I just did a test on web.com.ph on the website http://www.webpagetest.org/ and the result when it comes to response time and rendering time was great so I signed up for hosting with them )

Thanks for the tip!

Cheers

Fredrik

You seem to think that I am constantly having a dig at Total Cache..? yet if you read what I have actually said, I actually state that in my opinion it is the best cache plugin out there. However this does not alter the fact that it does not work correctly for everyone, which was the reason I posted the article about manual entries in htaccess files ! After all , the more options people have open to them, the better eh..?

I am not one of these idiots on the net that reads something , takes it as gospel & then plasters it all over his own site, stating it’s the best thing since sliced bread! Anything I write about, I test first on this site, that was what this site was originally designed for, a test case / guinea pig & yes there are certain aspects I have not written about, probably because I don’t feel I am knowledgable enough on that subject, to be able to advise my readers correctly.

Anyway, although I may not have mentioned it in my articles, I have minified my java script, but not my style.css as it is in a constant state of editing hahaha (I’m forever tweaking), I also setup & tried a free CDN which was an abismal failure & just slowed my site even more. I am now discussing with Timthumb creators about moving the cache folder that stores all their resized images, to the wp-contents folder,

I have an idea that if I can manage that, then the CDN side of Total Cache (using a sub domain on my own server), may well work as these images are hard copies & not created on the fly. As I mentioned previously, the problem I had when I last tried it, was the fact there was no way to tell Total Cache that there was a folder full of images tucked away in the wp-includes folder (seems a stupid place to put it from my point of view).
Oh and before I forget, Google Adsense..? Yes I would love to find some method of optimizing Adsense as it is slowing my page load by 1.62 seconds. The best I have found so far is to use a plugin called datafeeder which replace the bulky java script code with a single line of php code, but it still slows the page load times.

Regards Steve

PS can we keep any response a tad shorter or I’ll have to dedicate an area of my site solely for legnthy discussions hahaha

- The majority of shared hosts do not allow PHP based compression, shared hosting accounts dominate the marketplace as that’s what most care to invest in. W3TC compresses *once* and caches the compressed version and sends that until it expires. Making it compatible with the majority of shared hosting accounts and freeing resources to generate pages on cache misses, when comments occur etc. The technique you’ve put forward does satisfy some tests, but it will cause your server to buckle under load, which is what I hear time and again from users before using W3TC as their hosts complains to them about CPU usage. AFAICS, a remedy for this common user affliction is not mentioned in your tutorials.

Looking at webpagetest.org for this site, I suspect that this technique is also responsible for the nearly 1 second time to first byte, which is definitely a metric that I encourage users to pay attention to. Values like ~0.2s are fantastic (this requires the page to actually be cached on the server obviously). Clearly generating a page with PHP for every visitor is slower than allowing Apache (web server) to return the page to the visitor from the disk. Why do i care? Because this value affects the total time to start render, which is where the user experience comes into play. As an aside, there are lots of plugins out there that increase response time (bad) and your recommended method of compression often causes memory usage issues for some users, coupled with the fact that PHP’s gzip compression cannot be used in concert with the “flush early technique” with is something do it yourselfers typically implement to encourage an almost instant page render in the browser. AFAICS, implementation of the “flush early” technique or advice on how to reduce time to first byte is not in your tutorials.

- On this site in particular is not a good example of this case, however, if you were like most of the sites I’ve seen, they have often more than a handful of CSS and JavaScript files that would indeed benefit from the principles you’ve shown, but once again, the principle you’ve presented is circumstantial and incomplete. Those users with multiple CSS and JavaScript files can with W3TC, minify (concatenate, remove white space, comments etc and compress the resulting files) and then cache those results. Telling users to use mod_deflate to compress text based objects upon request is also a drain on resources since there is no caching there and ultimately under load response times will suffer. Hosts sometimes have issues with these, while others will do this for you (they do the caching on their side). Mileage varies. What I know for certain is that for a site like this one, there is a positive affect realized when the JavaScript files that are not used for navigation or other essential interactivity are moved to the or at least after so that they do not have to be completely downloaded before the browser can render the page. W3TC allows users to tweak issues related to page render (and therefore user experience) without modifying theme templates. AFAICS, there is no mention of CSS / JavaScript minification in your tutorials.

- Believe it or not minification of HTML in a given page actually produces significant (10-30%) gains in the compression that is ultimately realized via the technique of page compression you suggest or that in W3TC. W3TC does this for the users transparently without breaking screen readers or AdSense targeting code etc. The point of compression is to reduce file size to improve user experience via shorter download time. With HTML often a snappier page render is realized as well. AFAICS, there is no mention of this HTML minification in your tutorials.

- One of the larger wins for WordPress is the Object Cache, it’s basically a “runtime” cache that allows WordPress to generate new pages faster. Obviously this is not something the user interacts with directly, but it and/or database caching, improve user experience on a site with high traffic and lots of interactivity and lots of newly generated pages as a result. Faster response times are proven to make web sites more successful/profitable.AFAICS, there is no mention of how a user can implement this performance technique in your tutorials.

- Content Delivery Network (CDN) or at least increased pipelining (to improve progressive page render speed) of pages, is not made mention in of your list. Not all tools can detect every CDN, so most users fail this, but at least with YSlow it’s possible to change settings that allow users to see their true score (even if they are not benefiting from the low latency network of a true CDN). Anyway, with W3TC, you can create a subdomain on your own server (and use it) with multiple CNAMEs as well, and for image heavy or sites with lots of objects to be downloaded to overcome the max’ connection issues of browser and improve the “waterfall” download by having various types of objects downloaded in parallel to render the page faster. It seems to me the “self-hosted CDN” technique is one you could have in your tutorials as it allows users to improve page render for free, AFAICS, it’s not in any tutorials.

- This site also has AdSense, but your tutorials provide no guides on how to implement AdSense in a manner that does not block page render. That might be an interesting tutorial for your readers as it is a common affliction.

Side note: Lots of users have multiple themes today, that is one for desktop and other for mobile. W3TC allows users to optimize the above for all of those themes at once with all of the same benefits discussed above.

There are other subtleties, but these are the ones that come to mind at the moment. So I suppose the word optimized is in the end a loose term as one can look at lots of different tools in different ways to see if a site is consistent with the policies put forward by the creators of that tool. Or they can look at the various aspects of their site/server and respectively and optimize those pieces that matter to them. My plugin exists because most users need a quick fix or want to address more issues than those that satisfy test tools; what’s more there are fewer steps in deploying W3TC and achieving some performance wins than there are in DIY methods, which is why I bothered to create the plugin. You’ve made your preferences clear. Your readers may want to check out the work of Steve Souders or at least http://developer.yahoo.com/performance/rules.html for more information on some of the techniques I chose not to mention above or that were not already in your tutorials.

Great tutorials and thanks for helping the community.

Wow Frederick, do you often make a habit of coming out with such ridiculous statements.
Hell it must have been just pure luck that the entries I placed in my htaccess file to deal with “expires headers” & “file caching” just happened to work eh..? Oh & enabling file compression by whatever means is seen as optimizing ones site, by everyone apart from you it would seem.

Before coming out with anymore ridiculous statements, you may like to take a look at this article, http://www.my-blog-review.com/cache-plugins-wordpress/ which clearly shows the entries made in my htaccess file & at the bottom of the page also shows a screenshot of the test results carried out at http://www.webpagetest.org/ which I would think that (4 x A’s + 1 x F qualifies as optimized) especially as I reduced my site load times from 8.5 sec to an initial load time of 3.3 sec & repeat load of time of 1.6 sec (not optimized eh..?)

So thank you for your offer to optimize my site Frederick, but as I have achieved these results without the additional load of extra plugins, I’ll stick with what I have thanks.