Out Of The Mouths Of Clients

[suityou01]

This is what SY does all day - spout this stuff in emails and not only do no work himself, but distract everyone else from theirs.

It's a living.

[LondonManc]

I typed it while waiting for my chicken kievs to cook. I'm glad you're impressed.

"typed" is such a vague word - copy and pasted, plagiarised....

[DimPrawn]

No way did you write that all by yourself.

He got AndyW's mum to help him with the big words.

[suityou01]

He got AndyW's mum to help him with the his big words ****.

FTFY

[DimPrawn]

FTFY

FTFY isn't funny anymore. Ask Admin.

[suityou01]

FTFY isn't funny anymore. Ask Admin.

I'm so out of touch

What's the latest CUK vogue?

[darmstadt]

Seeing as most of the stuff I work with looks like:

SR    R15,R15  put zero into register 15 (return code)
BR    R14      branch to the address in register 14 (return to scheduler)

Then I have absolutely no idea what any one is talking about




(the above piece of code has been around for around 40 years and is probably executed hundreds of thousands of times a day)

[MrMarkyMark]

What's the latest CUK vogue?

There has been a couple while you have been gone:-



And this

[LondonManc]

I'm so out of touch

What's the latest CUK vogue?

Being offended by everything but reporting nothing

[SpontaneousOrder]

Agile can and does work well, when coupled with decent dependency injection framework, mocking etc.

That's an odd thing to say. Not really got anything to do with 'agile'.

The problem is that as there is no design per se, from a software architecture point of view what starts out as a good design can rapidly entropy into a ball of mud.

*can* being the key word. That aside, one of the main points is that it tends to be large up-front designs which lead to the big ball of mud. Without agility, it's hard to change direction when the design proves to be flawed, or simply becomes out of date. 'Agile' does not say that there shouldn't be a design. Just that commitments should often be deferred as late as possible. If anything, 'agile' mandates better design - i.e. with the future not being set in stone one must design with a mind to modularity & plug-ability being essential.

You cannot convince me that an air traffic control system is built using agile methods. Then again, we seldom work on projects large enough to push the boundaries of agile development.

The difference is that part of having agility is assessing the benefits & costs of the things we do. For many projects suited to agile development (which is a significant majority) the odd defect being released into production is a price one is very happy to pay - especially as the continual leak of small defects is often preferable to a rare leak of a defect which may, by that point in time, be very difficult to address & especially so without compromising the existing code architecture.

Air Traffic Control cannot afford to have any defects. It's much like the choice to be made between eventual data consistency, or some other model. There's a cost & a benefit - and the right price point is entirely context dependent. Calling it "[fr]agile" is silly. It's like choosing between a hard but brittle metal, and a softer metal, depending on it's utility. ATC are happy to spend the money required to make sure they never get any defects - and that price gradient isn't linear due to the fact, already mentioned, that more definitive design introduces more risk. If you want no defects and no risk in the design, then it costs.

Also Air Traffic Control is not competing in a highly dynamic market. So that's a whole hemisphere of agile development for which ATC sees no benefit.

Some of the biggest cluster****s I have ever seen involved agile and offshoring. That is the sweet spot for those "ajax and squeejy mop" clean up contracts.

Given that you're presenting an agile / not-agile dichotomy - then by 'some of' do you really mean 'most of'? Also worth noting that your typical offshoring setup is VERY likely to be antithetical to agile development. I.e. just because lots of people are tulip at it, doesn't necessarily mean that the paradigm is flawed.

and the industry is turning out scripters that can write unit tests that pass.

What does that bit mean?