HandledException

HandledException has moved

suniljagadish — Fri, 24 Jun 2005 14:17:00 GMT

I am moving to a new blog - http://geekswithblogs.net/suniljagadish/ after bearing with comment spam for quite sometime and also suggesting to the admins to include CAPTCHA to prevent comment spam. No response so far. Its irritating when you find an email saying - “XYZ has posted a comment on your blog” and then you find out that its only a list of sites where you can play online casino.

I know its really annoying to ask someone to open their blog roll or their RSS agregator to update the feed link to my blog/feed. Please take the trouble of updating your links. I thought it would be a good idea to use feedburner to manage my RSS feeds. So here is the new RSS feed url - http://feeds.feedburner.com/geekswithblogs/suniljagadish

With me moving to a new blog, you will definitely find me posting good amount of code apart from the tech bytes. Thanks to dotnetjunkies for hosting my blog all these days.

Cya!

Moral of the story: User experience is very important.

Visual Basic 6 "Dim" Days

suniljagadish — Thu, 23 Jun 2005 14:37:00 GMT

Sorry if you mistook this to be a Microsoft VB6 event which was coming up. :)

If you are a VB6 fan and feel that you have been over-shadowed or drowned in the world of the “uncivilized semi-colon” languages (as Raj calls it), check this out - Microsoft Visual Basic 6.0 Run.

This reminds me of the VB6 projects that I had done. It was about 3.5 years back when I thought of learning Visual Basic 6. I purchased the Visual Basic 6 Bible which is no way a recommended book for beginers. I tried to follow the samples given in the book but was not completely successful with it. After a short term course, I set out writing small applications including an Astrology software, IVRS Voicemail retrival system and stuff like that.

I had a very good time working on iCafe' Manager - an exhaustive Internet Cafe' management appllication which Sharath and I had done as a part of our academic project, for which we ended-up with an ultra-cool UI (3D bars, shadows, animated menus, etc.) . Thanks to Photoshop and the Visual Basic timer control. We had used almost 10 timer controls to animate the scrolling of menus and other parts of the UI which could be hidden/minimized. It was more of fun doing this though we didnt get much time to optimize the code we wrote.

I did some serious development using VB6 for my final semester project (BCA) which was a ready-to-market 'Email Management over SMS' suite. It was targetted at corporates who wanted to give their employees Email access on the move (when GPRS wasnt as affordable as they are now). This was the first time did some serious development with Winsock (in contrast to winsock chat applications where one usually starts off). Socket based communication with POP and SMTP servers are not always straight-forward. Thanks to Kurian and JK of Xtend Technologies, Cochin, I used their XSMS platform to communicate with the GSM world. I very badly needed to use multi-threading when I was developing this but a VB6 developer would know what that means. Thanks to .NET. Life has changed.

Offlate I have been trying to experiment with the Microsoft Windows Embedded Shared Source Tools - Bluetooth library. I have been facing some referencing problems with it, but will be out of it soon.

.NET Passport SDK & IIS Issues

suniljagadish — Sun, 05 Jun 2005 11:56:00 GMT

I installed the .NET Passport 2.5 SDK, to enable SSI on one of the portals I am developing. One good thing about version 2.5 of the SDK is that it has support for managed development in C#. I wanted to host this portal on my ASPXConnection webspace. Unfortunately ASPXConnection doesnt give access to some resources which may be required in special cases such as this. To install the Symmetric Key given by the Passport Services Manager, I needed command line accees to install the key. :-(

I had installed the .NET Passport 2.5 SDK on my system and uninstalled the same after realizing that I would not be able to do this with my webspace. The next time hit F5 in VS.NET 2003 to run my web application, it said:

“Unable to start debugging on webserver”

I run the web application without debugging and it was ok. Conclusion was that my Web.config was ok. I didnt want to research for a very long time on this issue, so went ahead with the good old “KB306160”. I reinstalled the .NET Framework and everything was back to normal.

Update: Article 306172 seems to be a better place to start off, if you have ASP.NET & IIS issues

World in your hands - NASA WorldWind on .NET CF !

suniljagadish — Tue, 31 May 2005 00:03:00 GMT

NASA WorldWind is worth a check.

“World Wind lets you zoom from satellite altitude into any place on Earth. Leveraging Landsat satellite imagery and Shuttle Radar Topography Mission data, World Wind lets you experience Earth terrain in visually rich 3D, just as if you were really there.”

WorldWind uses Managed DirectX (MDX) to acheive this. But what caught my attention was /cfWorldWind, when I was reading Mike Hall's weblog. Managed DirectX 3D Mobile (MD3DM) can do wonders and this is a proof of it.

One of the screen shots of /cfWorldWind:

Ready to go to Hawaii?

suniljagadish — Thu, 19 May 2005 01:14:00 GMT

Microsoft always has this tradition of code-naming their products in a unique way. “Hawaii“ is the code name of the Visual Studio IDE to be launched after Oracus. It definitely grabs the attention of developers. Sometimes it does scare developers when they begin to think - “Ok, so what about Whidbey? Is is already out-dated? Do I need to re-learn Hawaii“, “What about Oracus? Whats the difference in it?“. Hawaii is supposed to be a completely re-designed toolset.

So far I havent come across any resource on the Microsoft website which shows a complete product release and platform support diagram which talks about the upcoming releases. A consolidated view of the release strategy of Microsoft would be interesting to see.
[eWeek]

Update:

Microsoft Visual Studio Roadmap
James Curran
http://msdn.microsoft.com/vstudio/productinfo/roadmap.aspx

Who says Bluetooth is dead?

suniljagadish — Sun, 15 May 2005 11:45:00 GMT

In continutation to the subject of our (Team: BlueCQ - Sabarish, Sharath and myself) team's participation in the Imagine Cup 2005, yesterday we had the South Zone finals in Chennai. We were geared up with our idea to dissolve boundaries using technology, calling it “BlueCQ” (as in ICQ). It is a bluetooth based people networking platform with lots of innovative; fun and business oriented features incorporated in it.

No twists, straight to the point...
Our team didnt make it to the next round (National Finals) because of the following reasons sighted by the judges and the concerned people.

Bluetooth is going to die / Its already dead and 802.11b has already taken over
Bluetooth is unreliable
Bluetooth has a very short range (we had come up with an innovate solution for this but judges weren technically sound to understand and appreciate it... it was disgusting to note this)

When I first heard about Bluetooth I only knew 2 things about it - its a cool new wireless technology, its a substitute for Infra-red. The latter isnt obviously completely correct, which I realised after getting myself involved with Bluetooth programming for the Imagine Cup. Anil Dhawan's video on Channel9 is what motivated us to pick-up Bluetooth as the medium for communication. I started off by doing some initial research about what the technology is and how it works (the Bluetooth stack, the Bluetooth radio etc.). I didnt want to do a crappy job by starting off right away with some code.

According to the modules that we divided, I was the SDK for our platform which has been a very exciting experience for me. Wouldnt be explicit but, our project has a feature which the Bluetooth SIG (the body governing the Bluetooth standard) wants to incorporate into the Bluetooth stack after 2 years !!

We went ahead with this idea and landed up in Chennai on 13th May 2005 for the South Zone finals. We did have a few last moment preparations and sat coding in the hotel where we were staying. We were told that the judges would be from the “industry”. So, we were expecting a good technical interaction. However, expectations always do not turn in to reality. There was a tie with another team after the scoring and we didnt get through because we committed the crime of using Bluetooth (which according to them is dead).

Its really frustrating to note that people are not well informed of the technology and just speak what they “hear“ from others. Its the general notion that, when you ask someone about Bluetooth, his immediate reaction would be - “Naah, its dead.” (even without really knowing whats happening out there).

Even one of the Microsoft (India) folks said that Bluetooth has been replaced by 802.11b already !! Its unfortunate that probably the Microsoft guy hadnt seen this podcast from Microsoft MEDC 2005. To add to my arguments, Microsoft is also funding the Software GreenHouse initiative, which is currently working on a Bluetooth based social networking application which is hardly 50% of what our idea is. It called - “BlueJab”.

The war between different technologies will continue.

Hoping to see a better IC 2006.

Subjected to Comment Spam

suniljagadish — Fri, 29 Apr 2005 10:29:00 GMT

:( My blog has been receving a lot of comment spam. I wonder if there is any way to get across this highly irritating problem. If you have any feedback/comments you can mail me suniljagadish AT gmail DoT com

Hash Tables

suniljagadish — Fri, 29 Apr 2005 10:20:00 GMT

Most of the text books which I had read as a part of my ciruculum explain Hashing in a very abstract manner. Some text books wrongly interpret hashing as something only to do with files and searching in a file. A better way to project hashing would be as a data structure that allows faster and effcient data look-up.

Many programs that we write needs to store some temporary data which may have to be looked-up later during the execution for some purpose. The usual way one would approach this is to declare an array and search the array using a search algorithm like binary search or linear search. The problem with this method is obviously the cost pertaining to the time taken to search for the value. The time complexity of such methods are in the order of O(N) or O(N/2). This method is fine for programs which are done for an experimental or learning purpose.

So, is Hash Table some kind of a searching algorithm?
Well, not exactly. As I mentioned before, it is better called a data structure.

Consider this analogy:
Lets say that I go to a library and I am looking for “Windows CE.NET Programming by Doug Boling”.

Traditional Scenario
The conversation happnes this way in case of a traditional search algorithm, lets say, which is being used by the librarian.

Me: “Hi! I wanted Windows CE.NET Programming by Doug Boling.”
Librarian: “Ok. Give me 2 minutes. I will find out and tell you in which rack you can find the book.”
[Librarian opens up a register and looks thru page by page for the title of the book...]
[I am still waiting for an answer...]
Librarian: “Here it is. You can go to Rack 3 in the Technology section.”
Me: “Thank you.”

The hash table way...
Now lets say that the librarian is using hash tables to look-up the register.
Me: “Hi! I wanted Windows CE.NET Programming by Doug Boling.”
Librarian: “You can go to Rack 3 in the Technology section.”

Ok, so you saw how quick that was. How did the librarian manage to find out the location of the book so quickly. This brings us to a point wherein we need to know what *exactly* is a hash table and what is a hash function.

A hash table may contain objects that have keys, and values. The hash table is made of an array into which objects may be inserted, and a hashing function. The hashing function takes a key as an argument, and uses it to calculate some index in the range of the table's array. This function is used to map different elements to different arrays in the table. Hash Table approach is a more “predictable” way to look-up into a set of values.

When we want to insert an element in to the hash table, we calculate its hash value (the result of activating the hash function on the element's key). We then go to the location in the array whose index is equal to this hash value, and insert the item there. If this location in the table is already in use, we look for the nearest free location and place the element there (there are various other variations for how to handle such "collisions". the method we describe here is called "simple hashing").

When we want to locate an element in the array (given its key), we calculate its hash value, and then start scanning the table at the location whose index is equal to this hash value. If we find the element (by comparing the keys), we are done. If we find an element with a different key, we move to the next higher location (moving back to the beginning of the array if we are at the last element). We keep scanning until we either find the element, or find an empty cell. Such an empty cell indicates that the element was not found in the table, and thus the search fails.

The time complexity with hashing (in case of a perfect hash) would be O(1).

I know that if not the the concept of hash table, the actual implmentation would still be unclear. I will re-direct you to some links which graphically explain how hashing works.
1. Follow the images: http://www.sparknotes.com/cs/searching/hashtables/section1.html
2. Cool animation: http://ciips.ee.uwa.edu.au/~morris/Year2/PLDS210/hash_tables.html Very lengthy explnation, but scroll down and you will find a “Run Animation” button. Its a real cool animation which explains Hashing in a very neat way.

OpenNETCF to be acquired by Novel !!

suniljagadish — Sun, 10 Apr 2005 17:09:00 GMT

I wonder if this is good news or bad news but its true that OpenNETCF.org is going to be acquired by Novell.

Miguel de Icaza on a press note -

"OpenNETCF.org is key to Novell's business strategy in the embedded space. We realise the potential benefits for our customers to be able write code in C# and run this under Mono not just on a server or desktop machine, but also on PDAs, smartphones, games consoles or even internet toasters. When it comes to .NET technologies and knowing the constraints that embedded devices have, the OpenNETCF.org folks really know their stuff. And we just had to have them."

ASP.NET and IIS Problems

suniljagadish — Sat, 19 Mar 2005 15:47:00 GMT

Quite a few times I have faced the problem of running ASP.NET web applications on my IIS.

The various errors which I used to get are:

“Cannot start ATL or COM debugging“
“The webserver running your web application is in the Internet Zone“
“Cannot start ASP.NET debugging on this webserver“

Every time I did a search for help on these topics, I landed on the respective Microsoft Knowledge Base Articles, which explain these problems in a nice manner. One such article is “Visual Studio .NET has detected Web server is not running ASP.NET 1.1”.

The problem with my IIS was this one:
3. Web Server specified by ASP.NET Web Project/Web Service Project is stopped or IP address cannot be mapped.
A: Please check IIS's Web configuration, include IP Address and working status. If your network IP address already changed, you must change your Web Site's IP Address. (If you use "localhost" to create Web Project, you can do nothing.)

All I did was changed the IP address configuration in IIS to 127.0.0.1. I did try to use localhost, but IIS doesnt take it. Once I did this, I was able to create ASP.NET applications. The next problem was that I wasnt able to run the ASP.NET web application because the URL of the web application ie. http://127.0.0.1/WebApp1 was in the InternetZone. The solution for this is given in the MSDN Library.

Tough these are some small issues, it would be a real pain if one does not know that the MSDN KB has solutions for all these.

TAPI functionality is not supported in the Managed World!

suniljagadish — Sat, 29 Jan 2005 12:03:00 GMT

I had written a Dialer Application based on the Windows Phone Dialer that would pick-up numbers from a database and would dial 'em. From a CRM perspective, I also logged some stats about the telephone session. I did the coding in VB6 in the traditional way, by doing an API call to the tapiRequestMakeCall function of the “TAPI32.DLL”. My application would invoke the Phone Dialer and pass the phone number and callers name as the arguments. It works cool on Windows 98 and Windows 2000. I ran the same stuff on Windows XP. The Phone Dialer in XP is an updated one and supports both a Phone Call as well as an Internet Call. I need to prefix the numbers that I would be dialing by 0 (because I have a PBX). When I supply the parameter as “0, 24565687” as the number to be dialer, the dialer assumes that this is an internet call. I require the comma to introduce a delay before dialing the actual number because the PBX needs a second to switch to cease the PSTN line. This is quite irritating especially after I found out that when I dial the number manually in Phone Dialer (same number), it treats it as a Phone Number and does the job well.

With this problem, I thought that I will port my dialer to .NET. It all came to a halt when I found this KB article, which says that TAPI functionality cannot be implemented in managed code.

The reason:
Because of the complexity of the TAPI 3.x Component Object Model (COM) interface, the managed wrapper that is created by Microsoft Visual Studio .NET does not work. Therefore, you cannot call TAPI functionality from managed code.

The solution:
COM Interop. Hmmm... There is always a solution. :-)

Interesting note:
All the TAPI 2.x data structures are based on the DWORD data type. TAPI 2.x also requires heavy use of pointers. Therefore, TAPI 2.x functionality is difficult to use from any language other than the Managed Extensions for C++. Offlate I have been reading about the Managed Extensions for C++, which is pretty cool.

Security - the need of the hour

suniljagadish — Mon, 24 Jan 2005 23:46:00 GMT

I downloaded FreeSCO (Free ciSCO) a soft-router with firewall capabilities. I was on the lookout for a firewall, prefereably something which could be configured on a diskless PC. I preferred a diskless PC because of two reasons - easy and hassel free harware setup & not so easy for hackers to sneak in. Mr.Jayakrishna had suggested the use of FreeSCO quite sometime back. Now that I am getting a broadband internet connection, I definitely require something more than just a Zone Alaram. One can easily setup a custom firewall with two layers of firewalls. A FreeSCO box can be in the outer layer with an NT machine in the inner circle.

FreeSCO is an open source program that can run on a 1.44MB floppy disk and also it comes with lots of cool features and customizations that can be done. Being a command line utility it is not very user-firendly. All you need to do is to download FreeSCO and make a Floppy disk copy of the FreeSCO setup files. It is based on a 2.0.39 Linux kernel. And a sad part of it is that you need to tinker around with its code if you need to get internal modem support for it. I havent ventured too much into that direction but Skp was telling that there is a patch available for it. Yeah when I say modem, I reffered to a dial-up modem. It does support Leased line, Cable, DSL connectivity too. Support for PPPoE is also built into it. Another cool feature is that it has a web control pannel using which you can login from anywhere and control your firewall/router remotely. It can be configured as webserver to serve simple HTML content. Simple and cool implmentation of Linux.

Yeah obviously, it cannot be compared to a CISCO PIX or Checkpoint, which are real good enterprise ready and highly reliable products. Probably this is one drawback I see when you speak of many open source projects. They are not complete in the sense that not all features are built into it. Bug-fixes are a slow process. So if you wanna get a simple home-made low cost firewall, just go in for FreeSCO.

Imagine Cup 2005 - "Imagine a world where technology dissolves the boundaries between us.."

suniljagadish — Wed, 19 Jan 2005 22:56:00 GMT

Imagine Cup being one of the major global competitions for students hosted by Microsoft definitely draws the attention of anyone interest in computers, software, graphics and game programming.

I learnt about IC an year or more back that was when the IC 2003 was on. I hadnt worked on .NET at that time and I always thought that it is only for the geeks out there. "Change" is what is driving everything today. I saw the IC 2005 video when Mr.Tyson Dowd visited Bangalore and he spoke about IC '05. I still hadnt made-up my mind whether or not I should participate. Finally when the Round 1 dates for India was nearing, I teamed-up with Sabarish, Sharath and Deepthi Rao and we decided to get into it.

Round 1 is about a chess program, wherein we are given an SDK and we are supposed to code for a function to find the checkmate move. It is definitely a good task to do but I felt that we are wasting time when we think of what participants in other countries would be doing. They would be working on their Software that would be submitted for the finals. We are stiil writing some crappy code for a chekmate. Phew...

After lots of discussion amongst ourselves we decided to start off with a brute-force method to get cracking with the program. After facing a hell lot of problems with MovePiceUnconditionally, we chaged our approach and now our program seems to be working fine except that it needs more tuning for performance. I read in the spoke "...so that you can save some precious nano seconds..." Oh. Please I wasnt thinking in nano seconds to compute the Checkmate in 1 move. The SDK is undoubtedly crappy in terms of ease-of-use, clarity and in some cases even the functionality.

Inspite of the fact that my exams are going on, I am not able to resist from "Spoking" (as in visitng theSpoke and Poke to see whats up). I also decided to make a program that will test the code that we write for different TestCases. After having some problems with loading different XML files (which contain the position of the different pieces) I finally have a working verion of the tester, which all of us will be using. It would also help our team to test for different cases and easily store & forward the results of the test.

One of the test cases:

<ChessGame>
<White>
<King HasMoved="True" Position="b7" />
<Knight HasMoved="True" Position="e7" />
<Knight HasMoved="True" Position="f6" />
</White>
<Black>
<King HasMoved="False" Position="d8" />
</Black>
</ChessGame>

theSpoke was hot with lots of discussions about whether or not people should 'openly' post the strategies that they are using in their solution and also some of them wanted to even post their DLLs!! Sriram said it right... Why dont they post the code itself if they want to really help someone?...

Lots happening for IC... India stands at No.2 in the number of registrations for IC too!
Yo!

Mike Tyson and C#

suniljagadish — Thu, 11 Nov 2004 11:11:00 GMT

An interesting terminology that I came across while going through the C#LS was "boxing" and "unboxing". Though some may be familiar with these terms, I am new to it.

C# has a "unified type system". ie. all types including primitive ones derive from the type "object". A simple example would be:

class UTSDemo
{
static void Main() {
Console.WriteLine(333.ToString());
}
}

The above sinppet calls the method ToString on the integer literal 333 which results in the output "333".

class Boxing
{
static void Main() {
  int intVal1 = 123;
  object myObj = i;  // boxing
  int intVal2 = (int) myObj; // unboxing
}
}

From the above example it is clear that an int value can be converted to an object and then back to an int. This is nothing but boxing and unboxing [Mike doesnt know this one ;-)] When a variable of a value type needs to be converted to a reference type, an object box is allocated to hold the value, and the value is copied into the box. Unboxing is just the opposite. When an object box is cast back to its original value type, the value is copied out of the box and into the appropriate storage location.

So using this you could write a method which has object type as its parameters or as the return type when you want to make the method general irrespective of the type.

SMTP access using VC#.NET

suniljagadish — Sun, 24 Oct 2004 09:48:00 GMT

I developed the registration website for DWDN where I had to add a small functionality- when somebody registers for an event, an automated mail was to be sent to him/her to the email id specified during registration.

The harder way out was to use sockets and talk to the SMTP server directly and take the pains of handling various conditions. The simpler alternative is System.Web.Mail. The beauty of .NET lies in its feature-rich classes and the organized way in which everythings has been made available to the end-programmer.

The code was as simple as:

using System.Web.Mail;

MailMessage m=new MailMessage();
m.To=txtemail.Text;
m.Subject="Date with .NET : Registration Acknowledgment";
m.From="xyz@gmail.com";
msg="Hello "+ mytitle + txtname.Text +"\n \n Your registration for the event: \n " + lblevent.Text + " is under process.\nYou will be receiving a confirmation mail.\n Please feel free to mail us on xyz@gmail.com if you need any clarifications. \n Regards";
m.Body=msg;
SmtpMail.SmtpServer="localhost";//fill in the name of the SMTP mail server
SmtpMail.Send(m);//send the mail message

It provides properties and methods for sending messages using the Collaboration Data Objects for Windows 2000 (CDOSYS) message component. It can deliver mail through Win2000 SMTP mail service or any arbitrary mail server.

That’s cool enough. The site is hosted on WebMatrixHostitng and their server provides Authenticated-SMTP only. That’s because of their SPAM-control policy.

I had to add this extra bit of code into what I had done earlier.

m.Fields["http://schemas.microsoft.com/cdo/configuration/smtpauthenticate"] = 1;
m.Fields["http://schemas.microsoft.com/cdo/configuration/sendusername"] = "myusername";
m.Fields["http://schemas.microsoft.com/cdo/configuration/sendpassword"] = "mypass";

Here comes the problem… I was using VS.NET 2002 and the “Fields” collection for the MailMessage class isn’t defined in 2002. :-( I had no other option but to install VS.NET 2003 and get going with it.