Everyone should read this. I think it represents a shift in thinking that is (finally) taking hold in our industry. For a long time I’ve been arguing to make software less complex – less code = less bugs. Obvious code = maintainable code.
I once heard a rumor that air crew on commercial airlines had this quirk in their conditions where they got the public holiday loading of all the territories that they flew over. For example, if they were a Brisbane based crew flying to Adelaide and they flew over New South Wales on a public holiday then they were entitled to public holiday loading.
Now I don’t know if that is true or not, but given that I am often working interstate I sometimes have trouble figuring out which public holidays I should observe. For example, from time to time I’ve worked interstate whilst my home town had a public holiday so that I could fit in with a clients scheduling with a view of having a day off at some other point in time.
I don’t suppose it really matters. This coming Monday it is Labour day in New South Wales and my client’s office won’t be working. As a result I’ll be spending most of Monday in Melbourne. I figure that it might give me an opportunity to catch up on a few side projects and tidy up some loose ends on code I have shipped recently.
It looks like Facebook was listening. Thanks Facebook!
So I spend a little bit of time traveling. So far this year I can count catching around 59 flights and one of the little frustrations that I come across on a regular basis is when we are asked to board simultaneously from both ends of the plane based on the their seat number.
In theory this should lead to a quicker boarding process, but what invariably happens is that:
- Someone has a medical condition that board from the rear.
- Someone can’t following simple instructions.
If I was being kind I’d say that number one is the more common scenario, but more often than not its number two (because I see just as many people trying to get to seat 5F by boarding the rear of the plane!!
I’m not sure how they tested dual boarding procedures, but I suspect that they probably got a whole bunch of hosties and practiced boarding the plane. The problem with hosties is that they they probably got it right every time.
I think that if they built a computer simulation and put a random (1) and (2) amongst the passengers they could see that the boarding process is significantly slowed down in a dual boarding situation.
… what would it look like? I’m kind of inspired by these guys at Microsoft who are getting around with the blue monster card. Australian exile Frank Arrigo is the latest to get his batch.
This is the current front-runner:
Rocky Lhotka posts up an interesting thread rebutting some feedback from people about CSLA.NET. Reading through the post I can see that in a lot of cases people are comparing with EntLib.
I haven’t used CSLA.NET so I can’t really comment on that aspect, but I will say that anyone thinking of adopting any framework needs to consider this:
- How much leverage is the framework really going to give you?
- Can you already do 80% of what you need out of the box with .NET?
- Do you really need the other 20%?
- No really, do you need it?
Finally, I think that a lot of people when adopting frameworks seem to forget what the .NET framework has out of the box to allow you to build a lean mean application.
- System.Diagnostics (for all your logging needs)
- Trace.TraceInfo(…) is one of the most methods around.
- System.Data (for all your data access needs)
- I’ve been using System.Data.Linq lately and I love it.
- I can haz configuration data!
I’m tired of seeing simple applications made bloated simply because our “standard” is to use some random framework. Teams that have time to think about that stuff are probably building more than they need anyway.