IT Employment

General discussion


What Should I Teach?

By ITInstructor ·
What is the most important soft skill and technical skill for IT professionals to have? I teach at a small community college and this info help me prepare classes in the future?

This conversation is currently closed to new comments.

Thread display: Collapse - | Expand +

All Comments

Collapse -

Be fair

by Tony Hopkinson In reply to Great observation!

Things don't always get revisited because someone in IT made a cod of it. An important thing to teach is if you're are going to design a pit fall, knock up a ladder at the same time. There are many many reasons why the pit fall was a logical idea either from a design, current business model or even god forbid financial reason. Cutting some steps so you can get out of it yourself is going to get you a lot of brownie points when you get thrown in it though. They don't have to be good steps, lashed together accountants femurs will do.

Collapse -

Heheh your right of course

by dafe2 In reply to Be fair

What I was actually trying to well, you said it better. :-)


Collapse -

At least do it for a reason

by jdclyde In reply to Be fair

If you need to take a shortcut, at least do it for a reason not just because it is easier on you right then.

There is a BIG difference between having to make do with what you have and just doing crappy work.

Collapse -

As Merry said to Frodo

by Tony Hopkinson In reply to At least do it for a reas ...

"Short cuts make for long delays"
Unfortunately as a business software developer you cannot avoid making them. A designer who takes a short cut so he can spend more time browsing is a total a**e, one who takes one, because say sales set the project deadline, is doing his job. An experienced designer, would highlight the shortcut and leave some directions of how to get back to road just in case the it ended with him in Sauron's fist.

Collapse -

Not using object, component or container based architecture?

by james.chau In reply to Pride in a job well done

Real life problems require real life data, who can prepare these structured or un-structured data? You should realize that you cannot debug the various real life situations if you do not have the required test data.

OO, component or container based architectures all deploy an iterative development approach while what you are proposing here is the old classical waterfall approach, albeit on a smaller scale ? Regards.

Collapse -

RAD = OO ????

by Tony Hopkinson In reply to Not using object, compone ...

You can manage an OO development with a Classical Waterfall PM approach
Equally you can do an iterative project with Non OO tools.
One has does not preclude the other.
A big problem in the industry is the perception that non waterfall approaches are unmanaged ones, outright nonsense that. Its just that accountants and non-technical managers are happier with the large volume of numbers that come out of a classical design. The fact that these numbers were estimates of estimates based on guesses about a fantasy was unimportant.

Collapse -

are we crossing subjects ?

by james.chau In reply to RAD = OO ????

Waterfall is older and certainly it supercedes most if not all other development approaches since familiarity is best; but when you have that much quality definition why not just stick them into WebWhatever App Server and just generate the horde of code and voila, project finished (Information Engineering / Case Tools sound familiar to you ?)

What's at issue is the delivery of production code that works and is of good quality which can also be reused thru newer techniques like OO or component.

Waterfall is not un-managed, it is hyper-managed.

Again, we are crossing subjects since deliverables is the topic and you seem to be talking about project planning...

Collapse -

Cross threading

by Tony Hopkinson In reply to are we crossing subjects ...

Never happened before, but I may have read what I wanted to read instead of reading what you wanted to write.
It just seemed from your post you were linking OO to RAD. From your reply you weren't.
Did I say waterfall was un-managed ? Have to lay off the drugs, talk about reality check.

Where I work we choose a lifecycle based on the type of deliverable and the customer. Some methods work better than others, and sometimes development efficiency has to be sacrificed to external pressures. Neither method should ever be un-managed, short of extreme luck, that's your next financial disaster.
OO is not always the most efficient way of doing something. It lends itself extremely well to a lot of business IT, but there are cases when the overhead of objects is unacceptable.

Collapse -

You have hit the head of the nail right on !

by james.chau In reply to Cross threading

I wish to add my little 2 cents by saying that the budget should dictate the ADLC (application development life cycle) since budget is usually the ultimate reality check for most situations... cheers.

Although OO is not perfect, it is the closest to reality we have got through software simulation.

Collapse -

I love OO

by Tony Hopkinson In reply to You have hit the head of ...

I've been thru just about every goddamn modelling process there is for verifying your design with the requirements. Everything from flow diagrams to the Vienna Development method.
The best thing about it isn't the technical aspects, after all OO developers have very different ideas how that should be done. It's because, it's the way our customers think and the extensibility means you can cope with ambiguity in the detail, and even the odd out right snafu.

Obvious in retrospect, but damn clever.

I use OO models and then translate them to a classical design, it's easier to present and verify and it gives you a very clean foundation, the one you were trying to aim for with traditional concepts.

Related Discussions

Related Forums