crmcog has moved!

Moved to http://crmcog.com

(for what it's worth)

Siebel: A Simple Log Analyser

 

Deciphering Siebel log files is a chore. The basic premise has not varied since a long time now:

1. Look out for all the errors logged in the file. Even if most of them are benign, there is a chance (bad word to use here) that they are contributing to a larger error/system crash
2. Keep a look out for the performance of various SQLs thrown at DB. If a transaction is taking a long time, that’s where you want to start
3. Verify whether any Siebel transaction outside the DB is taking more time than expected (e.g. in cases where you have a really complex search spec that cannot be passed on to the DB)

Although I wanted to automate this somehow, for a very long time I just relied back on “proven” methods:

• Open log file, search for execution times around the functionality that is causing the issue
• Search for errors – most likely around the section that is causing the issue
• Navigate back and forth a lot, so that rest of the people think I already have a feel for how Siebel works and doing something great
• Rely on grep, tail and perl (when working on UNIX and developer-friendly). Sadly, Windows does not provide it out of the box, and to get open source/free tools in a restricted environment is no joy

This changed when I took a look back at  where .NET 2010 is, the culmination of which is here. “COGSiebelLogaliser” is a tool to parse Siebel logs generated by a client or server, and spool files to display useful information about:

• Performance times
• Errors
• An included editor is used to open log files (faster to open and navigate through large files)

High level requirements:

1. Tested it in Windows XP, with .NET 4.0
2. Application should have write access to the folder from which it executes, and read access for the log files. It will not write back anything to the log files

Of course this has to be treated as Alpha version since it has gone through limited testing. If you have blind trust in executables supplied from any person, and courageous enough to test out new things, you can try the application here. The tool must be tested against sample log files in a local, isolated environment until you are sure what it does, and how it works. Source code is pretty amateurish, and has to wait for a later date to get released to the wild.

Siebel: Trace Only When Needed

"Trace" statements in Siebel script provides a good way to debug and know what indeed is going on in the background. It is common enough to use the following statements to do debugging in scripts:

TheApplication().TraceOn("c:\trace.txt", "Allocation", "All");
...
...
TheApplication().Trace("Committed to the change " + rowId);
...
TheApplication().TraceOff();

While this is easy enough, an additional flexibility to trace only when required works out better. This can be easily accomplished using VB Script (not Siebel, stand-alone).

1. Open Notepad and type the following lines

'Reference SiebelApplication object

Set oSiebelApp = CreateObject("TWSiebel.SiebelWebApplication.1")

'Start tracing with appropriate parameters. Uncomment one of the options
oSiebelApp.TraceOn "C:\sbl-trace.txt", "Allocation", "all"
'oSiebelApp.TraceOn "C:\sbl-sql-trace.txt", "sql", "all"

'Continue tracing until stopped
set wshell = CreateObject("Wscript.Shell")
wshell.Run ("%comspec% /c title COGSiebelLog & echo Tracing in Siebel is currently on. & pause"), 1, true

'Stop trace and destroy object
oSiebelApp.TraceOff
set wshell = Nothing

2. Save the file and close it (assume the file is called "COGSiebelLog.vbs")
3. Open Siebel client as always, navigate to the specific screen of interest
4. Double click on the VBS file to run the script
5. Perform the transaction that needs to be traced, hit 'Enter' in the command prompt of the running script after the transaction

You will find that the trace happens only when the VB script is running. You would still need to have 'Trace' statements in the script for any custom tracing though. A similar approach can be applied for SQL tracing as well (refer to the commented line in the above script).

x2engine: A Good Lead Management Tool

Readers would have kindly (or otherwise) taken note that my views of the CRM products are narrow, immature, and hurried at best. I want to create more disruption by yet another tool/product, and pray I will not meet anyone really angry on the way to office.

This time, I will reverse the order - just for the fun of it.

Summary

     + A good CRM product with lot of potential
     + Nicely done integrated interface, with lesser prompts to navigate throughout
     + Integrated collaboration features including message board, chat, forums
     + Easy drag & drop development, ala Sugar CRM and the likes
     + Open source [BSD license]
     + iPad ready :) - well, this from the marketing material. Can't test this unless some kind soul sponsors an iPad for me
     ++ Support for INR. Truly, I was left flabbergasted

     - Limited feature set - targeted at lead management. Can only guess that the core industry support will find its way if the product survives for a few years
     - UI needs some getting used to (notice the bottom-most links for History, Actions etc), and I was far from delighted to scroll up and down when I click something

     - Some of the features appeared buggy on my computer, nothing that dedicated support or my own tinkering couldn't solve (no harm in boasting about that anyway)

Since it is open source, and the development team kindly provides a demo version and a trial version to find out more, the following is of real interest only to lazy people.

Now for some details.

Technology, Installation and Works

x2engine uses PHP, works on a standard LAMP/WAMP stack. XAMPP is a good place to start if you are completely new, or if you don't want to invest lot of time in tweaking tens of configuration files.The installation itself is a breeze, a trusty old install.php takes care of everything, provided you already have a DB user and schema created. I was not too impressed with click-click installs until I came across another CRM product with a really complex list of instructions to setup (another story for another day). The product uses Yii PHP framework for what it's worth, and clearly I am not spending enough time to understand PHP frameworks a bit more.

Features

Clearly, the company does not intend to throw all features under the sun in this product - their words, not mine. The core functionality is around lead management, and it sticks to the core.  The features don't get in the way, and enable users to concentrate on their business. However, this may also imply extensive customization even if you are a mid-segment business with slightly complex processes. The little things in the application are surely the highlights:

• Having an integrated screen where you can subscribe to a contact, lead (Chatter!), or chat with your fellow reps (new Chatter!)
• Promote a contact and make her watchable at the click of a button
• Search encompassing all entities
• Ability to see others activities so that you are well prepared to blame them at any time

Usability 

x2engine is easy-to-use, does not require complex navigation through every link to gain access to data, and tries to do a lot of things in one screen. If you are left wondering - yes, all these are really good things for the users (those are the pesky people who use the application).

In-place edits is the most I missed, the UI is old school for most part though some of the actions like emailing, and display of allowed state transitions are intuitive.

Performance of the application has been indicated as a plus point in the marking materials, but I did not see any reason to get excited. On my machine (which is antique btw), the performance was comparable to Sugar CRM. Demo version was a real improvement, I have to find out where the configuration was not ideal (clearly, throwing more hardware at it is not the solution :).

Configuration

The product allows power users or developers to configure everything of interest through drag & drop functionality.

 

The drag and drop of fields is impressive, and you are not expected to stick to ONE layout. The customization options are fairly limited. Other than adding fields and UI elements, you could define state models, basic assignment rules and do some Google integration.

That's about it for now folks. I am sure you are enthused about this a bit - so check out more at http://www.x2engine.com/live-demo-server/, or download and do your own setup.

Documenting Requirements - A Few Thoughts

Even after being in development projects for more than a few years now, I am always not far from being awe-struck at every requirement phase we have had (well, almost). The process of gathering and validating requirements, and converting everything to specs that a developer can understand, and maintaining traceability through all the rewrites and changes - are the most critical phases that always excite and worry me. Documenting all these in a language that finds common acceptance across user groups (& probably, across cultures) is, well, crucial to say the least (ALM tools not withstanding).

How well do you indeed document requirements? There may be entire books written on the subject, but where should one start? The following summary and lists should help one think in the right direction. A lot of it seems direct and too easy - I would think so too, but there have been instances where I was too busy to take note in a few of the projects.

Structure and Language

Two things that cause the max impact to the requirement document, and simple in hind-sight. The structure is fairly straight forward - you just group user requirements into buckets, and order it in a way that makes sense to the stakeholders.

Evaluate each statement of a requirement against the following list to apply the "few good rules".

• Correct: Speaks for itself, doesn't it?
• Complete: Should not be open-ended, and should carry forward on its own weight to functional spec
• Avoid use of "including ...", "not limited to "
• Don't make subjective or vague statements
• Avoid use of 'TBD', but do note that this is not disallowed
• Clearly state factors that have caused the exercising of that option
• Describe the conditions that will remove TBD

• Consistent: Consistency in description, language, and functionality expectations
• Bounded: Measurable conditions have to be described and relative statements have to be avoided
• "System shall provide response time less than or equal to 5 seconds" is better than "System shall provide quick response times"

• Unambiguous
• You cannot have a requirement listed but at the same time excluding from scope

While the use of language becomes a hot debate in some circles, I try to apply simpler rules:

• Use of "shall" (vs. use of "will")
• Use of definitive sentences that are deemed complete. Do not use "etc.", "including ... " and such terminologies which leave things to interpretation or imagination
• Avoid usage of double negatives in a sentence
• Simple to understand and well articulated

Each requirement should be unique in the set of requirements identified for a project or a program, and be identifiable as such. Inputs from all stakeholders - business and technical teams, and consideration of external environmental factors (regulatory policies, organization strategy etc.) are equally necessary. At the same time do not include designs to address a problem or solutions in the requirement document.

Handling Changes

This is a separate section since this is what we dislike the most. While change management by itself can be a clearly marked process, do you articulate on how to carry forward these changes to the requirements documentation? An indication of versioning of requirements document, and scoping of individual projects that align with the overall requirements go a long way in bringing enough clarity here.

Important Sections

While customers are free to exercise right on the template structure, it is imperative for us to make sure it has a place for some important sections.

• Business Rules and Business process diagrams: While not separate sections, making them part of the core requirements, or wherever needed, will help everyone to stay on the same page. Especially in case of COTS products, implementing business rules will demand a big share of development effort
• Non Functional requirements: Everything that is deemed important other than the core business requirements - like performance of the application, security requirements, usability requirements and supportability requirements have to be articulated. Even if the customer does not have this practice, it is never late to educate them and include such requirements in order to bring in enough clarity to business expectations
• Regulatory requirements: This is applicable and really important for certain industries like Life Sciences that are subject to intense auditing. For a newly proposed system, or even for changes to an existing system, capture and document requirements around regulatory considerations
• Assumptions: Clearly categorise assumptions while documenting them

The above points should act as a starter, but can no way be deemed comprehensive. Documentation standards available from IEEE standards talk about some of these and much more - for software requirements and specs. Evaluate them if you have access to the artifacts - that will be time well spent on establishing standards ourselves. Two of the documents that I liked while browsing through:

• 830-1998: IEEE Recommended Practice for Software Requirements Specifications. http://ieeexplore.ieee.org/servlet/opac?punumber=5841
• 1233-1998: IEEE Guide for Developing System Requirements Specifications. http://ieeexplore.ieee.org/servlet/opac?punumber=5982

Top CRM Blogs

Chris Bucholtz has listed the top CRM blogs of 2011. Well, as you can make out from this blog - I am not a good creator, but do read up from time to time. Even if you don't care about the blogs listed here, you should go through them at least four times in four weeks before outright rejecting any of them. They take up precious little time, but have the potential to give you something new (in CRM ofcourse).

My own list of sites (note - not only blogs) is not as lengthy, but does keep me occupied for a couple of hours in the week.

• Paul Greenberg's blog: Though Mr. Greenberg dropped the organization I work for in this year's CRM idol :). His posts are informative, full of CRM with a touch of humor
• Enterprise Irregulars: I am fairly regular on this site (through my reader). There are a bunch of really smart people who talk about a lot of things (not CRM, or IT alone)
• DestinationCRM: Well, this looks like a 90s magazine, and website. But, is a good source of information
• CRM Outsiders: This started more because of my interest in SugarCRM, but has continued ever since
• CRM Daily: A few good posts about CRM information can make one's day

What do you think, are you a regular anywhere else!?

vTiger - A Serious Open Source CRM Contender

I had been intending to explore vTiger CRM for sometime now, and am glad I finally came to it. vTiger CRM has been out there for a while, and I had been hearing about it everytime I hear "CRM" and "open source" in one sentence. What follows are bits and pieces of my first impression, a few cool things and some not-so-cool stuff.

Product and the Company

• vTiger is a small company, one of the two offices is in my own home town (= Bangalore, yeah!). The product is based on the LAMP/WAMP stack, and guess what, Sugar CRM. One of those instances of open source projects introducing competing products? Yes, this is (or can be) good for the customer, but do I want to see this as the Sugar CRM CEO? :). Some of us, open-source proponents, might want to think that this leads to a better all-round product, since there should be mutual learning for the two streams! 
• The entire product is open source, thankfully there are no "different editions" (yet)
• While going through the various screens, one cannot help but compare how this stacks with Sugar. Since vTiger exposes all functionality, it has an unfair advantage in some of the below points (I only have access to SugarCE)
• While the CRM product is completely open source and available for download, vTiger can also be tried online. There is a on-demand version available for $12/user (check price list). Again, no editions, no limits - except for space - 5GB per client.
• The overall product is polished, and works well. I liked the UI finish and overall functionality better than Sugar

Features and Functionality

• UI is very similar to Sugar, but some of the features are a sure welcome. A brighter UI (with the colour buttons thrown in) is not helpful on its own, but adds to the lustre

• Entities like accounts can be assigned to individuals or groups. This is a far cry from person-based/team-based visibility, but still something powerful for small or mid-sized businesses
• As in other web-based CRM products, by default the user lands in the list view. Drilling down on the record will show detailed information for the entity. Associated information (like activities or contacts for the account) do not show up by clicking on another 'tab'. But getting individual entities like activities, contacts by clicking each time is a pain.

• An interesting feature is tagging. All entities can be tagged with any custom attributes, and these show up in a 'tag cloud'. Users sure will love this feature, anything to find out 'all accounts that are important for this campaign' in a single click :)

• Users of this application are assigned a role, while roles can have one or more profiles. Access to entity views or the individual fields can be pre-determined

• Analytical reports can be customized by end-user. It kind of provides "actionable-intelligence" since any record in the report can be drilled-down to see more details

Development and Maintenance

• As in other web-based CRM systems, vTiger provides an administration screen where UI and fields can be configured. Changing UI and adding new fields is real easy and intuitive

• One of the interesting features is the ability to take backups. You can schedule a backup to be taken locally (where the server runs), or to any location through FTP
• There is also a provision to build workflows. OOB you get to do a few things like create an activity or send an email, and is also extensible. Workflows can run immediately upon satisfying the given criteria or wait for a pre-configured/customizable time to get initiated
• vTiger is extensible like Sugar, and getting new modules (mostly) is as easy as importing the zip file within the application. vTiger has its own market place, contributions have been mostly from the product development team. Some of the extensions sure seem interesting (keep a look out here for more :)

With continued development, vTiger sure is set to become more interesting. Sufficient to get you to try it? - head on over to vTiger site and try the click-click installation or the demo!

SugarCRM: Working with Google Maps

There was a time when if you had to show how Siebel integrates seamlessly with the world outside, it was considered cool enough to create a mash-up with Google maps. Sad to say, I feel stuck in that time warp sometimes and this post actually confirms that.  The premise is simple - integrate with Google maps as seamlessly as you can, the problem simpler - I am still on page 21 of the PHP manual. That brings us to the next best thing - JJWGoogleMaps! Again, the best screenshots had been provided by the products that charge a fee (e.g this), they simply would have to wait :).
The steps to integrate maps using this connector is simple/not simple as it is demonstrated below:

1. You install module in the standard fashion to see a new 'Maps' module showing up in the main menu
2. Navigate to the Maps screen and select the entity to which you want to map 
3. Select an existing map / create a new map 
4. If you have not already done that, click on 'Geocode Addresses' in the Maps module. This will fetch geocodes for the addresses entered in Sugar CRM
5. Finally, see the map in Maps screen and have a blast (while you can) 
6. The catch before you begin the party - don't forget to check the statistics and keep your requests within limit of 2500 requests/day. Though this module claims to cache the geocoded addresses to limit these requests, you cannot start testing it prior to a customer demo (to avoid effects of Murphy's 101st law)

SugarCRM: Being Social

SugarCRM is one of the few CRM companies that made all the right noises before Social CRM was known by that name. As I started looking into what the product can do, it was natural enough to have high expectations. Now that I have touched the surface, you will get a chance to sneak-peak as well.
Social CRM, as they say, is customer's territory. As the fact goes, the customer does not give a damn about your beautiful website and the associated paraphernalia. So being Social has been to a large extent providing seamless ways of integrating the product with the external websites that are in the business of getting people to socialize. If the product can achieve that without inflicting pain on its users, it is indeed a welcome development.

I can confidently say Sugar CRM does not disappoint one, although I hasten to add it will not pass in flying colours either. I would have to warn you - all my experiments are confined to  the 'community edition' version. To do justice I would have to go through the enterprise/standard editions one day, but I am not doing any implementation now and I am too lazy to follow-up for a trial version :). Here follows some tidbits:

• LinkedIn integration stands out indisputably, to such extent that I wondered whether there is any advertising revenue involved! Accounts within Sugar CRM are easily located, and can be processed through whatever LinkedIn offers 
• There was a lot of gaga about the Twitter integration sometime back. Was quite dismayed to see that it is only present in higher editions. However I did find something on Sugar Exchange which does exactly that, except that it just didn't do it. Well I am quite proud of writing a Twitter connector in Siebel, one of these days I should be repeating that here - with far better customization capabilities
• As one can expect on an open forum, Sugar Exchange does offer connectors to a few of the well known 'social' platforms out there. But they seem to be nothing but 'connectors' or connecting frameworks, with all the heavy lifting yet to be done to have something that will be of some use. One of the examples I quote is the Jigsaw connector.
• I tend to think about (& catalog) the 'social crm' part separate from the 'enterprise 2.0' part (sorry, Oracle Social CRM). Integration with Box.net tends to fall into both spaces, and it has a free connector for higher editions (what is with these guys and editions - not enough money in open source, sigh!). But the best and simple example is the Sugar Feed dashlet on the home page. It lets the user type a message and let the world (his world) see - ain't that cool? In the below screenshot you see the user feed below the blog feed.
• There are also more than a few small players which provide enterprise 2.0 capabilities, and have provided connectors to them. For e.g Broadchoice, and Quontext, but did not interest me enough to 'actually' test them

That is it for now I am afraid, watch this space to not see anything radical happening to Sugar CRM implementation space (you didn't miss the NOT there, did you?).

SugarCRM: Reporting

One big thing for any CRM application (or any enterprise application - sooner or later) will be the reporting capabilities of a given application. This is where custom applications often suffer, while packaged applications exhibit a hue of colours denoting their respective capabilities. Siebel (there we go again!) was heavily invested into Actuate reporting capabilities for online transactional (a.k.a OLTP) reporting. More recently we have BI Publisher replacing Actuate, the pain & pleasure to implement that is a story for another day. Siebel Analytics (alias Oracle Business Intelligence - OBI) has been the defacto application for analytical reporting (a.k.a OLAP reporting). Both Actuate (now BIP) and OBI are tightly integrated with the Siebel CRM application, which provides a strategic advantage to get started with reports that make sense quickly.
But, we are here to discuss more about how reporting plays a role in SugarCRM. Sugar does offer a reporting module, but that is limited to professional/enterprise versions, so it doesn't exist as far as I am concerned. When I started looking out for what can be done about the problem, I realized once more the open-source advantage. Almost nothing (barring my technical competence or lack thereof) will bar me from getting any BI suite with decent enough APIs to integrate with the Sugar CRM application. The whole problem can now be percolated to two major steps:

1. First, an application that can enhance/provide good reporting capabilities, and hooks/libraries to integrate. Since it makes obvious sense to first check out free (freer as in you get the code) software, the list is not bad at all.
   
   1. Jasper reports - transactional/analytical
   2. Eclipse BIRT - carried forward in association with Actuate, transactiona/analytical reporting
   3. Spago BI - if not anything, the website is quite well done. Again, provides transactional and analytical reporting capabilities
   4. Rapid Miner
   5. Pentaho
2. Next, apply the criteria of integration with SugarCRM, which is a PHP based application.

Now, for the actual decisions to address these two problems. As far as I am concerned, any BI suite will do if it lets me do things easily and produces beautiful reports. Due to my limited background in BI, it is a no-brainer - I have to give a bigger weightage to the next point.
The second problem about integration was supposedly big now. Almost all reporting applications are java-based, and integration can pose a problem due to the steep learning curve required. But, again here comes to rescue another most big aspect nowadays - 'third-party app development'. There are more than few components on SugarExchange that provide utilities/connectors to some of the BI suites mentioned above. After some searching around I selected one of the third party components  called Zucker Reports, for the obvious reason that it was available for development. This module deals with SugarCRM and Jasper reporting tool, so Jasper it is.
As simple as that, and without a line of code actually written, a bridge was established between SugarCRM and Jasper. Now, for some initial exploration:

1. ZuckerReports provides the ability to generate On Demand reports or schedule them at your own convenience
2. Reports designed and directly implemented in Jasper server can also be made available in the CRM application
3. Parameterized reports can be generated as well (though I might add, the functionality is largely limited)
4. Generated reports can be stored against a account/opportunity or some such record, and emailed to anyone interested

Below are two screen shots that demonstrate what it actually translates to in the application.





Yes, I understand this is no reporting nirvana. But there were one too many times I stifled 'a wow' when things fall into place almost automatically. Capabilities listed in other tools, and integration challenges certainly look promising to me and that is something I have to address sometime in the near future.

SugarCRM: Seamless Mail Sync

The disclaimer first - my knowledge with SugarCRM has been limited and I only try to dig into things when I find something interesting enough. I have been playing around the latest version (6.12) of SugarCRM lately, and was fairly delighted to see a good email integration feature.

How it works is pretty straight-forward.
1. The user can setup his own account to connect to an email account - gmail, yahoo, MS Exchange enabled by default

2. Connect to email box and check things in Sugar

3. Import the received/sent email within Sugar, and build other entities like account, opportunity etc. around it

I was also impressed by the level of integration a calendar entry has. For example, a divisional manager sends a meeting request to his boss and the rep working on a proposal. This is what the rep gets in the email.

Ofcourse, the point to note is that this was sent from the SugarCRM calendar. There is greater flexibility if it was google calendar (yes, there is a mod to connect with Google calendar).

The reason I was so delighted should probably be attributed to the way PIM Sync works. Until that is re-written, we probably would not miss out on 'complex' projects to implement email integration.

Siebel: Prototype in eScript - Part 3

In earlier posts, we have seen using prototypes to extend OOB objects. In other posts we have also dealt with user defined objects and scripts libraries. Time for seeing all three in action you say? And I hear you, we shall see how user defined object can be prototyped, and used across the application.
Problem: I need a function to lookup the description of a LOV provided its type and value
Solution: We address the problem in two steps:

1. Define a user object and make sure it is available for the session. One of the common ways of doing this is in Application object

2. Use the custom object to define a prototype for useful function(s) - in our case 'LookupDesc'

 Now, the how..

• Define a variable in the application script, for example: Siebel Life Sciences

[Application: Siebel Life Sciences, (declarations)]

var MyApp;

And, call a business service where MyApp is initialized. This is just to keep the application object, it works just fine within the application itself.

[Application: Siebel Life Sciences, Method: Application_Start]

function Application_Start (CommandLine)

{

 var bsStarter = TheApplication().GetService("COG Test Proto");

 bsStarter.Init();

}

• Write the custom business service 'COG Test Proto'

[Service: COG Test Proto, Method: Init]

 function Init ()

 {

    TheApplication().MyApp = new MyAppFun;

    MyAppFun.prototype.LookupDesc = LookupDescFun;

 }

[Service: COG Test Proto, Method: MyAppFun]

       function MyAppFun()
       {
 // this is a dummy, but funny function.
       }

  [Service: COG Test Proto, Method: MyAppFun]

function LookupDescFun (Type, Value)

 {

 try {

  var sDesc = "";

  var boLOV = TheApplication().GetBusObject("List Of Values");

  var bcLOV = boLOV.GetBusComp("List Of Values");

  with (bcLOV) {

     ClearToQuery();

     SetViewMode(AllView);

     ActivateField("Description");

     SetSearchExpr("[Type]='" + Type + "' AND [Value] = '" + Value + "'");

     ExecuteQuery(ForwardOnly);

     if (FirstRecord()) sDesc = GetFieldValue("Description");

  } //with

  return(sDesc);

 }

 catch(e) {

  throw(e);

 }

 finally {

  bcLOV = null;

  boLOV = null;

 }

 }

 That's about it, now MyApp is ready to have some fun. Again, a quick service to test how things work.

[Service: COG Test Proto Call, Method: Service_PreInvokeMethod]

 function Service_PreInvokeMethod (MethodName, Inputs, Outputs)

 {

    Outputs.SetProperty("out", TheApplication().MyApp.LookupDesc("blah", "blah"));

 }

A bit of theory on how it works:

1. MyApp is defined as a variable in Application, the same is initialized in a business service. It does not matter what the function [MyAppFun] contains, we will not be using that anyway
2. The prototype LookupDesc is defined as an extension of MyAppFun. Any object based on MyAppFun will have access to the LookupDesc functionality
3. We invoke LookupDesc as an extension of the variable in the application object. There are other blogs out there who just use TheApplication = MyApp, but that practice is not encouraged as I understand
4. When TheApplication().MyApp.LookupDesc is invoked, Siebel will look for the LookupDesc function for MyApp object. Since the functionality does not exist OOB, the prototype is consulted and in turn LookupDescFun function is executed.

If you are thinking this is an overkill, you may be right if this is all you want to do. Keep in mind that this coding practice will be more scalable, and ensures reusability.

Siebel: Prototype in eScript - Part 2

The last example in prototypes dealt with extending a OOB data type, but that extension was done within the business service itself. Although the prototype can be defined only once and used multiple times, it is not really demonstrated well since the entire thing is handled within the same code. Here, let me take an opportunity to show how the prototype definition itself can be handled elsewhere.

Problem
Check whether the given string is a valid email address

Solution
Define the prototype to extend the OOB string object, and use it as many times as needed. To achieve this we define the prototype in the Application object.

Go to "Server Scripts" of 'Siebel Life Sciences' or similar, and define prototype.

(declarations)

String.prototype.IsEmail = CheckValidEmail;

Add a function 'CheckValidEmail' to the same application object.

function CheckValidEmail()

{

var rEmailPat = /^\w+[\w-\.]*\@\w+((-\w+)(\w*))\.[a-z]{2,3}$/;

return rEmailPat.test(this);

}

That's about it! Any string can now be tested for a valid email. We will quickly build a test function to demonstrate how this is used:

Service: COG Test Email

Method: Service_PreInvokeMethod

function Service_PreInvokeMethod (MethodName, Inputs, Outputs)

{

var sEmail = Inputs.GetProperty("Email");

Outputs.SetProperty("Valid", sEmail.IsEmail());

return (CancelOperation);

}

When you run this in the simulator, this should either return a 'true' or 'false' value depending on whether the supplied string is a valid email address. Now, a bit about how it works.

1. First we define the prototype for the data type string with the statement String.prototype.IsEmail = CheckValidEmail'. This statement is executed when the application starts up.
2. The above prototype points to a custom function 'CheckValidEmail'. This function uses a simple regular expression to test the validity of the entered string. It returns a 'true' when the string has the specified regex pattern, false otherwise
3. None of the above has any effect until we actually invoke the 'IsEmail' function. When this function is used against the string object, Siebel checks the prototype since no such function exists OOB
4. Return value from the function denotes whether the string 'sEmail' is a valid email address

Watch this space for one more example!

Siebel: Script Libraries

Siebel 8 (or ST engine to be more specific) has a elegant way of calling business service methods. Script libraries provide the ability to call functions on a business service directly, no more is the need for building property sets and setting the parameters!

If you have not already used script libraries, do take a look. The reduced coding implies lesser mistakes, debugging time and easier maintenance! The simple demo below illustrates the usage.

Step 1:
Create the target business service 'COG Test Script Lib'. A custom function 'Echo' is defined here.

Service_PreInvokeMethod:

function Service_PreInvokeMethod (MethodName, Inputs, Outputs)

{

return (CancelOperation);

}

Echo:
function Echo (MyString)

{

return "Echoed: " + MyString;

}

Step 2:
Create the caller service 'COG Test Script Lib Call'. There is only one method here - 'Service_PreInvokeMethod', and the following code has to be input:

function Service_PreInvokeMethod (MethodName, Inputs, Outputs) {

var bsTest = TheApplication().GetService("COG Test Script Lib");

Outputs.SetProperty("Output", bsTest.Echo("abc"));

bsTest = null;

return (CancelOperation);

}

Compare this to the approach in earlier versions. In most probability you will be coding something like:

function Service_PreInvokeMethod (MethodName, Inputs, Outputs)

{

var bsTest = TheApplication().GetService("COG Test Script Lib");

var psIn = TheApplication().NewPropertySet();

var psOut = TheApplication().NewPropertySet();

psIn.SetProperty("MyString", "abc");

bsTest.InvokeMethod(psIn, psOut);

Outputs.SetProperty("Output", psOut.GetProperty("ReturnString"));

psOut = null;

psIn = null;

bsTest = null;

return (CancelOperation);

}

And, don't forget to add a line or two in the target service 'COG Test Script Lib' to set required properties in the output propertyset. Now, a simple 'bsTest.Echo("abc")' does the job. Isn't life simpler?

Siebel: Prototype in eScript

I had mentioned about using prototypes in one of the previous posts, here's a drilldown into how things work.
Tucked away in a remote corner in the eScript reference of the Siebel Bookshelf is a single paragraph referring to prototypes. Common to javascript developers, prototype is not something that we bother about too often. Prototypes provide a highly reusable way of extending out of the box objects, with an optimal memory usage as a bonus.

How do we use Prototypes?
The best way to understand will be through an example. Let us consider a problem where we need to add x days to the given date and get the next working day (!= Saturday or Sunday).
We go about this problem by extending the Date object to support a new functionality - add x days and retrieve the next working day. For simplicity sake we keep the definition of prototype and the actual call together, however they may be located in different objects as well.
To demo the prototype define custom business service 'COG Get Working Day' in Siebel Tools or Client. Paste the following code in relevant sections:
(declarations)

Date.prototype.AddDaysToGetWrkDay = AddDaysToGetWrkDay;

AddDaysToGetWrkDay

function AddDaysToGetWrkDay(iDays, Outputs)

{

var iDateNew; var iOneDay = 24 * 60 * 60 * 1000; // day in milliseconds -
(hrs/day * min * sec * ms)
this.setMilliseconds(iDays * iOneDay);
if
(this.getDay() == 0 this.getDay() == 6) { // if sunday or saturday

this.AddDaysToGetWrkDay(1); // add one more day and check whether that falls
on working day

}

return this;

}

Service_PreInvokeMethod

function Service_PreInvokeMethod (MethodName, Inputs, Outputs)
{
try {
var dSomeDate = new Date("1/28/2011");
dSomeDate.AddDaysToGetWrkDay(1);
Outputs.SetProperty("New Date", dSomeDate);
return (CancelOperation);
}
catch(e){
throw(e);
}
finally {
dSomeDate = null;
}
}

When executed through simulator (or a call from placed from another service), the service returns the next working day - in the above example 31-Jan-2011. Read on to find out how.


Code Explanation

1. 'Date.prototype.AddDaysToGetWrkDay = AddDaysToGetWrkDay' is the statement that tells the execution engine that there is a new prototype for Date object, and the extended functionality is provided by the function (or object) called 'AddDaysToGetWrkDay'. This is not synonymous with calling a function since there is no actual execution at this point. It is equivalent to a declaration.
2. Next, define the function 'AddDaysToGetWrkDay'. This will act on the provided date object, which is referenced using 'this'. All it does is to add the given number of days to the date, check whether the new date falls on a Saturday or Sunday, continue adding more days if that is the case, and return the final date. Recursion is used to keep on adding days until it is not required.
3. 'Service_PreInvokeMethod' just calls the extended function 'AddDaysToGetWrkDay' against the date object. When the actual call is made, the scripting engine looks at the Date object itself to check whether such a method is present, and then consults the prototype. Since such a prototype exists and points to a function, the function is executed to return the result

Note that the prototype itself may be defined only once, and the functionality is available thereon. Though this example is simple at its best, it does demonstrate how prototypes can be used to extend the objects.

Conclusion

Keep in mind the following:

• Prototypes share the same memory space. Multiple invocations do not mean multiple copies
• Prototypes can also be added later. Dynamic invocation can mean that you don't need to do everything in one place and at once
• A prototype can also be extended further. For example, I can very well define another which says 'AddDaysToGetWrkDay.prototype.ExcludeDay' to selectively exclude certain days. For example, myDate.AddDaysToGetWrkDay(21).ExcludeDay("Tuesday") can return the working day 21 days from the given date, but which is not a Tuesday
• From the above point you can observe that the methods can be executed in a 'chained' manner, which gives us a scalable solution that is easy to maintain (remember my OO reference in an earlier post?)
• Use prototypes for reusable functions, that is when they yield better results

So, how do you want to use prototypes in your application?

Siebel: Custom Functions for TheApplication

If you want to reuse code in Siebel, one of the common practices is to put it away in the business service and invoke those methods. This method, though is invaluable for complex operations, will become a painful process for simple things. When I say painful, it applies in some measure for both developers and the execution engine, reasons outlined below:

1. Additional business service objects have to be created during execution, this may put some overhead in terms of performance (business service caching kept aside for now)
2. Invoking most of the business services out there involves creation of property sets and calling them in a "proper" way. With the increased lines of code comes the increased complexity
3. Developers have to "know" about the business service. This is a big problem in the longer term since code again gets littered everywhere inspite of the same functionality in a service
   There is a way to solve some of the above issues for simple reusable code (at least partially) - custom methods on TheApplication object. This uses a simple fact - all functions written in the application object will be available in any entity and method. Let us illustrate this with an example.

Problem: I need to get the description of a specified LOV. Though developers swear by LookupValue (and other Lookup) functions, there is little one can do about the 30 (or 50) char limit imposed by the data model. It is common to use Description (or some such field) for this purpose, but the code reuse is limited.

Solution: Write a LookupDesc function in the application object. Just edit the 'Siebel Life Sciences' (or any application) that is being used by your object manager. Create a new function called 'LookupDesc' and paste this code:

function LookupDesc(Type, Value)
{
try {
var sDesc = "";
var boLOV = TheApplication().GetBusObject("List Of Values");
var bcLOV = boLOV.GetBusComp("List Of Values");
with (bcLOV) {
ClearToQuery();
SetViewMode(AllView);
ActivateField("Description");
SetSearchExpr("[Type]='" + Type + "' AND [Value] = '" + Value + "'");
ExecuteQuery(ForwardOnly);
if (FirstRecord()) sDesc = GetFieldValue("Description");
} //with
return(sDesc);
}
catch(e) {
throw(e);
}
finally {
bcLOV = null;
boLOV = null;
}
}

Now, invoking the above function is by using a single line in applet, bc or a business service -

TheApplication().LookupDesc("MY_TYPE", "My Value")

This is also available in ScriptAssist against TheApplication, to make others aware of this function ofcourse!
A note of caution though - don't overdo this. I would recommend keeping the code here simple and tight, and to those functions where lot of reuse is foreseen.

Siebel: User Defined Objects

Ever missed all the OOP goodness in Siebel? Although not used frequently, user defined objects provide you with some degree of control with hiding complexities in scripting. No, this will not help you attain OO nirvana - but you can start doing things with the custom objects and prototyping (another post for another day) to:

• Hide complexity
• Provide scalability
• Save memory while doing the above

Here’s a simple example that demonstrates use of user-defined objects. To test, you simply copy the code in a new client business service and you are all set.
First the Service_PreInvoke method:

function Service_PreInvokeMethod (MethodName, Inputs, Outputs)
{
var qryBC;

qryBC = new GetRecord("Contact", "Contact", "[Id] = '0-1'", "First Name", "Last Name");
Outputs.SetProperty("First Name", qryBC.Values[3]);

Outputs.SetProperty("Last Name", qryBC.Values[4]);

return (CancelOperation);

qryBC = null;
}

Next comes the constructor:

function GetRecord ()
{
try {
var arrValue = new Array();
var iCount;

if (arguments.length < 4) TheApplication().RaiseErrorText("I need minimum four arguments - BO name, BC
name, SearchExpr and at least one field name.");

var bo = TheApplication().GetBusObject(arguments[0]);
var bc = bo.GetBusComp(arguments[1]);
with (bc){
ClearToQuery();
SetViewMode(AllView);

SetSearchExpr(arguments[2]);
for (iCount = 3; iCount < arguments.length; iCount++){
ActivateField(arguments[iCount]);
}
ExecuteQuery(ForwardOnly);
}

if (bc.FirstRecord()) {
for (iCount = 3; iCount < arguments.length; iCount++){
arrValue[iCount] = bc.GetFieldValue(arguments[iCount]);
}
}
this.Values = arrValue;
} // try

catch(e){
throw(e);
}

finally{
arrValue = null;
bc = null;
bo = null;
}
}

Now, the explanation:
Purpose:

Provide ability to query any given BO / BC and retrieve the specified field values

How did we do that:

First, we create an object called “qryBC”, which becomes instance of a class “GetRecord” when the constructor “GetRecord()” is executed. Since this is the very first exposure to the class/object concept, we are letting the constructor do all the work rather than splitting it up. At this time we also pass the arguments to the object, whereby the constructor will query and return you the results. For simplicity in further processing, we return an array with the query results. Note the use of ‘this’ in the constructor and the reference to the set values when retrieving results in PreInvoke method.

Never understood why OO beats procedural programming you say? You need to depend your friend Google to find out what you missed.(cross posted elsewhere)

CRM 2011!

It is that time of the year folks - time when you, I and every IT guy worth his salt tends to get clobbered with analyst talk. This is all about the past year - how they prophesized just right, and how the world is converting to paradise or is getting destroyed in the nuclear winter next year. Since I too am finding myself putting together hard-to-comprehend words one too many times, I thought this is something ought to be addressed pronto (before you sink too deep in the holiday mood that is). Since CRM is what I am doing all the time (it is actually Siebel application development, but it feels liberating to be more generic), this is a short list of what I think should happen to CRM in 2011.

1. Fusion should come out of its wraps
   Inspite of Elison's jab at the implementable state of Oracle Fusion apps, I quite look forward to it. In the process, I also look forward to throwing out all the experience of developing lousy scripts using eScript/Siebel VB, half-baked COM interfacing programs, workflows with a clunky interface and all the configuration fun that can be had in any version of Siebel apps post v7. I will miss dismissing customer requests with a hard-hitting statement - "the functionality is tagged to SRF, and cannot be changed now". 2011 is a good year to brave up to J2EE, ODI, WebCenter, WebLogic and other hotchpotch. This is number one solely because I am associated with Siebel for too long a time.
   
2. SaaS triumps in Large Enterprises
   Yes, I know this has been said since the inception of SFDC and the glorious just-dream-forces unleashed upon us. But it is not hard to listen when businesses (not IT) is shouting at you about why they should be spending so much on CapEx and ongoing maintenance including upgrades. So there will be a slew of studies on SaaS take-over, hybrid architectures and how many millions one saved while firing the software and the dream-teams which were keeping them afloat. Well, this cannot happen in next year alone but will enable me to brag about how it came true to a certain extent at the end of 2011.
   
3. Mobile to the Rescue!
   Mobile phones with their ubiquity and affordability, should wield greater power over deciding the CRM strategy. How many of us wants to navigate the shiny website of Airtel when the transactions can be done through mcheck? CRM will get more focussed in the mobile space, true for developed and developing countries alike. App development, here I come!
   
4. Social CRM will bring down companies to their knees
   Well, I sort of made it up - companies don't have knees. Service companies especially will struggle to keep an eye on what is going on, and would have to resort to several IT adventures before they can get things right. A piece of copied advice - it is not about IT guys (a.k.a 'dont blame me'), set your focus on what you want to do when customers are talking all good or bad about you.
   
5. Paul Greenberg should be crowned the undisputed thought-leader of CRM
   I really like his way of writing - the book and the blog, even with my zero knowledge of baseball. Though I feel he vastly over-rates the power of getting Social, being a undisputed leader will push to the fore-front his genial thoughts on maintaining good customer relationship.

Siebel: Importing data from files

I am back after a not-so-brief hiatus, and am back hopefully for good :). Here I am with yet another utility - this time something that will help you to import data from files into Siebel, with zero coding required from your end. ‘COG File Import Service’ as it is known, will hopefully simplify your life a bit.

Contrary to popular (and let me add, buggy) excel macros / VB COM programs that are available to load data to Siebel, this is a business service that can be imported to Tools / Client. But at the same time it does not offer as much flexibility, given below are a few things that make this tool stand apart.

Why another tool?
File import is nothing new to Siebel developers. But why use another tool when there are dime-a-dozen you say? Thus goes the story:

• The first thing that hits you - most of the tools to load Siebel data have inconsistent interfaces, and may require some amount of ‘getting-used-to’
• External objects or platforms are required to actually see what is going on, and hopefully to change the behavior. For example, you need Excel or Visual Basic software, or have a dependency on Windows platform to execute the code. And all that can be done, provided you understand what has been written there
• Customers might have their own restrictions in connecting to production systems using third party software and utilities

The Mechanism

COG Import Service is just a business service that can be applied to Siebel v7.5+, but has only been tested with 7.7 and 8.1. The service works with CSV/delimited files and has two distinct ways to handle them:

1. Simple Update - This is just a wrapper to the OOB CSV reader service, with a couple of enhancements thrown in
2. Update - A more sophisticated updater, which can do a few more things than what CSV reader can do

The steps involved in both methods are simple:

1. Read file (all at once)
2. Map fields to provided content
3. Use a data map, if specified
4. Update

Where they differ is in the way they process files.

When to use which method?

.. Simple Update

• Uses the 'CSV Reader Business Service', and hence needs an integration object (IO) to be available for the input data
• When the processing is simple and there is no need for processing more than one component/entity (in other words, entity and its children) at one time
• Processing should be fast
• Any errors should roll-back all transactions

.. Update

• Can be used with or without involving IO
• Delimited with be comma or something else
• Process entity and one child with one operation

How to Use?

Import the attached XML into Siebel Tools and compile to SRF, or imported into the Client (Site Map > Business Service Administration). Parameters to the service specify the parameters that are needed for a successful run. The file formats/structure has been documented in the service itself.

License

None, you can use it or throw it away. Use it at your own risk, unintentional bugs are all but risks of the trade.

You can watch this space (at your own risk, of course) for new versions - improvements or bug fixes.

Fusion - at last

Well, the moment we have been waiting for is here - Oracle Fusion CRM finally finds the light of the day (so as to speak) in the recent Oracle Open World. This is a new technology framework folks, put on your best armour and start preparing for all new things coming your way. You can start with http://www.oracle.com/fusion! No, no downloads yet - you should satisfy yourself with the latest sales pitch. Alternately, you can take your frustration to the Siebel 8.3 Statement of Direction (available on support.oracle.com).