If you've been to Adobe Labs lately or keep up with the blogosphere you've probably heard the buzz about Spry. It's a new AJAX framework from Adobe.

It lets designers build rich Internet applications with little or no programming. The Adobe-Macromedia combine has been pushing RIAs for years and its vision is just now moving into the mainstream.

Adobe had three goals when creating Spry: open access, easy use, and enable innovation.

Open Access
Propriety tags or server-side code weren't used in the Spry framework, instead the capabilities of existing HTML tags were enhanced. Spry has been released under a BSD license and it's freely available on the Adobe Labs site.

Easy-To-Use
Spry was created with designers in mind. It uses regular HTML tags, CSS, and JavaScript. Very little coding is required, although its capabilities can certainly be enhanced with coding. The syntax reminds me of how ColdFusion integrates into Web pages as a tag-based language. You use the tags you're already familiar with, but with new properties. Adobe certainly accomplished its goal; Spry is very easy to use.

Enable Innovation
Enabling innovation is one of those warm and fuzzy goals. In this case what it's trying to do is give developers the ability to create rich applications while still allowing them freedom of design.

All that sounds great, but what is it really? Spry boils down to a couple of JavaScript libraries that you include in your Web page. Just a couple and you can be adding dynamic interactive content to your site. It can't be that easy you say, but it really is.

Loading Data in Spry
The two main JavaScript libraries that have to be included are xpath.js and spryData.js. The xpath.js is the Google implementation of XPath 1.0; spryData.js is the Spry library itself. Once these two files are included in your page Spry is enabled. Of course just enabling Spry doesn't do anything. You have to load some data and display it. Once you're Spry-enabled you can dynamically load data. This is one of the core features of the framework and is done by creating a Spry XMLDataSet. Right now XML is the only type of data supported, but there are plans to support other formats in the future. The sample below shows the code needed to enable Spry and load some data:

<script type="text/javascript" src="xpath.js"></script>
<script type="text/javascript" src="SpryData.js"></script>
<script type="text/javascript">
      var mydata=new Spry.Data.XMLDataSet("mydata.xml","/people/person");
</script>

Two pieces of information are need to load data into the Spry, the URL of the file you're loading and an xpath statement to select from the data that's returned. I used an XML file in the sample, but it could just as easily have been a dynamic file such as ColdFusion page that returns data from a database.

When using ColdFusion make sure you include a <cfcontent type="text/xml"> or Spry won't read the data. This has tripped server people up (include me) the first time they tried to retrieve data from a ColdFusion page. The URL can access data from anywhere, but it will generate a warning or may disallow loading data from other sites depending on the browser security settings. The xpath statement lets you filter the data that's retrieved. This is required even if the only thing you're selecting is the root node of the XML. Using the same data source you can select different XML nodes or search for a specific node or series of nodes in the XML.

Displaying Data in Spry
Once we have the data we want to display it. For this example I'm assuming we return a list of people with firstname, lastname, and e-mail as the fields. To output data you have to define a Spry Region. This is done by adding the attribute spry:region="dataset" to the tag such as a div that encloses the area using the dataset. This lets Spry know that it should update this area of the page if the specified dataset changes. The data output is very similar to data binding in Flex and has similarities to ColdFusion except with curly brackets and colons instead of hash marks. For instance, to output the first name returned in our data set the syntax would be {mydata::firstname}. Much like a ColdFusion query this would display the firstname field of our dataset from the first row of data. What we need to go with this is the ability to loop over a dataset when multiple data items are returned. This is done using the spry:repeat tag. But for a couple of minor differences, the syntax should feel familiar to you if you've been using ColdFusion. You don't use an output tag, you just embed it in the tag you want to loop with. The example below demonstrates how the tag would be use to display a table with our data:

<div spry:region="mydata">
<table>
<tr>
<th>First Name</th><th>Last Name</th><th>Email</th>
</tr>
<tr spry:repeat="mydata">
<td>{firstname}</td><td>{lastname}</td><td>{email}</td>
</tr>
</table>
</div>

This example is creating a table, but the same technique can be used to create other dynamic elements such as a select box by putting the SpryRepeat in the option tag. It makes chores like have multi-select boxes where the items in the second box depend on the item chosen on the first a breeze. You can update the second box without reloading the page by changing the URL of the data source for the second dataset. For instance, I could add a button that updated the data source to my table example:

<input type="button" onClick="mydata.setURL('mydata2.xml')">

When you click on the button the XML data would be read from a different file and the table would automatically update without reloading the page. You're not limited to text, you can bind the src attribute of an image to a Spry data field and the image will automatically be updated when the data changes.

Other Capabilities
The dynamic data capabilities are the core of the Spry framework, but this is still a pre-release of the technology. Though it's not covered in the documentation, there are two other Spry libraries included as part of the demos: SpryAccordion.js and SpryEffects.js. SpryAccordion.js contains code to produce an accordion-style menu. SpryEffects.js contains code for animation effects such as Move, Size, and Opacity. I expect these libraries will be expanded to include a full set of dynamic UI components and animation effects.

Limitations
All this dynamic functionality sounds great and you may be tempted to jump right in and start making Spry-based Web pages, but there are some limits you should keep in mind when working with Spry. Spry is AJAX-based and as such has the same limitations as AJAX. It can't work without JavaScript. If a user has JavaScript turned off your page won't work, not only that but the data might be inaccessible. If you were to browse to the sample used in this article with JavaScript turned off, you would just see a table with the variables in it. There would be no way for you get to the data.

This situation is of concern to the Spry team too and the latest release includes documentation on how to make your page usable even if Spry can't run. Since we're dealing with JavaScript you face the dreaded JavaScript-compatibility issues. Spry has been tested with Firefox 1.5 (Windows and Mac), Netscape 7.2 (Windows), Internet Explorer 6 (Windows), and Safari 2.0.3 (Mac). If you're not using one of these browsers, don't count on Spry working. It may not work at all or you may only have partial functionality. You also have the accessibility issues that come with dynamic Web content. Most screen readers do a poor job or just don't work with dynamic updates in Web pages.

Also keep in mind that Spry is in pre-release. The code could change at any time and it's not guaranteed to be backwards-compatible. There are already differences between the May and June releases. For instance, the format of the Spry Region attribute was changed from spryregion to spry:region. A minor difference, but it broke the code I had written and I had to update the examples in this article.

Where Is Spry Going?
The future of Spry is up in the air at this point. This is a pre-release of the code, not even an alpha or beta. The Spry team isn't event sure where it's going. They want feedback from the community to help with Spry's direction. Since many of the developers of Spry are on the Dreamweaver team I expect it to be included in Dreamweaver at some point. I also expect to see many more UI components and hopefully a component standard so that the community can create its own custom Spry components. This is just the beginning of a very interesting technology with a lot of potential. Hopefully we'll see it fulfill that potential soon.