back to article Google advises flushing your website

Steve Souders is the sort of person who spends his Saturday afternoons measuring website load-times. "You might watch football games," he says. "I watch websites load." After founding the Exceptional Performance Group at Yahoo! - an effort to improve website speed from the websurfer's point of view - Souders now does similar …


This topic is closed for new posts.
  1. Bryce Prewitt
    Thumb Up

    I'm not happy with the drag they impart either.

    In fact, I'm so not happy it, I block them.

    Thanks for supporting Firefox + Noscript, Google!

  2. Michael Xion

    I propose a new Web Standard...

    ..called Web 3.0 where a page loads super-fast because it contains only the HTML that is immediately required. The user then retrieves additional information into the page by clicking on embedded links that only deliver the content as requested, rather than frontloading ALL of the myriad JS, CSS etc calls at initial delivery and thereby slowing the page render speed...oh...hang on..wasn't that Web 1.0? At least it will work in Netscape 0.93b :-p

  3. ratfox
    Paris Hilton

    Whatever floats his boat

    I'd guess few people would notice the difference, especially if their connection is less than optimal.

    PH, because she floats my boat...

  4. iamapizza
    Thumb Down

    Let's also...

    ...serve websites as simple plaintext documents, get rid of images and ensure that each page is only fifteen words long. I'm sure this'll go a long way to overcome the coding ineptitude that the search crawler developers face.

    "Boo hoo hoo, the page is taking too long to load, let's just not fix our code to increase the timeout, let's get them to fix their websites."

    lol @ Michael Xion

  5. Anonymous Coward
    Anonymous Coward

    Hows about -

    not linking to google for trackign cookies or other sites for stats , wouldn't that improve page load because then you only load one site?

  6. Marvin the Martian

    wait or display?

    Do I read this correctly as saying, you better have a page's layout jump around for a few seconds (because it is re-displaying the changing input as it's receiving) than to not show much for those seconds?

    I'm not sure I agree with that.

  7. Anonymous Coward


    "You might watch football games," he says. "I watch websites load."

    Your medal is in the post mate.

  8. jake Silver badge


    "You can read all his esoteric CSS advice here (PowerPoint)."

    This muppet is trying to tell the world how to fix the Web, using Powerpoint as a method to convey his ideas?


    'nuff said ...

  9. Anonymous Coward
    Anonymous Coward


    @ratfox: Try loading the Times website with and without javascript enabled and compare how long you have to wait. Then say whether you think this is or isn't an important issue.

  10. Roger Greenwood

    me too, Bryce

    I also block adsense, analytics, doubleclick, quantserve etc. in firefox. Having said that, some sites are quicker in chrome without the blocks!

    Particularly anoying is how slow it is to go back a page on theregister. Perhaps you guys could get a copy of the book yourselves.

  11. Anonymous Coward
    Anonymous Coward

    "Tim 2.0'Reilly"

    that's really tickled me, thanks Cade

  12. Chris Ovenden

    @ratfox, you don't get it

    According to the article, "Only about nine per cent of the time the user waits for the page to load is spent getting that html document". Yes, some of the rest of the time is spent requesting external assets like CSS and javascript, but those are likely cached after the first page - the rest is processing time; nothing to do with how fast the user's connection is.

  13. Anonymous Coward

    right... web developers are only now just catching on (actually no - being *told* ) to do what normal developers dealing with terminal I/O have been doing for a very long time under heavy I/O loads.

    Seriously - what planet are web developers on!!

  14. Damien Thorn

    And another way

    ive always wondered why they do it the hard way, you add your code to a server, the server on request sends this code to the browser, in a manner the browser understands then the browser builds the code into a page. Surely thats 1 extra step.

    Imagine you write your code, and the browser requests the page, the server "streams" the page itself already pre-rendered. it rendered it virtually seconds after you uploaded it, it holds the exact image as a ... file. The browser then displays that, its faster because the browser has not much to do other that request the file, and show the file, its not trying to interpret every single tag etc, that has already been done for it, all it has to do is show the image. Dynamic content can be loaded AFTER the image, and even PRE-loaded so when you click a [more] the rest of the mores are then called, thats what the browser should be doing now too. and it doesnt put any extra strain on the server, probably reduces it actually because for a 4kb page the server has to send 4kb (an image might be 2kb)

  15. Charlie Clark Silver badge

    Flushing the code?

    Is it just me or is this an unfortunate and possibly misleading use of the word? I associate flushing with cleaning, ie. flushing the cache. I realise the use of the word seems to come from one programming library of another but surely the generic is streaming or buffering? Allowing the start of a message to be sent before it has been completely assembled?

    As for load time performance in general this is browser and server together. The more browsers can optimise for standards, the faster they can render. They may also run a lot faster if they can open more parallel connections. One weird suggestion about improving performance is solely related to the way IE handles multiple HTTP requests to the same server. Of course, avoiding multiple HTTP requests altogether (they are expensive and slow to set up) would be another technique. Or extending HTTP to allow it to be "always on" for the modern interweb.

  16. Dave

    @Charlie Clark

    The term comes from the fact that what is actually being done, is flushing the output buffer.

    I did read recently about some effort that had gone into testing the effect of different CSS selectors, and the long and the short of it was that for real world web sites, it made no real difference - most testing had taken place with massive documents, that were too large to be realistic.

  17. Anonymous Coward

    It seems a good arena for further research

    Most web based things seem to increase the granularity of stuff at the developer/designer stage and I suppose that it is good, required and important for it to be so.

    But does that need to be the published or sent to server information?

    Compiling server end (image to app) might be a good thing for the future especially if standards are upheld? With diversity of hardware, software accessing the web/net maybe a robust rethink to strategy is timely?

    Whatever happened to those refresh/change local content things in a way that avoids having to download a whole page for the sake of a few paragraphs of text?

    And, on a personal level, why flash white on content change? Flash grey instead or better still don't flash at all?

  18. This Guy


    Ever used C's "fflush" command in C to 'flush' (empty) an I/O buffer to wherever it's supposed to go?

    Same thing.

  19. Quantum


    > right...

    > By Anonymous Coward Posted Friday 3rd April 2009 09:21 GMT

    > web developers are only now just catching on (actually no - being *told* ) to do

    > what normal developers dealing with terminal I/O have been doing for a very long time under

    > heavy I/O loads.

    > Seriously - what planet are web developers on!!

    Real nice. What planet am I on? I am on the real estate development planet. You criticize without offering any suggestions, and that's the lowest form of life, AFAIC.

    Anyway, I don't get how to actually implement this flush. Researching it I find at least 15 different opinions on how it should be implemented, and as I don't know PHP I have no perspective.

    First of all, where does it go? In the header of each page, or in the body? Second, would it actually benefit a site like ours:

    Third, any perspective on best practice implementation?

  20. Anonymous Coward
    Anonymous Coward

    @ Quantum

    Judging by the state of that page, it's probably just as well you don't know PHP.

  21. Anonymous Coward

    @ Quantum

    "Real nice. What planet am I on? I am on the real estate development planet. You criticize without offering any suggestions, and that's the lowest form of life, AFAIC."

    Criticism - You are not really a web developer, your site is crap and some of the grammar and punctuation is awful, especially that shite about "Whatsoever ye would that others should do to you, do ye even so to them."

    Suggestion - Pay a REAL web developer to put your website together properly (possibly someone with English as a first language)

  22. Quantum

    Take a Leap

    > By Anonymous Coward Posted Sunday 5th April 2009 09:04 GMT

    Not interested in your opinion Coward, as your attitude is foul, as always.

  23. Anonymous Coward
    Anonymous Coward


    "Quantum Equities, LLC is proud to present its new web site. We are designing it to be helpful for tenants, prospective tenants and space users, and investors. We owe special thanks to Delphi Real Estate for the design and construction of the site."

    I would be quite suspicious that your site and your affiliates sites all follow the same format with the same graphics, layout and fonts

    The design is terrible - it looks like someones first attempt, thrown together by someone that had no idea what business you were in and no idea what you were trying to convey...

    did no one think to review competitors sites and look at strengths weaknesses of the design etc.

    few initial suggestions for free

    drop the background image

    drop the rotating gifs

    implement a decent menu - a 30 second web search could find you many decent examples and some of those will be free to use.

    and most important vary the design and at least try not to look like a fraudster with several registered companies

  24. jake Silver badge

    @AC 09:04 & @Quantum

    The phrase:

    "Whatsoever ye would that others should do to you, do ye even so to them."

    Is probably supposed to be a cross between:

    "Do unto others as you would have them do unto you."


    "An it harm none, do what ye will."

    Either way, it's a guaranteed ticket to one of the seven circles ;-)

    Quantum: ODFO, there's a good chap/ette. This place is for talking tech with the adults, not spoon-feeding children good web design practices. Try your local JC or Poly. I'd suggest searching the Usenet archives, but judging by your above commentary, you're not old enough for that yet.

  25. Quantum


    Just as I thought; no one here knows the best way to implement flush.

    So you nastily substitute criticism to try and cover up. Lowest form of life...

    You two -deserve- to be left behind.

  26. Anonymous Coward


    If you're a demonstration of what it's like at the front, I'll stay behind, thanks.

    You can't seriously expect not to receive criticism when you make comments regarding best practices and web design, having a site like yours. Why not take the feedback and use it to improve your lot, rather than trying to get into petty arguments with those who clearly know better?

  27. jake Silver badge

    @Quantum 13:47

    "Just as I thought; no one here knows the best way to implement flush."

    Ok, you webmaster, you. Please show us all up with your own site's perfect implementation.

    "So you nastily substitute criticism to try and cover up. Lowest form of life..."

    If you can't take the heat ... Side note: We aren't interested in doing your homework for you.

    "You two -deserve- to be left behind."

    Most of us have already graduated. I seriously doubt you ever will, not with that attitude.

This topic is closed for new posts.

Other stories you might like