I’m helping them deliver their collaboration and social media products to their customers, and couldn’t be more excited! More details to follow as I learn the ropes!

]]> http://mikeletulle.wordpress.com/2010/04/30/change-of-direction/feed/ 0 Mike http://mikeletulle.wordpress.com/2010/02/18/using-yui-and-json-with-vignette-collaboration/ http://mikeletulle.wordpress.com/2010/02/18/using-yui-and-json-with-vignette-collaboration/#comments Thu, 18 Feb 2010 16:30:00 +0000 Mike Letulle http://blog.full-leverage.com/?p=98 ]]> I’ve been doing some integrations of some of the YUI widgets with Collaboration and I really like it. They have so many useful widgets that are easy to snap in once you get some basic fundamentals down. (YUI2 Examples , YUI3 Examples) I decided to share some of the important things I learned here to give back a little to the blogosphere (I sure get a lot of help from it when I’m doing development work).

For the purposes of this post I’ll walk through some of how I integrated the YUI3 Drag and Drop Portal example with Collaboration. I created an ability to create portlets and page grids that one can drag and drop desired portlets onto very much like one can do with the RSS feeds in that example.

The YUI ( and many of the newer web APIs like JQuery) requires knowledge of working with the JSON format. For those that don’t know it’s a nice way of wrapping lists of objects with key attributes described that can be easily parsed. For example if I wanted to describe a collection of 3 lists of RSS feeds it might look like this:

{"list1":[{"id":"ajaxian","min":false},{"id":"ynews","min":false}],
"list2":[{"id":"slashdot","min":false}],
"list3":[{"id":"daringfireball","min":false}]}

Each list in the collection has a name (list1, list2 etc), an id, and a minimized state.

My integration required that these JSON formatted text strings get set and read as custom attribute values on objects in Collaboration.

Here are a few key things I learned about working with JSON strings:

1) The JSON object names cannot start with a number NOR contain periods

Because I was associating objects to grid sections, the natural thing that I wanted to use as a name for each item was the OID or object ID which looks like ’1.11.7862′ . I painfully discovered that I could not work with the JSON strings until I replaced the periods with underscores and prefixed the whole thing with non-numeric characters (I used ‘id’)

So I ended up building my ‘feeds’ variable like so:

var feeds = {
<ISPECT>while(portletsIt.hasNext())
	<ISPECT>TagObject portletObj=(TagObject) portletsIt.next()</ISPECT>
	<ISPECT>String oid=getOID(portletObj)</ISPECT>
	<ISPECT>String oidNoDots=oid.replace(".","_")</ISPECT>
	        'id<ISPECT>showText(oidNoDots)</ISPECT>': {
	            id: 'id<ISPECT>showText(oidNoDots)</ISPECT>',
	            title: '<ISPECT>showName(portletObj)</ISPECT>'
	      }<ISPECT>if(portletsIt.hasNext()),</ISPECT>
</ISPECT>

2) You need to URL encode the string before you can set it as a custom attribute value:
I grabbed a handy JavaScript URL encoder/decoder to do this from www.webtoolkit.info

So after my javascript code builds the proper JSON string  I set it as an attribute value like so :

var modifyObjUrl="ModifyObject?object=<ISPECT>showOID()</ISPECT>
&attrvalue_of_Grid+Mapping="+UrlCoder.encode(feeds);

3) Use the tag method ‘getStringLiteralQuotes()’  to present the string value properly in Javascript code:

var cookie= new String("<ISPECT>showText(getStringLiteralQuotes(
getAttributeValue("Grid Mapping")))</ISPECT>");

Iterating over objects defined by JSON on the server side

Yesterday I realized that I wanted to be able to iterate over the objects defined by JSON text in a custom attribute value on the server side. There are no out-of-the-box tag methods for working with JSON so I decided to write a custom one by adding a custom method to the CustomerReport.class extension hook.

The first thing I did was look for a good open source Java utility for working with JSON and I quickly found this nice library at www.json.org . I downloaded it, compiled to a .jar and tried to import it in CustomerReport.java but it kept  saying that package org.json did not exist. After a lot of headache I discovered that the real reason it wouldn’t compile is because that same package is already part of the base installation, and it was colliding because my classpath pointed to both. So after getting past that fun,  I made a  method called ‘getJSONIteratorByKey’ that takes the JSON text and the key as inputs. It looks like this:

    public Iterator getJSONIteratorByKey(String JSONText, String key){
    	 Vector v = new Vector();
    	 try{
	    	JSONArray a = new JSONArray(JSONText);
	    	JSONObject jo = null;
	    	String s = "";
	    	int len = a.length();
	    	 for (int i = 0; i < len; i += 1) {
	            if (i > 0) {
	            	jo = a.getJSONObject(i);
	            	s = jo.getString(key);
	                v.add(s);
	            }
	    	 }
    	 }
    	catch (JSONException e){
    		 v.add(e.toString());
    	}

    	return v.iterator();
    }

So now from a template I can grab a JSON formatted string from a custom attribute value and if one of the object keys is the object id or logical name, I can easily get an iterator for those object ids:

<ISPECT>String jsonText = getAttributeValue("Grid Mapping")</ISPECT>
<ISPECT>Iterator it = getJSONIteratorByKey(jsonText,"id")</ISPECT>
<ISPECT>while(it.hasNext())
	<ISPECT>String s = (String) it.next()</ISPECT>
	<ISPECT>showText(s)</ISPECT><br>
</ISPECT>

I hope someone out there finds this helpful!

A system like that which can surface and then roll up just the needed information to and from the specific users brings with it many advantages. But the actual interface for capturing that information has to be optimized as well. In many cases allowing the user to simply write whatever they want in a Word doc or rich-text editor may be perfect. In more cases some sort of template in which they just fill in fields is more appropriate. But in many situations the kind of information that should be on that small part of the whole document is very structured in nature and much better served by filling out a simple web form with form controls like constrained-choice drop-downs and check-boxes presenting the user with predefined choices to predefined questions.

I once learned about a customer’s auditing process and it basically went like this:

• Auditor informs company they will be doing an audit at a certain date
• Auditor comes in and audits those areas
• Auditor issues a findings document indicating they have x number of issues they need to deal with
• Responder from company creates an ‘issues document’ from this and emails it around to all people that might have input on any of the issues
• Issues document gets edited by several users simultaneously and they all email them to the responder
• Each issue needs to be responded to in basically one of three ways
  • “has been remediated” (referencing document number that shows this)
  • “will be remediated” (and date of expected remediation plus any notes)
  • “will remain non-compliant” (along with notes)
• Responder receives emails from users, repeatedly bugging them for responses as needed and creates response document
• Responder sends response document to Auditor

They were constantly losing information and repeating work due to this process.

It became obvious that what they should have been doing is creating and working with individual “issue” objects. I prototyped a solution for them on the Vignette (now Open Text) Collaboration platform that allowed them to create secure online audit workspaces for each audit. They were created by cloning a predefined templated workspace that had things like

• a common discussion thread around the audit
• a shared calendar with audit dates on it
• placeholders for the auditor’s findings document and the official response document (which  had its own approval workflow)
• an issues folder that lets one create specific issues from the auditor’s findings document

The issue object was based on a folder (the platform lets you extend the base object types like folders, documents, and discussions ) so that it could contain important reference materials, be subscribed to, be emailed into, and all the other things one can do with a base folder type in the platform. But I also gave it the custom attributes that it needed like ‘Due date’, ‘Assigned To’ ,’Response’ (with choices like ‘Has been remediated, Will be remediated, etc.),  ‘Notes’, ‘Reference #’, and ‘Expected Remediation Date’.

So someone would go through the auditor’s findings and then create an issue out of every issue on it. Each one would be assigned to the person that was best to answer it. That user would receive an email with a link to that online issue workspace that indicated its due date. The user would receive a preconfigured amount of email reminders until it was completed, as each issue object has a process defined around it.  The response document was then automatically created by a display template that rolled up the information that each user responded to into one nicely-formatted text document.

One of the more advanced and important features of the Open Text platform which we used here is that the workspace templates allow one to pre-map access rights so that the content in the cloned workspace  automatically has the correct permissions set on each of the cloned sub-folders and items.

With this solution in place the customer could respond much more quickly and accurately and with much less time spent by its valuable employees. Every audit has its own secure workspace with a full audit history of every response to every issue. This is also, I think,  a perfect example of optimizing a collaborative business process.

So far among other projects, we have worked on the following:

• Helped build some public-facing websites and their content management system for a division of the World Bank called CGAP (Consultative Group to Assist the Poor)
• Created a digital asset management system for a large hospital chain
• Enhanced a hospital policy management system that Mike Letulle originally designed while with Vignette
• Optimized the ability for users to comment on articles on a highly trafficked public website and the ability for the content managers to moderate those comments (which often go over 1000 comments/article)
• Added multimedia capabilities to the main CGAP website (videos, slideshows, and podcasts)
• Enhanced the implementation of a large consultancy’s intranet and extranet which provides online client engagment workspaces as well as helped transition the maintenance of the implementation from one company to another by training an offshore team

I’ll add more about some of these projects in subsequent posts.

Time to dust off the old blog. Shame on me for going 16 months between posts, though I am happy to share that is has been due to having too much billable work to do non-billable things like write blog posts and build my own websites. A high-class problem. However I am (once again) going to try to start making the time to share things that I think would be of interest. Wish me luck!

If you own these platforms and wish to leverage them better, or are simply looking around at technology options for improving competetiveness, please feel free to email Mike.Letulle@Full-Leverage.com .

Stay tuned for much more about specific solutions and offerings.

I’ll decline to list all the reasons why. There are enough blog posts out there that rant about how cool iPhones are. But what I think a lot of people underestimate is the storm of change they are bringing to the business world. Apple’s new iPhone Apps rolled out last Friday. I was just checking the available business applications and saw that Oracle has a free iPhone application that extends its business intelligence dashboards, KPIs, and alerts to the iPhone.

Salesforce.com has something similar. Now that is vision! I mean when most software companies are finally starting to grasp that they should probably consider a decent format option for mobile devices accessing content over the web, these guys offer rich interactive client apps on the world’s most popular new mobile platform.  Sure a lot of BI solutions can already send alerts to people via email and even SMS if certain thresholds are hit, but nice graphs and charts on your iPhone? Now that’s slick!

I seriously think iPhone’s use by the business user is going to start to snowball now. They took a lesson from the facebook apps and put it on the iPhone. The interface is ideal for a small-size device to view varying amounts of content, and the easy upgradeability of the firmware makes it easy for Apple to keep it up-to-date and competitive. And now it’s inexpensive and faster! What is their stock trading at now???…

 The idea of properly using technology to not only get around the isssues and enhance the process, but also to make customer relationships more profitable while you are at it has been somewhat of a devoted cause of my manager at Global360, Phil Ayres. He has a Blog dedicated to these concepts. He and I have been hard at work over the last several months building a blueprint (or solution template – call it what you want) on top of our Case360 product which is what I currently specialize in.

Global360 just posted the recent 1 hour webinar we held in which Phil presents the business value and architecture followed by my delivering a 20 minute demonstration of the solution.

I just looked up about when these tests are supposed to be scheduled. Interestingly enough, I read on wikipedia : “RMTs must be performed between 8:30AM and local sunset during odd numbered months, and local sunset to 8:30AM for even months” . And March is month number three – an odd numbered month the last time I checked.  So why am I seeing this test around 6:30am on odd numbered month conveniently right when all these primary voters are getting their kids ready for school before they go vote?

I’d be willing to bet that if you figure out who has the most to gain by voters voting with fear, you’ll find who is behind the EAS test I saw this morning.

Or maybe I’ve seen too many X-files in my day…