Tags are written as a comma seperated list, can have multiple words and contain commas if encapsulated by double quotes. Hope someone finds it useful!

< ?php
    function explodeTagString ($subject)
    {
        // split the string, populates the $tags array
        preg_match_all('/[^,]*\".*\"|[^,]\s*[^,]+\s*[^,]/',
            $subject, $tags);
	
        // loop through $tags and trims spaces and double
        // quote (\") chars this is a bit of a hack (I am
        // sure it could be built into the regex above).
        foreach ($tags[0] as $k=>$t) {
            $tags[$k] = trim($t, ‘” ‘);
        }
	
        return $tags;
    }
	
    // A little test case
    $test_tag_string = ‘test tag, “x, y and z”, foo, bar’;
    $tags = explodeTagString($test_tag_string);
    print_r($tags);
    /*
        Should output:
        Array
        (
            [0] => test tag
            [1] => x, y and z
            [2] => foo
            [3] => bar
        )
    /*
?>

So far I have looked at Propel, Biscuit, TaniPHP, WASP, Symphony, Cake, Solar, PHP on Trax, Prado and a few more to boot. The truth is, I just cannot find one that ticks all of the boxes for me.

Many suffer from too little documentation, several seem to be too complicated for their own good, and to be blunt there are just too many separate products trying to be the same thing. Of them all, Biscuit caught my eye but I cannot remember why … perhaps because I like the name and I wanted a cup of tea by that point.

I think the problem for me is that most of the frameworks try to follow Rails too closely, and end up suffering as a result. Now - I may well eat those words after some more research - so don’t blow any fuses yet. I will keep the feedback coming as I find out more over the next couple of weeks.

I would like to think its about accessibility and useful, easy to use standards-based web applications. But I fear its not about that at all. Its actually about search. More specifically its about developers building services with the sole aim of getting bought by Google. Or Yahoo.

Nifty search and sort algorithms can only do so much, but people can do an awful lot more. People are the sorting algorithms of the future, making meaningful associations between online content. Thats why we are seeing a lot of things like Squidoo, Rollyo, ma.gnolia.com and del.icio.us: they all ask *you* to layer over and draw themes across otherwise hard to link web content. Enough people put the same content under the same tag, and a pattern emerges. A pattern that the aforementioned search-engine lot will be very keen to get their grubby little mitts on.

Now, I am not necessarily against this. As long as the applications offer utility and are easy to use then I am all in favor. But what happens when every developer vies for the same space … not all will survive. Now thats my worry. Whats going to happen when services dissapear (read ‘cash runs out’)? What will happen to my stuff that was in it? Will it be lost? Time will tell.

While thinking about it I had a brain-wave. I can think of a way of doing it. But I dont know if I should be able to do it. What I mean is, I could well be exploiting an oversight hitherto undetected by browser developers. Or maybe its not their bug. Maybe its Macromedia’s. The solution relies on Flash.

You see, as of version 6, it has been possible to persistently store data locally on the client machine via Flash ActionScript. That data is linked to the domain of the movie: only other movies’ from that domain can re-access it. So it is kind of sand-boxed.

But, I wondered, what if a Flash movie from example1.com was embedded in web pages at example2.com and example3.com (just like a lot of those Flash adverts you see). As expected the movie from example1.com still has access to data from that domain. Now, realise that Flash movies can be accessed via Javascript. It would be possible to have a callable ‘getData’ and ’setData’ function. Kind of like a ’stealth Flash cookie’.

Now, this only works if both sites embed the Flash file served from the same location. But its still kind of scary. Many many people (I dont have the figures) have Flash 6+ installed. Many will have Javascript enabled. As far as I can see, there is no indication to the User that local storage of variables is occurring. Who knows if this is happening already…

It strikes me that users are forever asked to come to our community-spaces, to learn how to use yet another set of tools and rules. But why? Look at the take up of blog software. Why are we not building tools to integrate into these environments, these places where people already feel safe, feel in control and are more likely to invest time in? Why generate yet another online identity for someone who is already investing a lot of time in an existing online persona?

I guess where I am heading is to suggest that those tools that I would have built into the next online community platform should instead be developed as plug-ins to blog software instead. Perhaps the place to start with this will be to define an underlying API on which such tools can rely.

More on this to come…