test
Viewing 9 posts - 16 through 24 (of 24 total)
  • Author
    Posts
  • wiery Friend
    #319677

    I’m really liking this template too. The menu is the coolest part. 😀

    dawidt Friend
    #319735

    What about download? I’m allmost sure that yesterday Ja Halite was available for download. Unfortunatelly, today there is just an empty space, someone stole all 1 file 🙂 Please, give it back 🙂

    Hung Dinh Friend
    #319742

    Download is now available guys

    shapes;147716Hello,

    I noticed a little bug with my iPhone…

    On the first load of the demo on my phone, I had the special template, but I clicked on a button that switched the layout to full view like with a web browser. And after that, impossible to come back to the special phone template.

    Best regards,
    David aka Shapes

    This has been fixed. You now have a button to switch back at the bottom of the site

    jimg Friend
    #319765

    phill luckhurst;147744How about you design a template like that. I would be very interested to see how it is acomplished.

    From the tests on this template I have done locally, speed does not appear to be an issue.

    Phil: its done by reviewing and designing template solutions that adhere as close as possible to best practices set out by the big guys. Yahoo has become a central repository of such info lately, I would highly recommend taking a look at their guide Best Practices for Speeding Up Your Web Site. Some suggestions – like use a content delivery network – are a bit far off for a small site, the others are spot on when the goal is delivering a fast user experience.

    Local performance is always quick – the real challenge is what the site will look like when it goes into production. Yahoo’s research is interesting regarding the impact of page response time – 80% of the time is spent making http requests to fetch all the components in the html document (images, scripts, stylesheets) and 20% to actually download the html document. They posit that the biggest performance gains can be made by simply reducing the number of http requests. Detailed information can be viewed in this post what the 80/20 rule tells us about reducing http requests .

    But – shouldn’t everything be in the browser cache? Yahoos research also uncovers some interesting facts regarding caching, why you should consider far future expire records, and interesting finding on the percentage of visitors who actually have cached content. See Browser Cache Usage – exposed! Bottom line – tune for the empty cache experience, and benefit from browser side caching.

    Back from theory to practice – even looking at yahoos comments, they do make a disclaimer.

    Disclaimer: Design imperatives dictating visual richness need to be weighed against this request-reduction goal. When you need visual richness, additional steps can be taken – aggregating JS files, using CSS sprites, etc. – but visual richness does tend to run counter to a slender HTTP request pipeline.

    So – taking a balanced approach is best: one can create a visual richness like what JA products provide, but use tried and true optimization techniques to deliver a fast user experience.

    if you look at the majority of http requests in JA Halite, you will find that they are from css images. Using a tool like firebug and yslow will help point to where these files are. Then create a sprite file to combine these images – one file for vertical images, the other for horizontal. Using the online CSS Sprite generation tool tool will take your zipped images and create a single image + a coordinates map so you can reference the specific image inside the combo sprite image. One would then change out the css references to the individual image file and replace it with the sprite + coordinates. Result? no loss of visual richness plus a faster user experience.

    Past this, optimizations like combining css & javascript files can be done manually, or via 3rd party solutions using mambots. As a reference, a good example of a speed optimized joomla template would be afterburner, a free template. It uses sprites, eliminates mooltools, and is quite an interesting study in optimization techniques. It is spartan in contrast to what joomlart offers, so I would not consider it a competitor but as a good example on how to optimize a joomla template.

    Phill Moderator
    #319773

    I agree local tests will always be fast. But in my case I have a setup that allows me to simulate various download speeds from a virtual remote server. This has shown halite loads very well and with a little effort on my part I can improve this further.

    Afterburner is indeed fast but the tricks it uses to make it so fast really affect the way a user can customise it. Using sprites in your suggested way will also have the same effect. Most Joomla users simply do not have the skills to make effective use of sprites to speed up their site. Most simply edit the images as required. There is a balance between speed and usability and in this respect I think JA has got it right. One of the biggest sites in the UK ( http://www.bbc.co.uk ) gets the same yslow grade as halite so it cannot be that bad.

    A skilled developer will do many of the things you suggest once their site is more complete.

    jimg Friend
    #319787

    phill luckhurst;147863Afterburner is indeed fast but the tricks it uses to make it so fast really affect the way a user can customise it. Using sprites in your suggested way will also have the same effect. Most Joomla users simply do not have the skills to make effective use of sprites to speed up their site. Most simply edit the images as required.
    [/quote]
    I chose Afterburner as an example with the caveat that it may be too spartan for some. If the goal is reducing http requests, then using sprites in an intelligent manner to reduce the number of images a template loads is not reserved for just pros. A significant number of images (buttons, backgrounds, borders) do not change, regardless of the color scheme selected. Including a sprite version of the image AND the source images allows those who want to customize maximum flexibility.

    There is a balance between speed and usability and in this respect I think JA has got it right. One of the biggest sites in the UK ( http://www.bbc.co.uk ) gets the same yslow grade as halite so it cannot be that bad.

    I think we are comparing apples to oranges here. evaluate the two sites on a performance basis and you’ll see a different picture:
    The two are quite different sites:

    Results from:
    http://analyze.websiteoptimization.com/


    JA-Halite bbc.co.uk
    ---------- ---------
    Total HTTP requests: 128 25
    Total Size (bytes): 637235 285934

    Total HTML: 1 1
    Total HTML Images: 19 12
    Total CSS Images: 82 4
    Total Images: 101 16
    Total Scripts: 9 6
    Total CSS imports: 17 2

    Notice the larger number of http requests from the JA-Halite. bbc.co.uk does use sprites, does use a cdn and probably has a bigger hosting budget that most of us do, so I wouldn’t say this is a fair comparison. Also, I wouldn’t rely on the grade point as a value to compare against sites – its purpose is to evalulate how the single site does against a set of rules and help the user discover how to improve the end user experience.

    The goal of my post was to point out that known and proven techniques are out there that can make a significant improvement in the pageload performance. Joomla templates have a bad reputation of being bloated (google it) and it doesn’t have to be that way. One can deliver a visually rich template AND not require a large number of http requests.

    nefar Friend
    #319788

    How about some comparrisions between JA VS it’s competition — like rocketheme, garvick et cetera.

    New framework Vs their framework. 😉

    Hung Dinh Friend
    #319794

    jimg;147877I chose Afterburner as an example with the caveat that it may be too spartan for some. If the goal is reducing http requests, then using sprites in an intelligent manner to reduce the number of images a template loads is not reserved for just pros. A significant number of images (buttons, backgrounds, borders) do not change, regardless of the color scheme selected. Including a sprite version of the image AND the source images allows those who want to customize maximum flexibility.

    I think we are comparing apples to oranges here. evaluate the two sites on a performance basis and you’ll see a different picture:
    The two are quite different sites:

    Results from:
    http://analyze.websiteoptimization.com/


    JA-Halite bbc.co.uk
    ---------- ---------
    Total HTTP requests: 128 25
    Total Size (bytes): 637235 285934

    Total HTML: 1 1
    Total HTML Images: 19 12
    Total CSS Images: 82 4
    Total Images: 101 16
    Total Scripts: 9 6
    Total CSS imports: 17 2

    Notice the larger number of http requests from the JA-Halite. bbc.co.uk does use sprites, does use a cdn and probably has a bigger hosting budget that most of us do, so I wouldn’t say this is a fair comparison. Also, I wouldn’t rely on the grade point as a value to compare against sites – its purpose is to evalulate how the single site does against a set of rules and help the user discover how to improve the end user experience.

    The goal of my post was to point out that known and proven techniques are out there that can make a significant improvement in the pageload performance. Joomla templates have a bad reputation of being bloated (google it) and it doesn’t have to be that way. One can deliver a visually rich template AND not require a large number of http requests.

    Dear jimg,

    My personal opinions on this (not to be taken as official remarks) :

    1. If JA is to give stripped templates, then enhancing the designs would be a problem to users.
    2. If JA is to combine CSS and JS in one go, again that would be a hindrance in further development.
    3. CSS Sprites will be difficult to understand and manage for users, as it would restrict customization. Besides, not all would be well acquainted with css sprites.
    4. Providing pre-optimized templates would be as good as giving product sans the source code.
    5. Providing fast loading basic designs, will result in sacrifice of all possible customization, which i dont think would be in favor of anyone.
    6. JA Templates are fully loaded with scope for customization

    What’s brewing :

    1. New framework is a big step in performance related issues.
    2. Plugin to provide option to compress JS and CSS (with option to select the files), the JA Compress Plugin is under development, BUT as is the backdraw of such options, it wont work in all scenario’s and everywhere, and would be released with such warnings, post testing.
    3. Selective loading of css files is under consideration, but again needs restructuring the template and needs through testing.
    4. Implementation of css sprites and its results are under implementation and hope to release a template with such options.
    jimg Friend
    #319800

    Hung,

    Thank you for your response – I’m certain that your efforts will result in a better product for all users – beginner and advanced.

Viewing 9 posts - 16 through 24 (of 24 total)

This topic contains 24 replies, has 15 voices, and was last updated by  jimg 15 years, 2 months ago.

We moved to new unified forum. Please post all new support queries in our New Forum