Discussion Locked

The Programming Life: Experiences in Development

0 helpful
68 replies
Closed to new repliesThis discussion is closed. It stays readable, but it takes no new replies.
blog root

68 replies

by george.p.alexander

Some time back our development team of three got a requirement for the enhancement of a database application for one of our operations team who are based some 500 miles up north. This enhancement was a request made by their client. The success of this enhancement would lead to a bigger contract with the client. This was a small application project but yet very significant for our operations team.

Programming is great. But what do you when you're given a summary list of requirements at a broad level, an existing application to work on with absolutely no design document and an unreasonable timeline? Compound this with a barrage of frequent requirement changes during the development phase that affect dependencies. This leads to chaos. Eventually escalations start kicking off due to bugs and delays while you end up being part of the problem instead of the solution. This is a programmer's worst nightmare. And I am sure many can testify on living through this scenario.

I found myself in the front lines of this unwitting situation that started around a month back and which thankfully ended yesterday evening (I thnk). After a day of staying up late night and working a week's worth of coding and having done loads of emailing and call responses to escalations that went all the way up to the top management of operations and our group, I'm wondering if this is the limit of my endurance.

However, after the entire situation cooled down, we had an informal yet fruitful run-through on identifying key problem areas with respect to the way we handled this project on the very same day.

Many issues were discussed but one of the fundamental aspects that led to most of the fireworks was that we did not stick with a frozen requirement document up to the low level details from the start.

Can we really stick with frozen requirement documentations during the normal tenure of SDLC? Consider a small project with a short duration - Is it justifiable, as a developer or a project manager to skip certain standard software development procedures for a short term project in favor of winning a contract on behalf of the top management?

Even though we have standardized on proper software development processes, this was one occasion where we jumped the fence in favor of winning a contract. I'm also wondering what role an IT CRM could have played here? I am looking forward to hearing similar experiences and comments which would be of immense value.

by george.p.alexander

Yesterday I was required to prepare a document articulating recommendations against limitations of working on a VBA/Access 2000 system so that our client could have an objective idea on why moving for a OOPS/RDBMS such as .NET/SQL Server 2000 would be a more efficient alternative for his requirements over an Access Database system.

One of the points I brought up was that Access 2000 does not support a multi-tired approach to programming. Hence, this introductory exposition; not on the limitations of an Access Database but on a multi-tired approach to programming.

Another name commonly used for multi-tier architecture is "N-Tier". One of the fundamental starts while developing projects is to create a structure following the N-tier architecture. It greatly enhances the productivity of any development team which I will be expounding later in this article.

Now creating applications that require N-Tier architecture is nothing new. Developers' way back during the VB4 days used to develop using N-Tier architectures. It is now the norm for developing any large scale application.

Being a resource for .NET, I found that Microsoft has provided an excellent mechanism to create, build and implement n-tier applications. However, I would like to just give a run through over here on what an N-tier application is and how it can benefit the project development team on an overall basis. Later, we'll have a look at .NET implementation of the N-Tier architecture.

So what is an N-Tier application?

The fundamental characteristic of any n-tier application is that the entire application is split into several pieces - both logically and physically. Each one of these pieces performs a specific task (displaying UI, Data Access, Business logic etc). There can be a number of layers/tiers of such pieces and therefore the name that it is commonly called today: N-Tier.

The three most common layers in any n-tier application are

1. Presentation Layer
2. Business Logic Layer
3. Data Access Layer

The names in themselves imply the purpose. There can be more layers based on requirement and fore-sight.

The presentation layer is nothing but a piece of software that deals with the user interface of your application. Displaying data to the end user and allowing them to interface with it is the core functionality of this layer.

In any large scale application, there will always be cases where data entered by the end user will need to be validated or some process will have to be executed. This is where the business logic comes in.

The core of any application would end with data being stored and retrieved. This task is handled by Data Access Layer.

To a beginner or a novice, this architecture would seem really unnecessary and might even be considered as an overhead with respect to time and other factors (I have no idea?but initially even I was skeptical).

But the benefits are amazing...

1. The application is divided into logically isolated pieces therefore ending or reducing tight coupling between the User Interface, business logic and database.

2. Table and column names can be effectively eliminated from the client-side code.

3. This one is my favorite: It's pretty easy to change or add to your application new functionalities, without breaking or recompiling the entire client-side code.

4. Client application will no longer have SQL statements embedded in it. Therefore it remains de-coupled from rest of the application.

5. A change in one tier or layer does not affect the rest of it.

All this sounds pretty cool; however, there is one disadvantage: you would probably need to create many isolated classes and pieces of software. Does this outweigh the benefits? I think not.

If you have any more experiences using an N-Tier application, please share it with me. So far, I haven't noticed any drawbacks. N-Tier architecture probably wouldn't be required for small projects but considering a large scale application where changes might be required to be implemented later, an N-Tier architecture would defiantly make it feasible considering that we have already isolated the program into units and therefore we can identify a particular layer to be changed without affecting the entire application as a whole therefore ruling out the possibility of bugs or errors being created due to dependency issues else where in the application.

hmmm....enough of N-Tiering for now... Baby's N-Tier day out!

by george.p.alexander

"New virus hitting this Friday!", "Yeah? Okay. Yaaaawn...". My firm has given us an incoming warning of a so called deadly warning. My friend has also just sent me a warning.
But why am I least bothered this time? Me - the guy who used to keep track of hot virus news. Me - who used to sometimes observe the little magic they used to do (Pakistani brain virus was once my favorite!).

This post originally appeared on an external website

by mrizvandi

Play the game with VS 2005 software development. programming and development of software solution goes to be a game. not a technical or hight technology with vs 2005. you can play this game on weekend. All new programmer can develop a solution such as game.

by george.p.alexander

Have you noticed that the offshore business model is gradually changing in India, with big players acquiring smaller offshore based companies? Texas based IT Services giant EDS recently offered to buy a 52% stake in Mphasis BFL Ltd, an Indian IT Services and BPO solutions firm for around $380 million. Both sides look at this as a win-win situation.

This post originally appeared on an external website

by george.p.alexander

Why did Bangalore, the IT capital and the fastest growing city in India, have its worst ever riots anyway? I'm looking for a real good reason on why personal grief of the natural death of a popular veteran actor would mutate into public rage and violence. That this happened in Bangalore, the 21st century face of India to the rest of the world, with all its fancy shopping malls, all the biggest names in every industry setting shop here, with all its, offcourse, my favourite part: fast food joints

This post originally appeared on an external website

by muthukumar.g

For ASP and PHP projects, we dont find any issues... but for .net and j2ee based application, the deployment of components becomes an issue sometimes. soemtimes the DLL library files, sometimes the ocx control registrations.. etc.. for websites.. hmm..not that much.. if windows based application.. it is all perfect, but please check your CD with QC department for installation test..

by duckboxxer

Environment differences between development and production can cause issues.  Also any sort of production restrictions, if you don't control the production environment.  Also check with the host if you have something important coming up for you application.  An example might be if you are planning to show off your app to the corporate office and the host is planning upgrades to the server then.  Maintenance will be an ongoing battle as well.  One may upgrade the hosting server and that may not be compatible with all of your application (features depreciated, etc.).  Be sure that your system is fully documented also in case something dies or if the hardware is changed.  

Be sure that your project plan includes time for hosting to take place.  A prime example that I deal with daily is that with some centralized goernment hosting, people assume that we can get their app up and running that day.  Sorry, but we have review policies before we put anything on our servers and this could take weeks based on application size and workload here.  Jsut be sure you aren't planning a press conference or something and have essentially no time left for deployment.

by george.p.alexander

We'll be taking a look at different .Net Deployment techniques which are part of .Net development framework. What are the techniques that .Net developers have at hand when it comes to the important requirement of deploying their project - a Web based project, a windows based project or even a web service? What about the level of customizability in these deployment alternatives?

This post originally appeared on an external website

by sirlanse

The bias against VB is based on old stereotypes.  If you tried VB prior to 7, the advantages for C++ were great.

C++ was more complete and more consistant.

I am wierd, I was programming in VB and liked the changes for VB7 but was not thrilled with .Net.

So if you are comparing the languages in 2006, VB is more readable.  The newest versions of VB are as complete and consistant.  Unless you are writing "unmanaged" C,it all goes back to CLR.  Your C++, C#, J# etc will not run any better than the VB version. 

VB has also adopted a lot of the object.function syntax.

 

This discussion is closed to new replies.