1. Advertising
    y u no do it?

    Advertising (learn more)

    Advertise virtually anything here, with CPM banner ads, CPM email ads and CPC contextual links. You can target relevant areas of the site and show ads based on geographical location of the user if you wish.

    Starts at just $1 per CPM or $0.10 per CPC.

Making This Accessibility Wreck Semi-usable

Discussion in 'Support & Feedback' started by deathshadow, Jan 22, 2013.

  1. #1
    This 'new' layout being even worse than the previous one, I sat down and put together some user.css to at least TRY and correct some of the more annoying bits. If you have a user.css extension for chrome or Firefox, or like I do just use Opera since it's built in, this makes things a LITTLE more tolerable... though it's still not great given what absolute bloated rubbish the markup is.

    * {
      font-size:100% !important;
      line-height:140% !important;
      font-family:arial,helvetica,sans-serif !important;
    }
     
    body {
      font-size:85% !important;
    }
     
    .pageWidth {
      max-width:72em;
    }
     
    .messageUserInfo {
      width:12em;
    }
     
    .messageInfo {
      width:auto !important;
      margin-left:13em !important;
    }
     
    .navTabs {
      font-size:16px !important;
    }
     
    .navTabs a {
      line-height:25px !important;
    }
     
    .blockLinksList a {
      line-height:20px !important;
    }
     
    .socialBar {
      display:none !important;
    }
     
    .newIndicator {
      color:#EFC !important;
    }
     
    .privateControls span a {
      line-height:2em !important;
    }
     
    .button,
    .privateControls a.item,
    .publicControls a {
      height:auto !important;
      padding:0.2em 0.5em !important;
    }
     
    .publicControls a {
      display:inline-block;
    }
     
    .scrollable {
      height:2em !important;
    }
     
    .pageNavHeader{
    padding-right:0.4em !important;
    }
     
    .button,
    .publicControls a,
    .privateControls a.item,
    .PageNav a {
      color:#000;
      background:#EFD !important;
      border:1px solid;
      border-color:#749B85 #749B85 #618872;
      -webkit-border-radius:7px;
      -moz-border-radius:7px;
      border-radius:7px;
      -webkit-box-shadow:
          0px 1px 4px 0px #C8C8D2,
          inset 0 4px 8px #FFF,
          inset 0 -4px 8px #BDB;
      -moz-box-shadow:
          0px 1px 4px 0px #C8C8D2,
          inset 0 4px 8px #FFF,
          inset 0 -4px 8px #BDB;
      box-shadow:
          0px 1px 4px 0px #C8C8D2,
          inset 0 4px 8px #FFF,
          inset 0 -4px 8px #BDB;
    }
     
    .button:active,
    .button:focus,
    .button:hover,
    .publicControls a:active,
    .publicControls a:focus,
    .publicControls a:hover,
    .pageNav a:active,
    .pageNav a:focus,
    .pageNav a:hover {
      background:#CEC !important;
    }
     
    #AlertsMenu_Counter .arrow,
    #header_data,
    #header_middle,
    .underPost {
      display:none;
    }
     
    #AlertsMenu_Counter {
      position:relative !important;
      top:0 !important;
      padding:0.1em 0.4em;
    }
     
    .formPopup {
      width:16em !important;
    }
     
    #QuickSearchQuery {
      width:15em;
    }
     
    #QuickSearch * {
      line-height:1.4em !important;
    }
     
    #QuickSearch input.button.primary {
      width:auto !important;
      margin-right:0.4em;
      padding:0.2em 0.5em !important;
    }
     
    #QuickSearch a.moreOptions {
      overflow:hidden;
      display:block;
      margin:0;
      padding:0.2em 0 !important;
      height:auto !important;
    }
     
    #commonSearches {
      margin-left:0.4em;
      padding:0.2em 0.5em !important;
      height:auto !important;
    }
     
    .pageNavLinkGroup a {
      width:auto !important;
      padding:0.1em 0.2em !important;
    }
     
    #messageList li:nth-child(even) {
      background: #fcfcff; /* Old browsers */
      background: -moz-linear-gradient(left, #fcfcff 0%, #f4ffe0 50%, #fcfcff 100%); /* FF3.6+ */
      background: -webkit-gradient(linear, left top, right top, color-stop(0%,#fcfcff), color-stop(50%,#f4ffe0), color-stop(100%,#fcfcff)); /* Chrome,Safari4+ */
      background: -webkit-linear-gradient(left, #fcfcff 0%,#f4ffe0 50%,#fcfcff 100%); /* Chrome10+,Safari5.1+ */
      background: -o-linear-gradient(left, #fcfcff 0%,#f4ffe0 50%,#fcfcff 100%); /* Opera 11.10+ */
      background: -ms-linear-gradient(left, #fcfcff 0%,#f4ffe0 50%,#fcfcff 100%); /* IE10+ */
      background: linear-gradient(to right, #fcfcff 0%,#f4ffe0 50%,#fcfcff 100%); /* W3C */
      filter: progid:DXImageTransform.Microsoft.gradient( startColorstr='#fcfcff', endColorstr='#fcfcff',GradientType=1 ); /* IE6-9 */
    }
     
    #messageList .messageInfo {
      background:transparent;
    }
    Code (markup):
    Like most user.css it's annoying all the !important -- but that's specificity for you... and of course in many ways is throwing good code after bad.

    I've also found the site a LOT more useful once I disabled scripting and went in and blocked ALL script files from even being allowed to load. (again, Opera's 'block content' can be a godsend). I'm going to play with targeting specific scripts and see if I can find a happy middle ground between pointless slow bandwidth wasting garbage (like that stupid malfing social bar) and useful functionality (like Ajax posting).
     
    Last edited: Jan 22, 2013
    deathshadow, Jan 22, 2013 IP
  2. Nigel Lew

    Nigel Lew Notable Member

    Messages:
    4,642
    Likes Received:
    405
    Best Answers:
    21
    Trophy Points:
    295
    #2
    If I may, there are ways to go about things, and there are ways to not go about things. You seem to be a smart guy but got stuck in the mud of the latter.

    Getting work by being a jackass is not a tenable situation.
    Nigel
     
    Nigel Lew, Jan 22, 2013 IP
    ryan_uk likes this.
  3. deathshadow

    deathshadow Acclaimed Member

    Messages:
    9,732
    Likes Received:
    1,998
    Best Answers:
    253
    Trophy Points:
    515
    #3
    1) Who says "getting work" has anything to do with ANYTHING -- this is me sharing user.css so other people who want to use it before (or even if) things are fixed can do so. You do know what user.css is, right?

    2) If something is a turd, call it a turd -- if there's ineptitude call it ineptitude; that's how the world gets better. In the words of George Bernard Shaw, "The reasonable man adapts himself to the world; the unreasonable one persists in trying to adapt the world to himself. Therefore, all progress depends on the unreasonable man."

    Pussy-footing around it, slapping the rose-coloured glasses on everyone's head to lead them further down the garden path, patting everyone on the back with an ill-deserved "well done" like some soft namby-pamby limp-wristed drum circle wussy -- well, that's a hefty chunk of everything wrong with society today. Most of the time you have to walk right up to someone and say "Hey buddy, take a shower"... but no:

    This test is too hard! (So they lower the standards)
    I’m not good at sports! (So they give them all trophies)
    My dad used to spank me (So they lower the standards)
    I’m too fat for this seat (So they widen the standards)
    They say no 'cause I’m black (So they lower the standards)
    They say no 'cause I’m white (So they lower the standards)
    They say no 'cause I’m Asian (So they lower the standards)
    No hablo Inglais (So we all become Spaniards)

    If said attitude qualifies as being a jackass, so be it! Better than the sleazeball shmucks, shmendriks and nudniks that have pretty much destroyed the functionality of the Internet by preying on the ignorance of others the past five or six years... to the point that forums, webmail, information websites and even shopping isn't as usable, functional or practical as it was a decade ago. We've been given the tools to make things better, and instead everyone just shoves their heads further up 1997's backside.

    Though when it comes to the current state of web development, we have reached the point of needing a Gordon Ramsey, Simon Cowell or Charlie Frattini to give the entire blasted industry an enema. I have reached the point of disgust to the point of nausea with the sleazeball scam artists, half assed nitwits unqualified to even be writing HTML much less CSS or PHP, and practice of throwing "Millions of lines of code" at the job of thousands or even hundreds of lines; especially when they then have the cojones to call their fat bloated half-assed overcomplicated solutions "easier" to work with.

    See statements like "over 1.5 million lines of code" when something like a forums and CMS based website realistically shouldn't break that many QWORDS... so unless it's averaging 8 characters per line... Again, see using 130k of markup and 775k of scripting to deliver 6.23k of plaintext.
     
    deathshadow, Jan 22, 2013 IP
  4. digitalpoint

    digitalpoint Overlord of no one Staff

    Messages:
    38,333
    Likes Received:
    2,613
    Best Answers:
    462
    Trophy Points:
    710
    Digital Goods:
    29
    #4
    Obviously you are free to use whatever CSS or JS blocker you want, but I really have no clue where you are getting your numbers.

    After flushing browser cache, the main page weighs in at 250KB for me on first load. Subsequent visit (without flushing browser cache) comes in at 16KB.

    Even measuring traffic at the switch level is impossible with the numbers you are throwing out. With the number of visitors we have, with your numbers we would need to be passing about 200Mbit/sec overall. We are under 10Mbit right now with thousands of visitors on (and many, many page views per second happening).

    Also, if the 1.5 million lines of code thing is directed at me, how many full bidding style advertising platforms or digital goods marketplace have you built with a few hundred lines of code? As far as I'm concerned, HTML, CSS or even JavaScript are not "code" and aren't part of that number. Building the backend for a mini AdWords+AdSense advertising system is though (that part was probably around 300k worth).
     
    digitalpoint, Jan 23, 2013 IP
  5. deathshadow

    deathshadow Acclaimed Member

    Messages:
    9,732
    Likes Received:
    1,998
    Best Answers:
    253
    Trophy Points:
    515
    #5
    Firefox "Web Developer" extension.
    https://addons.mozilla.org/en-US/firefox/addon/web-developer/

    
    	Size	Uncompressed Size
    12 Documents	447 KB	491 KB
    16 Images	85 KB	85 KB
    0 Objects		
    15 Scripts	266 KB	775 KB
    2 Style Sheets	152 KB	152 KB
    45 Files	950 KB	1503 KB
    Code (markup):
    The document count is high because of all the iframe contents, NONE of which are caching, and that the document is for some reason being loaded twice in the count -- likely due to a faulty reflow (I've seen that before). Likewise a number of scripts are hanging the pageloads here which is why even after a waterfall is finished, you can't type or even scroll in Opera or Chrome for >10 seconds here. (painful after 15 seconds page-load on subsequent pages).

    Subsequent pages here coming in at ~380k in 40+ files. It's hard to peg down an accurate number since the scripting keeps on loading even more garbage (adverts) in the background -- to the point that if you accidentally leave a tab/window open for more than a minute or two Firefox is ready to blow it's lid due to it's piss poor memory management, Opera is thrashing to the disk, and Chrome gets painful to scroll the page.

    Well, much of that CAN be explained -- firstload vs. subsequents -- much of the scripting NOT being on the host, much of the image bloat (adverts) not being on the host. Breakdown the scripts being loaded:

    
    https://connect.facebook.net/en_US/all.js	57 KB	174 KB
    https://x.dpstatic.com/j/xenforo/xenforo.js?_v=e2cbe1b0	42 KB	137 KB
    https://ajax.googleapis.com/ajax/libs/jquery/1.5.2/jquery.min.js	29 KB	84 KB
    https://apis.google.com/_/scs/apps-static/_/js/k=oz.gapi.en_US.qnskwPQ6cxA.O/m=plusone/am=iQ/rt=j/d=1/rs=AItRSTMp-COqtlKeux7vie0Db-3pG6Sxkg/cb=gapi.loaded_0	25 KB	72 KB
    https://pagead2.googlesyndication.com/pagead/expansion_embed.js	25 KB	69 KB
    https://platform.twitter.com/widgets.js	25 KB	79 KB
    https://pagead2.googlesyndication.com/pagead/js/r20130116/r20121214/show_ads_impl.js	20 KB	57 KB
    https://www.google-analytics.com/ga.js	15 KB	36 KB
    https://apis.google.com/js/plusone.js	9 KB	19 KB
    https://pagead2.googlesyndication.com/pagead/osd.js	7 KB	17 KB
    https://pagead2.googlesyndication.com/pagead/show_ads.js	6 KB	13 KB
    https://www.gstatic.com/bg/ao7cfjproEEAn5OijXEgfdHyA_5BKye5jml3JCZ7KvU.js	3 KB	8 KB
    https://platform.twitter.com/js/tfw/hub/client.js	3 KB	8 KB
    https://tpc.googlesyndication.com/pagead/js/r20130116/r20110914/abg.js	366 bytes	819 bytes
    https://cdn.api.twitter.com/1/urls/count.json?url=http%3A%2F%2Fforums.digitalpoint.com%2F&callback=twttr.receiveCount	102 bytes	102 bytes
    
    Code (markup):
    How much of that is local? 137k for xenforo (42k compressed)... and that's it. First load it's fun watching it struggle with all those DNS lookups, that can take as long or longer than a handshake. Blocking all the google crap and pointless social networking nonsense works wonders for speeding things up. Social to draw people in is worthless if it also drives users away because it takes too damned long. This is REALLY true once you start ajax-ing in even more garbage during and after pageload.

    Advertising scam bull I don't put on websites in the first place and actively block... Of course such nonsense being "take 2" of the same things that led to the dotcom bust.

    I based my numbers off existing projects like SMF, SiteSense, and Magento, and combining their TOTALS together.

    JAVASCRIPT isn't code? Wow... Just... wow... HTML and CSS might not qualify in some people's minds as programming but they most certainly ARE code. If you're not counting those towards it -- that's even more frightening.

    So I imagine that this:
    
    #header_data,
    #header_middle,
    .underPost {
      display:none;
    }
    Code (markup):
    ... really hurts -- much less script and advertising domain blocking. 300k lines of code neutered by a fraction that?

    Though I'll admit that everything outside the forums you have means Jack to me because I'm not a sleazeball scam artist -- which are the types who seem to populate things like the marketplace or deal with online advertising.... the latter in particular being such a bunch of untrustworthy sleaze I cannot believe anyone would voluntarily place other folks ads on their sites. You'd think people would have learned their lesson during the dotcom bust... but NOOOO....

    Not that people who build websites just to sell them or the PSD jockeys who crap out pretty pictures and call them templates are much better -- or the people who write a 5 page pamphlet of marketspeak and have the Huevos Rancheros to charge money for them as a e-book -- again, sleazy scam artist bull that does nothing but prey upon the ignorance of others.
     
    deathshadow, Jan 23, 2013 IP
  6. jimmyn115

    jimmyn115 Member

    Messages:
    145
    Likes Received:
    7
    Best Answers:
    0
    Trophy Points:
    33
    #6
    Has the internet stopped working? Shit.

    I thought my problem with "likes" was as bad as it got.
     
    jimmyn115, Jan 23, 2013 IP
  7. civilwartalk

    civilwartalk Well-Known Member

    Messages:
    16
    Likes Received:
    4
    Best Answers:
    0
    Trophy Points:
    133
    #7
    Sounds like DP is headed for a disaster of biblical proportions.

    Fire and brimstone coming down from the skies! Rivers and seas boiling!
    Forty years of darkness! Earthquakes, volcanoes...
    The dead rising from the grave!
    Human sacrifice, dogs and cats living together... mass hysteria!
     
    civilwartalk, Jan 23, 2013 IP
    jimmyn115, newbie191 and Helvetii like this.
  8. Nigel Lew

    Nigel Lew Notable Member

    Messages:
    4,642
    Likes Received:
    405
    Best Answers:
    21
    Trophy Points:
    295
    #8
    This site gets a 95/100 on the google speed test https://developers.google.com/speed..._3A_2F_2Fforums.digitalpoint.com&mobile=false

    with a few things taken into consideration that is pretty fast but more noteworthy, well delivered. In general the site is much faster for me. That statement assumes however, I am somehow able to measure "loading immediately". Its hard to make a stock wordpress site to perform as well.
    Nigel
     
    Nigel Lew, Jan 23, 2013 IP
  9. digitalpoint

    digitalpoint Overlord of no one Staff

    Messages:
    38,333
    Likes Received:
    2,613
    Best Answers:
    462
    Trophy Points:
    710
    Digital Goods:
    29
    #9
    Well, really have no clue why the site is so sluw for you, but I suspect it's something with your connection or computer somehow. Overall the site is less http requests under the new design, less KB needed to download, less DNS lookups and more importantly, pages are loading roughly 3x faster (for example this pageview of this thread took 0.31 seconds). And that is on my computer that is 5 years old almost to the day.
     
    digitalpoint, Jan 23, 2013 IP
  10. deathshadow

    deathshadow Acclaimed Member

    Messages:
    9,732
    Likes Received:
    1,998
    Best Answers:
    253
    Trophy Points:
    515
    #10
    Don't even get me STARTED about Google's pageSpeed "ratings" -- since that's the pile that will give this at 991k (after compression) in 61 files, taking 4 seconds (check the waterfall, or as they call it "critical path explorer") on google's "more bandwidth than God" connections a 96/100 (currently), yet will give a 67k page (after compression) in 22 files taking 900ms on the same connection a 71/100... Oh wait, it's dinging because the smaller site isn't using a CDN, and isn't second-guessing browsers's caching behaviors, isn't wasting time and effort on whitespace compression becuase it doesn't need to... and of course is silently dropping ten points on it because god forbid you run a site without their fat bloated analytics crap. (that any useful information is redundant to your server logs via tools like analog and webalizer, the rest being micromanaged BS people waste time on for nothing).

    Of course it's always suspect for the people selling space on CDN's and trying to get you to use their tools to ding you points for not using CDN's or using their tools... but of course nobody takes the time to rub the brain cells together and say "HEY!" which is why so many seem to take tools like Google's Pagespeed or Yahoo's ySlow seriously; when they're kind-of a joke compared to just saying "How big is it", "How many files is it", and "How much does the scripting mess with it during loading?"

    Basically, gives higher ratings to slower loading pages -- oh yes, that makes sense... While back someone was asking me about it since their site was taking THEM over a minute to load -- was around 30 seconds here (geography in relation to the host ALWAYS has an impact) -- and was 22 seconds according to Google PageSpeed -- but was getting a 99/100 because he was doing every single goof assed trick in the book they base their rankings on. Meant absolutely nothign for the actual speed of the site since it was over 2 megabytes in a couple hundred files, some megabyte plus of that being endless pointless scripts for nothing, multiple scripting frameworks, that 'Less' garbage...

    Lemme format the output from web developer's "document size"
    
                    Size    Uncompressed  Size
    11 Documents    168 KB  310 KB
    38 Images       404 KB  404 KB
    0 Objects   
    14 Scripts      266 KB  774 KB
    2 Style Sheets  153 KB  153 KB
    65 Files        991 KB  1641 KB
    Code (markup):
    Oh yes, no clue why it's so slow. Really? REALLY? Just ... wow.

    Though it DOES seem miraculously faster out of nowhere since I got back to my workstation a few minutes ago.. if a ten second pageload once things should be cached qualifies as 'faster'. Two and a half times slower than Google's magical mystery connect -- sounds about right. Though the most painful part remains hitting 'reply'' ...
     
    deathshadow, Jan 23, 2013 IP
  11. Nigel Lew

    Nigel Lew Notable Member

    Messages:
    4,642
    Likes Received:
    405
    Best Answers:
    21
    Trophy Points:
    295
    #11
    Its blazing fast here. That is largely my point. I suspect you have what we call a problem between the chair and the keyboard.

    That being said, if things are so dreadfully ill equipped around here to keep up with the times I am sure DP won't mind if you simply take your ball and go home.

    I will further note that I am not trying to argue with you. I just don't quite get the basis of your vitriol.

    Nigel
     
    Last edited: Jan 23, 2013
    Nigel Lew, Jan 23, 2013 IP
  12. digitalpoint

    digitalpoint Overlord of no one Staff

    Messages:
    38,333
    Likes Received:
    2,613
    Best Answers:
    462
    Trophy Points:
    710
    Digital Goods:
    29
    #12
    Honestly, you really probably want to have your computer looked at or something. I'm having flashbacks to when the old site took 30 seconds for you to render too.

    Yet for everyone else, it took about 1 second (even when clearing the browser cache completely)...

    https://forums.digitalpoint.com/thr...ent-really-wrong.2517416/page-2#post-17909140

    And again... I'm not on some fancy crazy super computer. I got my computer in January of 2008.

    Even this thread, look at the Timing/Memory/Queries numbers at the bottom... that's the time it took the server to do everything on it's end and output the page:

    [​IMG]

    Overall the site is using about 40% *less* bandwidth than the old design as measured from the best measurement point, the network switch (more visitors too). And on top of it, if you are using a browser that supports the SPDY protocol (the HTTP protocol is inherently slow and insecure (which makes sense because it was designed over 20 years ago). So if you are using Chrome 13+, Firefox 13+ or Opera 12.10+, you are actually not using HTTP here, you are using the SPDY protocol (does things like eliminate the need for roundtrips for HTTP requests, can compress the "HTTP" headers, which can't be done with the HTTP protocol, etc.)
     
    digitalpoint, Jan 23, 2013 IP
  13. ryan_uk

    ryan_uk Illustrious Member

    Messages:
    3,983
    Likes Received:
    1,022
    Best Answers:
    33
    Trophy Points:
    465
    #13
    I have to wonder if you're using an old version of Opera and refuse to update as it is (in your view) inadequate somehow? Have you heavily customised it? Are you running Opera through a proxy so you can do packet inspection and this is causing a slowdown?

    I'm finding it insanely faster in Chrome, so you might bite the bullet and try it, never mind how much it pains you (the latest version, too). You're located in NA, right? I'm in the UK, but not experience any issues despite the greater distance.
     
    ryan_uk, Jan 23, 2013 IP
  14. N_F_S

    N_F_S Active Member

    Messages:
    2,475
    Likes Received:
    56
    Best Answers:
    0
    Trophy Points:
    90
    #14

    It has nothing to do with speed of page load.
     
    N_F_S, Jan 23, 2013 IP
  15. digitalpoint

    digitalpoint Overlord of no one Staff

    Messages:
    38,333
    Likes Received:
    2,613
    Best Answers:
    462
    Trophy Points:
    710
    Digital Goods:
    29
    #15
    It's not about browser's caching behaviors. A browser caches based on what the server tells it to cache. For example if a site would be insane to serve static content without Cache-Control headers, and here's why...

    With the Cache-Control headers, you are telling the browser when to check for an updated version of the content. For content that is truly static (images, JS, CSS, etc.) we tell browsers to not even bother checking if it's updated until a year from now. If you don't tell the browser how to cache something, the browser will make an HTTP request to get the item with the "If-Modified-Since" request header... if the item hasn't been modified since the date the browser is telling it, the server won't send it.

    But here's the kicker... the "expensive" part of an HTTP request isn't the body of a document being return, it's the actual network connection and the fact that HTTP cannot compress HTTP headers (both to and from). So say you have a page with 10 static items, every page view you are adding probably around 20k in network traffic just to do the If-Modified-Since checks that never yield any new content or changes (again... HTTP HEADERS CANNOT BE COMPRESSED like the payload of an HTTP request can, unless you are using SPDY protocol instead of HTTP).

    While it's pretty clear we aren't going to agree on some things, believe it or not, I *am* a borderline expert on making things as fast as they can be with servers and stuff. :)

    Either way... how browsers cache something is something the server that is serving the content dictates, not the other way around. So a lack of Cache-Control headers for static content is silly at best.
     
    digitalpoint, Jan 23, 2013 IP
    civilwartalk and ryan_uk like this.
  16. deathshadow

    deathshadow Acclaimed Member

    Messages:
    9,732
    Likes Received:
    1,998
    Best Answers:
    253
    Trophy Points:
    515
    #16
    Uhm... no... A browser caches what it has room to cache regardless of the headers, unless you send it an header telling it NOT to cache. Otherwise there would be no reason to have it and nothing would be cached without those headers -- and that's 100% bull.

    Which is only rechecked every 3 to 8 hours depending on the browser by default on non-document files! With only 20-50 megs of disk cache and a varying amount of memory cache, meaning setting it to more than that is typically a waste of time as it's gonna get flushed to make room for other sites! (and that will happen quite often if you visit a heavy content site like a wakaba/futaba board or facebook unless you turn images off).You can find the setting for the recheck in Opera under Tools->Preferences->History -- recheck documents always, recheck everything else 5 hours, for FF it's 3 hours somewhere in about:config... and I believe (don't quote me on this) it's a static 8 hours for IE8/newer. (IE7/earlier behaves as you describe)... no clue as to chrome, but given it's speed I suspect it's the same.

    Yes, it's called handshaking -- and it's WHY I keep harping on the number of files people seem to want to put on their sites... and why I personally make 12 total files my ideal target, with an upper allowed limit around 20... and try to restrict myself to just one or two cookies (which is usually overkill) since they increase the size of those pesky HTTP requests. That's why many people move statics to subdomains - so the cookies aren't sent in the header.

    Again, it is NOT -- otherwise files without cache-control headers wouldn't be cached -- when they QUITE OBVIOUSLY ARE!!! (go to a page without them in IE, look in the cache directory... it's in there!)

    Making it a waste of time at best given that cache is a limited resource that is going to flush due to being filled ANYWAYS... much more so with people vomiting up megabyte sized monstrosities. It is not checked EVERY time on most files -- that's what the existing delay of 3 (FF) to 8 (IE) hours is for... all you are doing is changing that delay.

    Though if browser makers started increasing the size of their cache you could see an improvement using such settings; but at the same time with the anemic amounts of memory and not caching to disk on less capable devices like tablets and phones (since they're flash media and writes are bad) it really is a silly bunch of misinformation and 'hosting voodoo'... at least, when compared to keeping the code to content ratio below 3:1 and not throwing endless pointless megabytes worth of images and scripting at pages that don't even have 10k of actual content, reducing the number of files by combining together scripts and images, and other legitimate methods for making a page load fast -- and render fast.
     
    deathshadow, Jan 24, 2013 IP
  17. jimmyn115

    jimmyn115 Member

    Messages:
    145
    Likes Received:
    7
    Best Answers:
    0
    Trophy Points:
    33
    #17
    Loads fine for me.
     
    jimmyn115, Jan 24, 2013 IP
    civilwartalk likes this.
  18. WebDev Solutions

    WebDev Solutions Well-Known Member

    Messages:
    1,644
    Likes Received:
    80
    Best Answers:
    2
    Trophy Points:
    170
    #18
    Insanely fast compared to the older version here, OP may want to take a look at this own set up!

    WebDev
     
    WebDev Solutions, Jan 24, 2013 IP
    ryan_uk likes this.
  19. ryan_uk

    ryan_uk Illustrious Member

    Messages:
    3,983
    Likes Received:
    1,022
    Best Answers:
    33
    Trophy Points:
    465
    #19
    319MB cache here, but that is with Chrome, so no problems caching on DP. Your cache should be more than 20-30MB for Opera, though (AFAIR, 50MB). Perhaps you need to reset Opera's settings, to get the latest recommend defaults. Have you been mess about with opera:config a lot? Anyway, any good browser will follow the HTTP header instructions, I think you know that? So I'm not sure why you're disagreeing. Sure, if you clear the cache (or the browser does), the content needs reloading, but it still checks the HTTP headers (unless you've set your browser not to) and won't later reload until it has to (or you explicitly do a reload that bypasses the cache, but I don't think many people do that often). Have you overridden Opera's expiry settings? Perhaps you did it once and forgot. It could explain the issues you're having with the site.
     
    ryan_uk, Jan 24, 2013 IP
  20. digitalpoint

    digitalpoint Overlord of no one Staff

    Messages:
    38,333
    Likes Received:
    2,613
    Best Answers:
    462
    Trophy Points:
    710
    Digital Goods:
    29
    #20
    Stuff is cached, what I'm talking about is preventing the browser from rechecking if a item has been modified continuously.


    Absolutely not the case... Here's what's going on for the page in your signature upon immediate reload...

    [​IMG]

    As you can see, the browser is checking for if the static resources have been modified every request. And while the body of the request is only a few bytes because it didn't change, the expensive part is the connection itself and the uncompressed HTTP headers.

    Now, let's compare that to this page for example (obviously there's a lot more items to build the page, but when it comes down to it, there's actually LESS connections going on because browsers aren't checking if static content has been modified.

    [​IMG]

    If you want a browser to cache for 3-8 hours, the web server needs to tell it that. While there might be special plug-ins you can install that will override what the web server does (or doesn't tell it) about caching, most users are not going to have those installed.

    I'm totally in agreement... the less files the better. It's why we did exactly that anywhere we could. If your browser supports CSS gradients, it uses CSS rather than gradient image files, all smile images in posts are a single sprite image, all icons on the site are a single sprite image, etc.

    For example:

    [​IMG]

    We also minimize cookie usage and all our static content is on a cookieless domain (dpstatic.com). We couldn't use a sub-domain for that purpose on this site since we have various sub-domains for the normal site (www, tools, forums, etc.) so we had to make the user login cookies domain-wide otherwise users would have to log in every sub-domain they go to.

    But yes... cookieless domains (or subdomains) are huge.

    Not sure what to tell you other than you are absolutely not correct on how browsers cache... caching and checking if a document was changed are not the same things. You can cache something but still check if the document has changed or not continuously as you can see in the screenshots above.


    Check your raw web logs... EVERY HTTP 304 response means it was a request that wasn't needed because someone was checking if the resource was modified and it wasn't.
     
    digitalpoint, Jan 24, 2013 IP