TechnologyTales.com

I am in the process of further linking the content of hillwalking blog into the fabric of other parts of my website. To date, this has included adding a list of all the posts to the site map and a dossier of latest entries to the site’s welcome page. One things that started to crop up were a series of warnings of the form:

Cannot modify header information - headers already sent by…

The cause of this was my calling a PHP script that was part of my hillwalking blog installation and this caused Bad Behaviour to be invoked. The result was that a HTTP header was being sent in order to create a cookie after one already had been sent to display a web page. A spot of conditional coding and the use of an extra flag resolved the problem.

Apparently, there is another surefire way of getting the same result: whitespace before or after the <?php … ?> tag in an included script. The cookie issue is one that I can understand but it does seem strange that an attempt is made to send HTTP header information when the latter arises. It causes loads of questions, though…

When I first started it, my online photo gallery started out simply as a set of interlinked HTML pages. Over time, I discovered frames (yes, them!) and started to make use of JavaScript to make the slideshows slicker. In those days, I was working off free webspace provided by my ISP and client-side scripting was the only tool that I had for enhancing functionality. Having tired of the vagaries of client-side scripting - the browser wars were in full swing and incompatibilities reigned supreme, I went with paid hosting in order to get access to tools like Perl and PHP for server-side processing; their flexibility compared to JavaScript was a breath of fresh air to me and I am still a fan of the server-side approach.

The journey that I have just described is one that I now know was followed by a lot of website builders around the same time. Nevertheless, I have still held onto JavaScript for some things, particularly for updating the DOM as part of making the pages more responsive to user interaction. In the last few years, a hybrid approach has been gaining currency: AJAX. This offers the ability to modify parts of a page without needing to reload the whole thing and that has generated a considerable amount of interest among web application developers.

The world of AJAX is evidently a complex one though the underlying principle can be explained in simple terms. The essential idea is that you use JavaScript to call a server-side script, PHP is as good an example as any, that returns either text or XML that can be used to update part of a web page in situ without the need to reload it as per the traditional way of working. It has opened up so many possibilities from the interface design point of view that AJAX became a hot topic that still receives much attention today. One bugbear is efficiency because I have seem an AJAX application lock up a PC with a little help from IE6. There will always remain times where server-side processing is the best route and that needs to balanced against the client-side and vice versa.

Like its forbear DHTML, AJAX is really a development approach using a number of different technologies in combination. The DHTML elements such as (X)HTML, CSS, DOM and JavaScript are very much part of the AJAX world but server-side elements such as HTTP, PHP, MySQL and XML are also very much part of the fabric of the landscape. In fact, while AJAX can use plain text as the transfer format, XML is the one implied by the AJAX acronym and XSLT is used to transform XML in HTML. However, AJAX is not limited to the aforementioned technologies; for instance, I cannot see why Perl cannot play a role in place of PHP and ASP can be used for the same things.

Even in these standards-compliant days, browser support for AJAX remains diverse, to say the least, and it is akin to having MSIE in one corner and the rest in the other. Mind you, Microsoft did introduce the tools in the first place but they used ActiveX and Mozilla created a new object type rather than continue this method of operation. Given that ActiveX is a Windows-only technology, I can see why Mozilla did what they did and it is a sensible decision. In fact, IE7 appears to have picked up the Mozilla way of doing things.

Even with the apparent convergence, there will continue to be a need for the AJAX JavaScript libraries that are currently out there. Incidentally, Adobe has included one called Spry with Dreamweaver CS3. Nevertheless, I still like to find out how things work at the basic level and feel somewhat obstructed when I cannot do this. I remember perusing Wrox’s Professional AJAX and found the constant references to the associated function library rather grating; the writing style didn’t help either.

My taking a more granular approach has got me reading SAMS Teach Yourself AJAX in 10 Minutes as a means for getting my foot in the door. As with their Teach Yourself … in 24 Hours series, the title is a little misleading since there are 22 lessons of 10 minutes in duration (the 24 Hours moniker refers to there being 24 lessons, each of one hour in length). Anything composed of 10 minute lessons, even 22 of them, is never going to be comprehensive but, as a means for getting started, I have to say that the approach seems effective on the basis of this volume. It has certainly whet my appetite for giving AJAX a go and it’ll be interesting to see how things progress from here.