Centralized info for web page optimization. I wonder at the efforts spent on page speed improvements, versus letting tech improve and publishing tools handle much of the implementation details.
I think web masters, sites and developers should focus all their efforts on content quality,and check a box that the page loads fast enough now.
To the contrary of this, I consciously admire optimized page loads, that flicker into existence. But this is secondary to the "knowledge payload" of the site.
The most critical components of page load optimization are simple tick-the-box items when setting up your application that need to be automatic for every web engineer, but aren't. They are mostly not a matter of having the tech improve. They are mostly not involved with "publishing tools" at all, insofar as by the time your CMS is in the picture it is way too late.
Partial checklist:
It takes something like four lines in your web server config to turn on gzip. Did you turn on gzip?
It takes perhaps ten lines of code to splat all your Javascript and CSS files into one file at deployment. If you were running Rails, this would be about fifteen keystrokes or so. Did you do it?
You should be loading your static assets from 2-3 hosts to maximize how many can be loaded in parallel. This is, again, two minutes of work and doesn't even require you to actually possess multiple hosts since you can just create multiple DNS entries pointing to the same one. Did you do it?
Don't forget CSS spriting; static assets from multiple hosts is great, but when you minimize your dependencies AND their size the loading time appears even faster.
1) may not be as slow as getting adverts from the same server
2) depends on use, if nearly all images are provided as thumbs as well and most images are viewed full-size in the end then this can save both the overhead and the size of the thumbnail get
3) requires additional searches for discovery of dotted quads, why make this search manual when if can be done more efficiently in the background?
3[b!]) sadly flash banners may be paying for the site, but I'd agree apart from that
1) Maybe, but at least you have control over it now. A lot of sites, for me (with slow Ireland internets...), seem to be slowed down by between 5 and 15 seconds by external adverts.
2) True, but I've rarely seen this be a win (for me, YMMV).
3) What? I don't understand, explain.
3[b! emm typo oops]) Why do they need to be flash when a simple image is just as good? Need animation? gif images have supported that since roman times. No need for hefty flash for these. Lots of sites (again YMMV, maybe I just go to all the retard sites) have useless crap that supposedly makes the site look prettier. Well, it doesn't and it makes it slow. Throw it out and go for user experience. (You can still make a site pretty without making it slow)
Sweet! An open source project that I have (kind of) contributed to before made the list! I wrote some code for Pylot to support dynamic fields (in my case it was my ASP.NET "__VIEWSTATE" field, but it could be adjusted to support any dynamic field). I don't think my contribution has been committed to the trunk, but Corey added me to the list of contributors anyway, so that was nice :)
I love how google constantly contributes to the entire internet community in spite of the lack of immediate returns. I mean it takes some serious cojones to bet on the long term with stuff like this. I really hope companies start following this lead in the future.
just for the record since nobody else mentioned it: if you find this kind of material interesting/relevant, you may very well check with "velocity" conferences and their presentations (which are available online--or were for last years conference)
http://code.google.com/speed/articles/optimizing-php.html
http://groups.google.com/group/make-the-web-faster/browse_th...