying away from
    Internet Explorer. Come on, get it right; I know you CAN!

    See, I didn't pull any punches this time. Enough is enough.

    Jody
     
    Wrothchild, Jun 30, 2006
    #1
    1. Advertisements

  2. Eric Lawrence [MSFT], Jun 30, 2006
    #2
    1. Advertisements

  3. Wrothchild

    Tombraider42 Guest

    I don't understand your point. That page seems to explain to web page
    developers how to write their web pages. How does this solve Jody's problem?

    I don't use Internet Explorer anyway, I use the newest FREE Opera 9.0 which
    is really a great browser...even use its mail system which is really super
    too. Bit Torrent, Chat, and RSS built right in and ability to take notes,
    save them and search for them means I never have to leave my browser!

    Of course, it had tabbed browsing before Firefox and IE and I LOVE the
    "create linked" feature and easy way to add any search on any page to your
    drop down search box.

    Opera 9.0 works superbly in Vista beta2.

    (Heh, I just selected the text of my post and right-clicked to save it to a
    note in Opera in case the posting gives an error message - really great
    stuff!)
     
    Tombraider42, Jun 30, 2006
    #3
  4. Wrothchild

    Wrothchild Guest

    Looking down the page (from the link you suggested), when you go to the
    section titled "Table 3. Cache behaviors", (my point is that) when the
    Setting is NEVER, there should be "No Request", no matter whether the "Cache
    copy is fresh" or "Cache stale" (see table 3 of the web page). There
    shouldn't even be a conditional request. There should be no request
    whatsoever. I KNOW how the browser works; what I am saying is that the way
    it works is WRONG, and HAS BEEN wrong for a long time, in my opinion. No web
    site should have any control of whether the cache is to be refreshed or not
    on returning to a previously viewed page, WHEN the setting is set to NEVER;
    no matter what. Control should be completely up to the end user ONLY, not
    the web site designer. Opera does it that way; when it's set not to refresh
    the cache, it does just that; it obeys the settings, not the content of a
    web page. WHY IS THIS SO HARD TO UNDERSTAND?

    Jody
     
    Wrothchild, Jun 30, 2006
    #4
  5. Wrothchild

    Wrothchild Guest

    The page shown by the link, although it is (as you said) meant to
    instruct web page developers, does graphically illustrate the problem. See
    my earlier post, and take a look at "Table 3. Cache behaviors" on the web
    page. BTW, I am also using Opera 9. I'd like to use I.E. also, but this
    problem really inhibits me from it. I have posted about this problem before,
    and got similar responses (which is why I wrote certain things in my post
    this time). It's not a bug, and it's not how I have my cache settings set;
    it's the way the browser is designed. The designers apparently saw fit to
    allow web designers to have control over end users cache behavior. I simply
    do not agree with that. More than just not agreeing however, in my case (and
    on some web sites) it makes browsing extremely frustrating. Oh well, I tried
    (again)...

    Jody
     
    Wrothchild, Jun 30, 2006
    #5
  6. I understand what you're asking for, but it's simply not a scenario that the
    cache settings are meant to accomodate. The "Work offline" feature on the
    file menu is intended for the "don't make any http requests" scenario.

    In general, we know that the offline browsing scenario isn't one that IE
    supports very well (although there are some good plugins that help). This is
    something we'll be looking at for future releases.

    Cheers,
    --
    Program Manager
    Internet Explorer
    Trust & Networking
     
    Eric Lawrence [MSFT], Jul 1, 2006
    #6
  7. Wrothchild

    Wrothchild Guest

    Just to be sure you DO know what I mean, I'm not talking about the "Work
    Offline" function. I'm talking about being ONLINE, and downloading a sites
    home page, then clicking a link on that home page to go to another page,
    reading the page and then clicking the Back button to return to the original
    (previously viewed) home page (all of this being done while in ONLINE mode
    and CONNECTED). On many (most) sites, to return to that page entails
    downloading the page (or parts of it) over again. That is my problem. Even
    when the cache setting is set to NEVER, it still does this, IF the web page
    in question contains data (meta tags) telling the browser to do so. My point
    is that when the setting is set to NEVER, web sites should not be able to
    override it. I don't CARE if a web site changes content and the web master
    wants me to see it; I want to see exactly the same thing I saw the first
    time I viewed the page, and I want to return to it INSTANTLY, without having
    to download it (or any part of it) again. I have my browser set to delete
    the cache (Temporary Internet Files folder) when I close the browser anyway,
    so offline browsing isn't what I was referring to. I don't do offline
    browsing anyway. Any page that I've viewed in a session, I should be able to
    return to, instantly, without downloading it again, while I'm still in that
    session (and I'm talking ONLINE here, and connected, not in offline mode),
    and I should be able to do this whether the site designer wants me to be
    able to or not (meta tags and/or other commands embedded into HTML documents
    should not be able to take or influence control of the cache behavior *IF*
    the setting is set to NEVER check for newer versions of stored pages). That
    is what I'm asking for. When the setting is set to AUTOMATIC, then the web
    site (or HTML page) should be able to dictate/influence how the cache
    behaves, but NOT when set to NEVER. If any of this isn't clear, I'd be happy
    to try and clarify it further because I consider it very important. I
    appreciate your attention to it.

    Jody
     
    Wrothchild, Jul 1, 2006
    #7
  8. ....

    Eric,

    The OP has not given any specific examples unfortunately to support
    the case. However, coincidentally yesterday I saw one in a separate
    incident which could be used as an example of deficient processing
    given Cache-Checking: Never.

    http://www.pcl.com/projects/Active/0900343/index.aspx

    The poster noted that there was a lot of cache-checking when
    the window was restored after being minimized. In fact, there
    is a lot of cache-checking almost any time the user does anything
    with that page. ; )

    I think that it could be due to this bit of scripting in it:

    <extract>
    <SCRIPT type="text/javascript">
    window.onresize = function(){
    if (!document.all){
    ddmx.init(); // this only works in mozilla, IE calls multiple
    }else{
    //document.location.reload();
    window.history.go(0);
    }
    }
    var ddmx = new DropDownMenuX('header_menu1');
    ddmx.delay.show = 0;
    ddmx.delay.hide = 400;
    ddmx.position.levelX.left = 0;
    ddmx.init();
    </SCRIPT>
    </extract>

    I suggested disabling scripting and in fact that avoids the symptom.


    I don't have a script debugger installed in this partition to understand
    exactly when that would be called or what it would do but if that comment line

    //document.location.reload();

    was the real thing happening then I suspect it would mean that the implementation
    of it would induce a Refresh (e.g. as if a user pressed F5)?

    I can understand that the implementation of a real F5 might bypass
    a user's preference for Cache-Checking: Never and actually
    force cache-checking on all elements but perhaps it would help
    if you would explain why some scripting should be allowed to do so?


    <eg>

    Robert Aldwinckle
    ---
     
    Robert Aldwinckle, Jul 2, 2006
    #8
  9. ....

    I think you're missing Eric's point. He understands that you want
    to read *only* from the cache and is saying the designed mechanism
    for doing so is setting Work Offline. When you are online, even with
    Cache-Checking: Never there are still going to be some checks made,
    especially when scripting is involved or other mechanisms associated
    with Refresh.


    BTW you haven't mentioned if you have disabled (in Security, Custom Level...)

    Allow META Refresh

    but understand that by doing so you may have to provide manual
    intervention for some site's implementation of redirect? (Etc.)


    And that is when you could be setting Work Offline before going back
    to ensure that there is no HTTP traffic.


    If it's all cached, Work Offline will find it. ; )


    HTH

    Robert Aldwinckle
    ---
     
    Robert Aldwinckle, Jul 2, 2006
    #9
  10. Wrothchild

    Wrothchild Guest

    OK, I just checked my settings and found that "Allow Meta Refresh" was
    set to "Enable"; I have now set it to "Disable". I do use custom settings,
    for all security zones. While on the subject of working in Offline mode, I
    want to mention that I have a few times tried to return to a previously
    viewed page by first going to Offline Mode and then clicking the Back
    button, and Internet Explorer (at least some of the time, and I haven't
    experimented with a lot of web pages to see just how often) will tell me
    that the page is unavailable Offline. This is in Internet zone that I'm
    talking about here, so the pages would be cached (I have the browser set to
    not cache pages in SSL but that's not what I'm referring to here). BTW, I
    found a problem with IE7 Beta 3, and I uninstalled it, so now I'm back to
    I.E.6. What was happening, was when I ran the Disk Cleanup utility (whether
    I.E. was running or not), one of the selections in the cleanup window (which
    I remember said something in it about occache.dll) was corrupting the "View
    Files" button. The button, instead of saying View Files on it, showed some
    text that was part of the item that was selected in the cleanup window (the
    selection that said something about occache.dll). After I uninstalled Beta
    3, the problem went away.

    Jody
     
    Wrothchild, Jul 3, 2006
    #10
  11. ....

    First of all when that happens click on Stay Offline and keep clicking
    on Stay Offline as long as you seem to be getting further into the
    rendering process. E.g. frames may be rendering and each frame
    may have some hidden non-cached .gifs (such as advertisements),
    etc.

    Alternatively try disabling Active Scripting or toggling off Images.
    Tip: you can use the IE Developer Toolbar to make those changes
    easily.

    Also use FiddlerTool to help show you (after you reluctantly allow
    a connection) exactly what the problem was.

    Please share some URLs of problem examples if you would like to
    discuss them. It's much easier analysing particular cases than trying
    to generalize without even citing any examples. I think understanding
    particular cases would make you more tolerant of the problem too.
    As I pointed out there are some apparent deficiencies in the rendering
    process regarding cached elements but often a lot of time
    they make sense once you can break out all the details.



    Apparently this has previously been noted:

    http://blogs.msdn.com/ie/archive/2006/06/29/649910.aspx

    (MSN Search for
    "internet explorer" beta occache
    )

    I suggest a separate post to discuss this further if you wish.


    HTH

    Robert
    ---
     
    Robert Aldwinckle, Jul 3, 2006
    #11
  12. Wrothchild

    Wrothchild Guest

    If you want me to give you a URL so we can discuss it, I can, but the
    one I have in mind, that REALLY gets me frustrated, is a local newspaper,
    and you have to register as a user to get into it (there is a free version,
    which is what I use). However, if I am to give you that particular URL, I
    would want to take this discussion to email (for privacy reasons). I assume
    you're using a valid email address (I'm not), so if you want, I can send you
    the URL that way. The home page of the newspaper is about a megabyte or more
    in size, and has to be returned to after reading each front page article
    (which is what got me started on this). On a slow dialup connection (and
    with some kind of a phone line problem that the telco refuses to try to
    solve, which slows things down even more, and which isn't apparently caused
    by something easily solvable like noise, but by some other technical
    phenomenon, which I have fought with them over for years), trying to reload
    a megabyte sized page every time I read a newspaper article is a royal pain.
    Now, as to clicking Stay Online, I have done that, but the rendering
    stops on the first try, so I don't get anywhere with that. I'm not familiar
    with FiddlerTool and I don't know where to get the Developer Toolbar, so you
    might let me know about that. I will continue to monitor this thread, for as
    long as you want to work on this. Or we can start another one.

    Jody
     
    Wrothchild, Jul 3, 2006
    #12
  13. At some point in the IE5.x timeframe, cache-correctness was made a top
    priority and IE started respecting cache headers if present. The problem is
    that a large % of sites have no directives, and when a site does specify
    caching headers, the values are typically ~bad~ for performance due to
    overzealous use of the "no-cache" directive rather than "private" which is
    typically more appropriate. (On the other hand, see www.google.com or
    office.microsoft.com for counter-examples).

    Before cache-correctness became such a priority, work offline worked quite
    well-- the user was able to view any content, even that which the server had
    specified should not be saved. Unfortunately, this would lead to unexpected
    results if the user was viewing outdated content and didn't realize it.

    In a future version of IE, we're going to be looking at how to provide a
    better pseudo-offline experience where net-traffic is minimized.
    Unfortunately, for best results, site owners have to help us out by making it
    clear which content should be cached, compressing content, and so forth.
    We're evangelizing this to site owners and I've written some code (Fiddler)
    and documentation (E.g.
    http://msdn.microsoft.com/library/default.asp?url=/library/en-us/dnwebgen/html/ie_introfiddler2.asp) but it's slow going to put it mildly.

    Generally, when I find a site making terrible use of the cache and not
    compressing large content, I send them a friendly pointer to the Fiddler
    article and explain how they could save a fortune in bandwidth while reducing
    load on their servers all while improving the customer experience.
     
    Eric Lawrence [MSFT], Jul 6, 2006
    #13
  14. Wrothchild

    Wrothchild Guest

    You wrote:

    "At some point in the IE5.x timeframe, cache-correctness was made a top
    priority and IE started respecting cache headers if present.

    This is precisely what I have been complaining about, both in this
    thread and in another old one on another Microsoft USENET group. My point
    being -- That IE shouldn't be designed to respect cache headers when set to
    cache: never. When set to automatic is when it should obey cache headers,
    and only then. It should come "out of the box" set to automatic, and if
    people set it to never and then miss out on something on a site because of
    it, then that's their problem. People shouldn't change settings until they
    know what they're doing, is how I see it. Why make it impossible for those
    who do know what they're doing, to have it behave as they want it to, just
    so others won't trip on their own mistakes? I have a big problem with that
    kind of thinking. I'd like to see it changed, and that's been the whole
    point of my complaint all along; I wasn't looking for help, (as some seemed
    to think) I was looking to get things changed. I already knew how it worked.
    How I'd like to see it work, it should have a cache in memory (as well
    as on the disk), and when set to cache: never, it should (when the Back
    button is clicked) instantly re-display a previously viewed page, from the
    memory cache without downloading any additional data. When you make it do
    that, you will win me over.

    Jody
     
    Wrothchild, Jul 6, 2006
    #14
    1. Advertisements

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.