MVSFORUMS.com

raj051076 · Beginner Joined: 05 Dec 2002 Posts: 64 Topics: 21

Well this may be a different type of question but I am a bit curious to know what the experts think on this. Here is the question..
Suppose you have five software development methodologies available to you to execute a project and they are

Pure Waterfall
Modified Waterfall
Evolutionary prototyping
Spiral
Staged Delivery

Now while choosing one out of these methodologies
what factors will you consider and how much does each factor contribute(in a scale of 1-5, 1 being minimum contribution and 5 being maximum) to the selection of a particular methodology.

While answering this question may I also request you to let me know your experience in software industry and your role(like project leader,team member etc) in the industry.

Thanks
Rajib
_________________
Rajib

semigeezer · Posted: Sun Feb 23, 2003 10:54 pm Post subject:

I don't know. In 20 years, I've never even had to write a specification (at least, not until the code was done). We always just sat down and coded until it worked Smile

The code wasn't perfect, but I doubt the addition of formal design procedures would have changed that. Granted, ours was a small team with experienced people and straightforward requirements, but at least in that environment, the "tell me what you need on Friday and I'll have it done by Monday" method worked great. Razz

raj051076 · Beginner Joined: 05 Dec 2002 Posts: 64 Topics: 21

Semigeezer that is another methodology called "code and fix" and it also has its own advantages and disadvantages. In big projects "code and fix" cannot be implemented.
_________________
Rajib

Manas Biswal · Posted: Mon Feb 24, 2003 3:58 am Post subject:

Rajib,

The answer to the question that you have asked is too long to answer here. I can however suggest you a good book for that. Go through "Software Engineering - A Practitioner's Approach" by Roger Pressman.

Regards,
Manas

Dibakar · Posted: Mon Feb 24, 2003 4:43 am Post subject:

Rajib,

I don't know about these methods but I found a Software Life Cycle Comparison Table.

Diba.

raj051076 · Beginner Joined: 05 Dec 2002 Posts: 64 Topics: 21

Hi Manas and Dibakar,
Thanks for the suggestions but my aim was not to go through books. I apologize here for not telling the actual purpose of posting the question here. I am trying to do a research on "how can you select the best software methodology" and so wanted to take some feed back from the experienced softaware professionals here.

Thanks
Rajib
_________________
Rajib

taltyman · Posted: Mon Feb 24, 2003 3:48 pm Post subject:

There's an old saying about software projects.
1. You can get it done on time
2. You can get it done on budget
3. You get what you ask for...
Unfortunately you can only get 2 out of the 3

However for those using software development methodologies then maybe you don't get any.....

Dibakar · Posted: Mon Mar 10, 2003 6:06 am Post subject:

Sorry Rajib,

I could not figure out the best method. It seems Taltyman is right, if you try to get all 3 then you don't get any.

zatlas · Beginner Joined: 17 Dec 2002 Posts: 43 Topics: 4

I am not sure what are these methodologies but if we believe our friends in Symantec (and allegedly other companies) the best methodology on earth is eXtreme Programming (XP). Basically good ol' "Code and Fix" with a lot of spin. But the real difference between this methodology (XP) and anything else our bosses have ever dreamed about and imposed on us is what they call "Pair Programming".
Imagine that instead of sitting alone in your cubicle and do some creative coding, the two of you are sitting in that cubicle and advising each other. Just please do not leave a kitchen knife in that cubicle.
I believe that all these methodologies are somewhat stupid if you look at them more than as guidelines. The XP shows that these methodologies could evolve to something evil so we should beware!
And as we were all have been warned by Antoine De Saint-Exupery: "Children, watch out for the baobabs!"
ZA

taltyman · Posted: Wed Mar 12, 2003 9:02 am Post subject:

This is Startrek and the Enterprise is in trouble. Captain Kirk says to Scotty 'I need to override the molecular readjuster'. Scotty then does an internal brain scan and says 'I can link the alpha particle generator to the scuzzy IO port but it's possible that we vaporize the warp drives when I do'. Captain Kirk then pauses, weighs the pros and cons and then says ' make it so'. A tweak may need to be added, but in the end the Enterprise is saved (or at least has a fighting chance).

Uh oh, in this episode the star fleet admiral and a consultant are onboard. The consultant suggests that they should study this problem using the Feebler Truxan methodology. Since this consultant is highly paid and the star fleet admiral has just finished reading about the Feebler Truxan methodology in 'Star Fleet Today'. He overrides Captain Kirk. Minutes later they are all vaporized.

In my world the boss says to me 'I need xyz, I need it now, make it so'.
As long as the admirals, consultants and methodologies stay the heck out of the way then we've got a chance.

raj051076 · Beginner Joined: 05 Dec 2002 Posts: 64 Topics: 21

Well Taltyman what you are saying may be true when you are supporting a system . I am talking about developing a system. What I have felt during my 3 years of developing system and 2 years of supporting a system that a methodology is very much required while developing a system otherwise following things happen

Schedule Slip - Methodology helps in tracking the progress of the project and tries to prevent schedule slip
Feature creep - Due to the unrealistic expactations of the users feature creep might happen and without a methodology(e.g, staged delivery or evolutionary prototyping) a project may never get complete with feature creep.

There are other benefits also for choosing a methodology. Code and Fix is not possible while developing a big system..thats for sure.
_________________
Rajib

taltyman · Posted: Wed Mar 12, 2003 11:30 am Post subject:

All software is code and fix at my level but then I'm pretty low on the scale.
Maybe a forum with a more 'project manager' audience would get a better response.
Just curious, is the 3 years developing and 2 years supporting a system... is that 5 years on the same system?
Could you give a very brief overview of those methodologies you refer to?

Frank Yaeger · Posted: Wed Mar 12, 2003 11:38 am Post subject:

[deleted by poster]

taltyman · Posted: Wed Mar 12, 2003 12:05 pm Post subject:

deleted by imposter

raj051076 · Beginner Joined: 05 Dec 2002 Posts: 64 Topics: 21

Well most of my projects were in mainframes. We in our shop has always followed the pure waterfall model while developing an application. I know 5 years is very less an experience to try some other methodologies and with my mainframe background I feel that pure waterfall methodology is what I will mostly use.
_________________
Rajib