If you're reading about bloghacking, chances are you've got at least one blog and you've tried out a couple of the hacks presented here. You've probably also got a broadband connection, a big monitor, the latest version of your browser and plenty of patience. But what about your readers? Do they share your enthusiasm for the customisations, neat tricks and experimental tech on your blog? Or will delays and errors cause readers to click-off, never to return? As an avid bloghacker, I'm as guilty as the next blogger of squeezing more into my templates than I probably should. But here, I offer some tips and tools for finding out just how far you can push it.
The first step is to know your readership. Whether they're accessing your blog via low-powered handheld devices or high-end corporate machines makes a difference - as does the security environment they're likely to be in. (Webstats can tell you a lot, particularly the user agent string.) With this insight, you can design for your "baseline": the bandwidth, screen-size, plug-ins (eg Flash? Java?) and security settings (eg cookies?) of a reasonable worst-case reader.
Next, not all browsers respond in the same way to warnings and errors, so you need to catch them all. The World Wide Web Consortium's validator service is a very thorough scrutineer of your page. While not every warning or error warrants your attention, it is worth a review. Remember: this service is not for the faint-hearted or those with fragile egos!
http://validator.w3.org/check?uri=http%3A%2F%2Fblogfresh.blogspot.com
A simple check of file size is a good next step. Remember to check post pages, the main page and archives separately as they'll have different sizes. While "reasonable size" depends on your content and readership, it's worth remembering that every 10KB will take 2 seconds over a dial-up connection. I'd suggest that a 100KB page is too much and most blogs should manage coming in at around 30KB. If you're oversize, you may want to consider splitting up your headers (style and script declarations) into separate files. (Blogspot members may need to look at hosting these files off-site.)
A more nuanced view is afforded by the good people at webperf.org. I've been using their free service to monitor website performance for five years. You get a great breakdown of the different page elements - scripts, images, style sheets - being loaded. This is great for diagnosing problems on your blog and benchmarking against others. (I've included a bookmarklet at the end of this post to help with this.)
http://webperf.org/breakdown.html?URL=http%3A%2F%2Fblogfresh.blogspot.com
You should note at which point your page makes a call to your webstats provider (eg sitemeter). Up front? Half way? At the end? This is because readers who click-off (or abandon your blog during loading) before this point won't appear in your stats. If you've got a heavy page, moving your webstats script further up the template could help you pick up any missing traffic. Alternatively, as an exercise you can temporarily move it to the bottom to find out just how many readers will hang around for the page to finish.
Of course, this tool views your blog like a search engine spider: it will load but not run the scripts on your page. This is not necessarily a bad thing. In fact, I recommend viewing your blog with javascript turned off just to see how it will appear to search engines. But still, it artificially limits the page rendering time, especially if you're bold enough to run scripts that invoke other scripts.
While most of us preview our blogs on our own machine (possibly trying out two browsers), we're still not recreating that baseline experience. Why? Well, for starters our machines tend to be newer, with later browsers. Secondly, the images and other files are probably cached locally, giving a snappier response. Thirdly, if you've ever crafted a bloghack or even just coaxed one into life, you'll be more forgiving of "minor" glitches, teething problems and other shortcomings.
What's the answer? For me, it's relatives' broken computers. The only silver-lining to another malware infestation call-out is the chance to see how my blogs render on first-load on an underpowered PC at 800x600 over a 4KB/sec dial-up, using an ancient browser with a host of "helper" plug-ins and bizarre security settings. Even better is when I see the reader reaction - what do they tolerate? What do they notice? What's the first thing they click on or read? This is invaluable feedback and frequently an eye-opener.
Once your blog's pages are being delivered to (and rendered by) the browser in an acceptable way, you can start focusing on optimising the content-side of blog usability to attract and retain readers.
What else can you do to understand reader expectations? What about other tools for monitoring or profiling reader experiences? Or does RSS and syndication obviate the need for this? Feel free to share your tips and thoughts on getting the most into your blog.
-Greg.
Bookmark: Drag and drop this link onto your links bar to make a button. When browsing, to profile the current page simply click the ClockThis button.
ClockThis
Filed in: blogtech, blogger-hacks, how-to
-Greg.
More, and more I appreciate web standards, because it speeds up web development time.
Zeldman (Yeah, the same guy who designed a couple of blogspot template) lobbied IE and Netscape to join a standards consortium during the browser war that ensued, which of course led to them joining W3C.
His book on standards is enlightening .
Valid or not, it is good to check for errors on codes:
HTML Validation
CSS Validation
Link Validation
Feed Validation
As for usability, I find Vincent Flanders, to be a more flexible person on the issue than the rigid Jakob Nielsen. You can judge for yourself, who is better on their debate.
I was particularly interested in the link checker. I figured it would just do a recursive wget thing but it looks like it actually had a go at parsing out the URLs. It wasn't very successful with the dynamic javascript invocation (ie a javascript that writes out another javascript), as can be seen by this URL:
http://del.icio.us/feeds/json/tags/'+del_user+'/'+anchor+'?sort=freq&count=100
Still, it at least parsed the "outer" javascript!
Cheers,
-Greg.
ps Do you know of a site that will fire off a URL to all these different services and combine the results?
I only know of one, and its not free: netmechanic
hope that helps