Making This Accessibility Wreck Semi-usable

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

  1. deathshadow

    deathshadow Acclaimed Member

    Messages:
    9,732
    Likes Received:
    1,999
    Best Answers:
    253
    Trophy Points:
    515
    #21
    I just set up a clean room XAMPP and Win7 installs in VM's to test this out -- as what you're saying doesn't jive with my experience. Turns out, it's a little from column A, a little from column B.

    What I was saying about not sending rechecks appears to only apply to images, and javascript in SOME browsers. After a first-load, I tried just hitting 'enter' in the address bar, a refresh, and navigating to a sub-page on the same site that uses all the same files. The results were kinda interesting.

    Again, these are all after firstload, without cache headers.

    "Enter" on address:
    Opera: HTML resent, CSS 304, JS not requested, images not requested
    FF: HTML not resent, CSS 304, JS 304, images not requested
    Chrome: HTML resent, CSS 304, JS not requested, images not requested

    "hard" Refresh:
    Opera: All except images resent... images not even requested
    (Opera's a bit of a pain on convincing it to update images, this can work against you during development at times!)
    FF: All resent
    Chrome: All resent

    Navigate to subpage: (of course HTML is sent on all)
    Opera: no subfiles requested
    FF: CSS 304, JS and Images not requested.
    Chrome: no subfiles requested

    ... and not requested means quite literally they're not even asking for the file again -- WITHOUT setting cache times in the response headers.

    So ok, it might make sense if you have a fat bloated page filled with endless extra stuff that it doesn't need... but really it still strikes me as hosting voodoo instead of fixing the ACTUAL problems.

    Though it's even SLOWER tonight -- you running anything special in the background? It also seems like all pageloads are hanging now waiting on an element that never resolves.

    NOT that this thread was supposed to be about speed -- but the accessibility wreck; but then code bloat goes hand in hand with speed. I mean, four to five dozen separate files totalling over a meg and a half, and few if any others have a problem with this? Whiskey tango foxtrot is wrong with developers these days?!? Are people just getting used to websites sucking? Oh wait, fixed widths, fixed metric fonts, non-semantic bloated markup and more 'gee ain't it neat' scripting than you can shake a stick at... yup, standard practice these days.

    I mean with absolute silliness like:
    
    								<li class="node forum level-n node_27">
    	<div class="unread">
    		<h4 class="nodeTitle"><a href="forums/adsense.27/" class="menuRow">AdSense</a></h4>
    	</div>
    
    Code (markup):
    What makes those headings? Are they the start of a subsection?!? No, because you've got a bunch of them in a row with no content... LI's a perfectly good block level container, so why the div? (particularly since you could also just display:block that anchor) -- its obviously in a perfectly good list, why does it need so many classes? You look at the list tag it's an OL... when there really doesn't seem to be any order to it... What is being done with that which could not be done with just the LI and the A? The answer is not a blasted thing.

    The menu ALONE is ten times the needed markup -- with such laughably bad bits of nonsense as fieldsets without forms, endless pointless DIV when there's plenty of perfectly good semantic tags in use or could be used, endless pointless classes doing inheritance and the 'cascading' part of CSS' job, spans just to add the pointless HTML 5 bloat attributes like itemscope which even if they weren't a waste of code, could/should be applied to existing elements BEFORE diving for the pointless span.

    It's stuff like that which really makes me want to pimp-slap every last person promoting the use of HTML 5.
     
    deathshadow, Jan 24, 2013 IP
  2. digitalpoint

    digitalpoint Overlord of no one Staff

    Messages:
    38,334
    Likes Received:
    2,613
    Best Answers:
    462
    Trophy Points:
    710
    Digital Goods:
    29
    #22
    Yeah, there's something really weird going on for about the last hour. It actually looks like some sort of upstream connectivity issue. Talked with an admin that is upstream of me, and he said the same thing. Not sure *where* the issue is, but it's unfortunately upstream of my network, so not a lot I can do about it other than maybe route around it if we can pinpoint it if we have to.

    Servers are running fairly idle like they always do... highest load on any server is 0.08 right now, queries are running so fast you can't even see a single one as actually running.

    A bit ago the entire network just disappeared for about 30m, and while it seems okay for *me* now, others are reporting what you are seeing.

    [​IMG]

    I think there is some packet lossing going on somewhere on a route that is not between myself and the servers.

    If you ping the servers continuously, so you see any timeouts on the pings?

    This is what I'm seeing from the servers if I try to ping www.nytimes.com...

    PING newyorktimes.com (170.149.168.130) 56(84) bytes of data.
    64 bytes from 170.149.168.130: icmp_req=1 ttl=237 time=71.6 ms
    64 bytes from 170.149.168.130: icmp_req=3 ttl=237 time=71.6 ms
    64 bytes from 170.149.168.130: icmp_req=5 ttl=237 time=71.6 ms
    64 bytes from 170.149.168.130: icmp_req=7 ttl=237 time=71.6 ms
    64 bytes from 170.149.168.130: icmp_req=9 ttl=237 time=72.6 ms
    64 bytes from 170.149.168.130: icmp_req=11 ttl=237 time=71.4 ms
    64 bytes from 170.149.168.130: icmp_req=13 ttl=237 time=1309 ms
    64 bytes from 170.149.168.130: icmp_req=15 ttl=237 time=1273 ms
    64 bytes from 170.149.168.130: icmp_req=17 ttl=237 time=1261 ms
    64 bytes from 170.149.168.130: icmp_req=27 ttl=237 time=1168 ms
    64 bytes from 170.149.168.130: icmp_req=34 ttl=237 time=71.4 ms
    64 bytes from 170.149.168.130: icmp_req=36 ttl=237 time=71.3 ms
    64 bytes from 170.149.168.130: icmp_req=38 ttl=237 time=71.4 ms
    64 bytes from 170.149.168.130: icmp_req=40 ttl=237 time=71.3 ms
    64 bytes from 170.149.168.130: icmp_req=42 ttl=237 time=71.5 ms
    ^C
    --- newyorktimes.com ping statistics ---
    42 packets transmitted, 15 received, 64% packet loss, time 41217ms
    rtt min/avg/max/mdev = 71.317/386.658/1309.532/523.135 ms, pipe 2
    Code (markup):
    64% packet loss, and 30% of the non-lost packets are over a second to return. lol wtf... it has to be some route between the east and west coast somewhere.
     
    Last edited: Jan 24, 2013
    digitalpoint, Jan 24, 2013 IP
  3. digitalpoint

    digitalpoint Overlord of no one Staff

    Messages:
    38,334
    Likes Received:
    2,613
    Best Answers:
    462
    Trophy Points:
    710
    Digital Goods:
    29
    #23
    deathshadow - I'm curious what you think about the new version of vBulletin (vBulletin 5) from a UI, speed and everything else point of view... Maybe we should have stuck with vBulletin?

    http://www.vbulletin.com/forum/
     
    digitalpoint, Feb 20, 2013 IP
  4. deathshadow

    deathshadow Acclaimed Member

    Messages:
    9,732
    Likes Received:
    1,999
    Best Answers:
    253
    Trophy Points:
    515
    #24
    Worse than 4 in most regards -- your upgrading to 5 was just as painful as the switch to Xenforo -- it's a step sideways, not a step forward. You rightly pointed out my displeasure with 5 here -- and that's why the switch to Xenforo was so annoying as it was more of the same crap addressing NONE of the actual problems!

    Your years I've said they have giant brass monkey balls charging money for vBull -- and systems like vBull, Xenforo and phpBB often leave me wondering if the people making them know the first blasted thing about actually making websites.

    Let's take the vbull you just linked to -- on the page there's 23k of plaintext -- but easily half of that is content cloaked due to some 'gee ain't it neat' javascript vbulleshit. Even accounting for that and the two dozen content images that are nothing more than avatar thumbnails, how exactly can one explain 125k of markup -- three to five times what should be needed? Only answer is either ineptitude or ignorance on the part of the developers. Ignorant bull like HTML 3.2 slapped into a 5 doctype -- Ignorant bull like abusing lists around headings and tabular data -- Ignorant bull like javascript only markup in the markup instead of generated by the scripting -- Ingorant bull like broken/nonexistent scripting off behaviors -- ignorant bull like clearfix or classes like 'left' -- Ignorant bull like 90 to 200 bytes of classes on elements that probably don't even need to EXIST -- SPAN with endless classes and non-breaking spaces doing list and padding's job (NOT that I'm a fan of that useless idiotic 'tag cloud' crap), hidden inputs that aren't even inside forms, buttons that aren't even inside forms, static scripting inlined in the markup, that stupid malfing x-jquery-template bullshit, shall I go on?

    ... and that's before we get into the inaccessible fixed metric fonts, broken pathetic attempt at responsive layout that doesn't work -- though at least that skin doesn't have illegible color contrasts and TRIES to be semi-fluid.

    There's a reason I stopped coming here after the last update on vBull-- and this update I only hang around because I've got scripting blocked and enough user.css in place I might as well have written my own skin.

    Again, Xenforo is a step sideways in terms of actually being a forums. When you make a new release it helps to FIX things that are wrong -- I saw no indication of that as it's still the broken inaccessible poorly written crap it was under vBull.
     
    deathshadow, Feb 21, 2013 IP