How do you know what needs to be tested after you modify a line of code?

I've asked myself this question many times both as a developer and as a manager. Well - I just got an idea :)

Today we have code coverage tools, that can tell you what gets executed when you run something (like tests). Imagine that when a Quality Engineer would run a test, the system would automatically detect all the methods that were executed for that test, and it would log this in a database for further use.

Then, when a developer commits some new code into the versioning system, it would automatically identify the methods that the developer modified, and mark the matching test cases for regression.

What do you think about this?

On this blog, we have a page where we want to list all our job openings. But these job openings already exist on different websites, and we do not want to maintain the list each time the list changes.

So the first thing that crossed my mind was the HTMLDataSet from Spry. It allows you to build a Spry DataSet from an HTML Table.

As the list of jobs is on another website, we had to build a proxy to get it. We do it this way because of cross-domain scripting restrictions.

So after about 20 minutes of "hard" work, we have managed to implement it.

http://myadobe.ro/angajam/ (again, in Romanian) retrieves the job list from http://bestjobs.neogen.ro/?1=1&bj_nl_p=show_company&idco=29890&clid=1 through the proxy: http://myadobe.ro/includes/jobsproxy.php.

Bogdan

Today I went even further and improved the help pages by bringing in comments from LiveDocs.

So right now, when I look at a specific help entry, I do not only get the "default" static content, but I also get comments if I'm online. So it looks like this:

(notice the comments at the bottom of the page)

How did I do this? Quite simple, using Spry. Again - the code is not perfect, but it works:

I opened the "help.js" file that is linked from all the help interfaces from here:

c:\Program Files\Common Files\Adobe\Help\en_US\Dreamweaver\9.0\help.js

And I added these lines at the end of the JS file:

document.write('<script language="JavaScript" type="text/javascript" src="../../../../../../../Adobe/Adobe Help Viewer/1.1/includes/xpath.js"></script>');
document.write('<script language="JavaScript" type="text/javascript" src="../../../../../../../Adobe/Adobe Help Viewer/1.1/includes/SpryData.js"></script>');
document.write('<script language="JavaScript" type="text/javascript" src="../../../../../../../Adobe/Adobe Help Viewer/1.1/includes/SpryEffects.js"></script>');
document.write('<script language="JavaScript" type="text/javascript" src="../../../../../../../Adobe/Adobe Help Viewer/1.1/includes/getParams.js"></script>');

function unescapeHTML(str) {
    var div = document.createElement('div');
    div.innerHTML = str;
    return div.childNodes[0] ? div.childNodes[0].nodeValue : '';
}

window.onload = function() {
    var related_in = document.getElementById('related');
    if(related) {
        var webLink = related.getElementsByTagName('a')[0].getAttribute('href');
        var m = webLink.match(/http:\/\/help.adobe.com\/(.*)\/(.*)\/(.*)\/(.*)/);
        var locale = m[1];
        var product = m[2];
        var version = m[3];
        var page = m[4];
       
        related.innerHTML += '<div style="font-weight: bold; margin-top: 10px;">Comments - from livedocs</div>'+
        '<ul spry:region="dsCO" spry:repeatchildren="dsCO">'+
            '<li style="margin-bottom: 10px;">'+
                '<a href="{link}">{dc:creator} said on {dc:date}:</a>'+
                '<div>{description}</div>'+
            '</li>'+
        '</ul>'+
        '<button type="button" class="submit" alt="Add Comment" name="AddComment" value="Add Comment"  onclick=\'javascript:window.open("http://www.adobe.com/cfusion/livedocs/add_comment.cfm?filename='+page+'&book='+version+'&product='+locale+'&version='+product+'", "addcommentwin","directories=yes,toolbar=yes,menubar=yes,location=yes,scrollbars=yes,resizable=yes,status=yes");\'>Add Comment</button>';
        dsCO = new Spry.Data.XMLDataSet(" http://livedocs.adobe.com/cfusion/rss/livedocs.cfm?product="+locale+"&version="+product+"&book="+version+"&full=true&page="+page, "/rdf:RDF/item");
        dsCO.sortOnLoad = 'dc:date';
       
        var obs = {onPostLoad: function(ds) {
            for (var i=0;i<ds.data.length;i++) {
                ds.data[i].description = unescapeHTML(ds.data[i].description);
            }
        }};
        dsCO.addObserver(obs);       
    }
}

This way, I have added this feature to the Dreamweaver CS3 help, but you can add it to other products as well (as long as they use the Adobe Help Viewer) buy modifying the help.js file from their corresponding folder. For example, for the "Extension Manager" you would have to modify this file:

"c:\Program Files\Common Files\Adobe\Help\en_US\ExtensionManager\1.8\help.js" 

Enjoy!

Bogdan

So while I was playing around with the Adobe Help Viewer today, I realized 2 things:

1. The search results (at least) is built using Spry
2. You can modify the search results page and add results from other sources

So I said - wow! I can split this in 2, and add some results from Google on the right side of the results screen!

So I opened the results.html file from C:\Program Files\Adobe\Adobe Help Viewer\1.1\ and I started modifying it.

I modified the width and the float for the results list (in red is what I added):

Line 353: <div id="nav_content" width="46" style="width: 46%; float: left">

And added a new DIV at the end of the file, just before </body>:

Starting from line 463

<div style="width: 50%; float: left; border-left: 1px solid black; padding: 0 2px">
<p style="margin-top: 8px; font-weight: bold">Search results from Yahoo!:</p>
<script>
      var dsGG = new Spry.Data.XMLDataSet("http://search.yahooapis.com/WebSearchService/V1/webSearch?appid=YahooDemo&query={dsProducts::@name} {dsQuery::query}&results=20", "/ResultSet/Result");
</script>
<ol spry:region="dsGG" spry:repeatchildren="dsGG">
      <li style="margin-bottom: 10px;">
            <a href="{Url}" target="_blank">{Title}</a>
            <div>{Summary}</div>
      </li>
</ol>
</div>
</body>
</html>

As you can probably see, I ended up using Yahoo! search, not Google, because Google does not export XML results unless you have a client code or someth.

And the result is:

Bogdan

P.S. Ignore the CSS layout - it is just a proof of concept

P.S.2. It only works on Windows. I did not had the time to make it work on a MAC. Again, it's just a proof of concept

"Adobe is announcing plans to open source Flex under the Mozilla Public License (MPL). This includes not only the source to the ActionScript components from the Flex SDK, which have been available in source code form with the SDK since Flex 2 was released, but also includes the Java source code for the ActionScript and MXML compilers, the ActionScript debugger and the core ActionScript libraries from the SDK. The Flex SDK includes all of the components needed to create Flex applications that run in any browser - on Mac OS X, Windows, and Linux and on now on the desktop using "Apollo".

Developers can use the Flex SDK to freely develop and deploy Flex applications using either Adobe Flex Builder or an IDE of their choice."

More on this topic:

• The announcement: http://labs.adobe.com/wiki/index.php/Flex:Open_Source
• Digg: http://digg.com/tech_news/Adobe_Flex_Goes_Open_Source
• Podcast: http://www.podtech.net/home/podtech/2826/breaking-news-adobe-flash-fle
• More on Flex: http://www.myfeedz.com/articles/tags/flex

Bogdan

It is basically the same product, fully compatible with Dreamweaver CS3, but misses a few features like KTML, MX CSS Menus, and XML Import. The ASP.NET and PHAkt support were also dropped.

On the other hand, we have spent most of our time testing the product and making sure all remaining bugs were fixed.

From an entrepreneurial point of view, this is a dream come true. As InterAKT, we have spent a lot of time designing, developing and selling this product, and the fact that Adobe decided to continue its support proves that we were on the right track.

Bogdan

Great! One can now build an offline version of Yahoo! Mail.

Well... not exactly. The APIs give you access to almost anything BUT reading a specific e-mail. So you can get the list of e-mails, list of folders, etc, but then, if you want to view the contents of an e-mail, you are redirected to Yahoo!'s website.

So my dream of having an Yahoo! Mail that is both online and offline has vanished. Unless I pay for the premium account to get POP3 or access to the full API.

mail.google.com says: "We have identified the root of this problem and are still working as
hard as we can to find a solution. We want to make sure you know that
we haven't forgotten you, and we sincerely apologize for the ongoing
inconvenience."

More here: http://www.readwriteweb.com/archives/a_glitch_in_the_web_office.php

Bogdan

http://www.philterdesign.com/dev/myfeedz/

It's a tag browser that can be used to discover new things. To use it, type a tag name (ex: myFeedz) in the upper-right corner.

Fresh is one of the samples that come with Apollo. It is an RSS reader, based on yui-ext.
It uses Apollo's offline storage capabilities to remember the list of feeds you read (opml), and the articles you have already viewed.

Download it and give us your feedback. Make sure you install the Apollo runtime before Fresh.

The bottom line is that now I use myFeedz as an RSS leafer, and Fresh as an RSS reader :)

Bogdan

P.S. Did I mention that Fresh is 100% Ajax, and that you can modify it as you wish?

Some new features:

• Nested Datasets
• • Spry now understands nested data. No need for extra XML files.
    
• JSON Datasets
• • One classical way of sending data over the wire when developing Ajax applications is JSON. Now Spry datasets support those as well.
• HTML Datasets
• • If you want to explain web designers what a database is and how they can start using it, they get scared and run away. With the HTML Datasets, HTML tables can be treated as datasources. Even more, degradability issues can be addressed this way (datasource in the same file)
• Session Handling
• • When developing Ajax apps, one thing you should care about is the session on the server. Using this feature, you can control it.
    
• Form Submission
• • Use this feature to do programatic form submits.
    
• Paging Datasets improvements
• Effects rewrite
• • The effects were re-written to be more flexible.
    
• Radio Button Validation Widget
• • New validation widget
    
• Auto Suggest Widget
• • This is one of the most visible and common Ajax use out there. Now Spry implements it as well.

Bogdan

A simple way to install Flex, and give it a try.

Say I have IIS or Apache on my windows development machine and I want to develop in PHP. All I have to do is download the installer from php.net, then start writing PHP code and preview it in a browser. Same with ColdFusion, and other server technologies.

With the "Flex Module for Apache and IIS" you can do the same thing with Flex. Download the installer, go through the wizard, and that's it.

After installing it, you can go and modify the MXML and the AS code in the samples, hit "refresh" in your browser and preview the changes.

Now - of course Flex is a client technology, not a server technology, but bringing the two together within the same editing/compiling workflow is just great.

Enjoy!
Bogdan

• We started to integrate with the new company (Adobe)
• I'm a Director of Engineering at Adobe Romania, managing all extension teams
  
• We have created a few extension teams here:
• • Dreamweaver
    
  • Flex
  • Apollo
  • Experiments (myFeedz, JSEclipse)
  • Web Infrastructure
  • Globalization
• We have released a few technologies on Adobe Labs
• • JSEclipse - on 12/15/2006
  • myFeedz - on 2/15/2007
    

Stay tuned, more to come!

Bogdan

As you probably already know, this week we have sold our company to Adobe. To make it even more challenging, I'm getting married this Saturday. So Anca is going to be my wife as of tomorrow.

After that, we'll going to fly to Spain for our honeymoon in the morning after the wedding. We'll leave from the reception directly to the airport.

I'll post some pictures when we get back.

You've probably noticed that we've been kind-of quiet in the last period, and this is one of the reasons.

We have spent a lot of time on this deal, making sure that our current customers won't be disappointed. Even more, I think most of them will be happy, because we provide them with really good upgrade packages.

Four of our most successful products are now bundled together as "Kollection". This new bundle includes:

• MX Kollection Pro
• KTML
• MX CSS Menus
• XML Import Export

and will be sold for $399. Of course, existing customers get a special upgrade price.

myFeedz will be posted on Labs, PHAkt and JSEclipse will continued to be shipped as free products, and the rest of the products will be discontinued.

Our customers do not have to worry about the products they have already bought, as we will continue to support those on our forums.  

Everybody here at InterAKT is very excited about joining Adobe, and looking forward for new challenges and opportunities.

If you want to find out more on the acquisition, please visit our FAQ page.

And by the way - remember to take a look at our new website.

After about 10 months since we had the myFeedz idea, the service went public.

Here are some details about myFeedz:

myFeedz is like a personalized newspaper. It gives you the latest news, filtered by your interests. Still, it is NOT an online RSS reader.

It finds what's important from all the information out there and shows you what you need to read.

It learns from what you like and keeps up with your interests.

Don't forget to read the blog post describing what's new.

There are still some limitations, as we're still in beta:

• You cannot yet see any news if you're not logged-in
  
• We are not allowing too many users to create accounts, as we have to tweak the application to handle as many connections as possible

Bogdan

After months of hard work, we have managed to release what we hope is going to be one of the most successful product in InterAKT.

MX AJAX Toolbox is a Dreamweaver extension that enables AJAX development in Dreamweaver. Not only it has a perfect integration with Dreamweaver, but it is also based on one of the most revolutionary AJAX platforms, supporting features like degradability, web services and implementing lots of useful controls and widgets.

Here is a quote from one of our customers:

"This toolbox is nothing short of revolutionary!  I've been playing around with several of the open source AJAX libraries, but they can't compare with the simplicity of implementation that InterAKT's MX AJAX toolbox enables." Timothy Rand- MX AJAX Toolbox customer

And here are to top features:

• AJAX dynamic table
• Graceful degradation
• Categorize items with drag and drop
• Edit-in-place
• AJAX site wizard
• AJAX photo gallery

More links:

• MX AJAX Toolbox homepage
• Product live demo
• Whitepaper

After ~ 1 hour of de-bugging, I managed to pass all the problems and integrated the 2 products.

What I tried to do was putting some Spry regions and repeaters in a panel. The problems I encountered were mainly some Spry miss-usages of Arrays. There are a few places where they use "for i in array", and we inherited some Array prototype modifications. But I replaced the []s with {}s and now it's working.

Also, they attach onload events to the window object to initialize the regions. I had to move that in the page in <script> tag:

            <script>Spry.Data.initRegions();</script>

Anyway, I'm glad it worked. I'll come back with more details on Monday.

http://www.interaktonline.com/Support/Articles/Details/Sneak+preview+into+MX+AJAX+Toolbox.html?id_art=42

Also, we have written a nice white paper detailing different aspects of the product:

http://www.interaktonline.com/files/White%20Papers/MX_AJAX_Toolbox_WhitePaper.pdf

Bogdan

P.S. For those of you who don't know what the product does, MX Ajax Toolbox enables Ajax development in Dreamweaver.

But why are these websites loading so slowly?

The main reason is that for the website to function correctly, all the Ajax application logic must be loaded into the client. And sometimes, we're talking about a lot of JavaScript code...

On the other hand, we have the degradability concept that says that a website should function correctly even if JavaScript is disabled.

Well... ok, so the website will function if JavaScript is disabled, then again it will wait for the whole Ajax logic to load before actually functioning. This does not really make sense...

Moving one step further, I realized that if a website would degrade while the Ajax functionalities are loading, then you would get both speed and Ajax.

What does this mean: When you access an Ajax-based site, the site loads very fast, and all the functionalities are available for you (the Ajax part of the site is not yet loaded). Then, if you click on a link, it would degrade and the whole page would reload. In the meanwhile, the required JavaScript files are asynchronously loaded from the server, and once everything is loaded and we're ready to go, Ajax comes into action and the site is automagically transforming into an Ajax website.

So, using this technique you would get the speed of a "classic" website (and here I mean the 1^st page loading time), and the power of Ajax.

Does this make sense?

We've just updated our Ajax integration demos (http://mail.interaktonline.com/demos/ajax_integration2/) using this technique and it seems to work. Even on a dial-up connection, after a few clicks, the JavaScript is loaded in the client and the site is 100% Ajax.

And one more thing: If you are interested in the MX Ajax Toolbox, just let us know: http://www.interaktonline.com/demos/ajax/?Content__state=Register

Bogdan