On GameFAQs: Is it OK to lay my Wii down on its side?
BNET Business Network:
BNET
TechRepublic
ZDNet
TalkBack 8 of 34:
Next »
« Previous
Like handcuffing your code to specific hardware
That Roger Ramjet guy had it right: It's just another trick. In
nearly 30 years of programming I've been thru lots of
"revolutions" - structured programming, OO, scripting, Use
Cases, Patterns, Test-First, XP. The only one that yielded more
than a few percent productivity improvement was the big jump
from assembler to compiled languages. That's it! And even
THAT was not a 5-fold increase! 2X, if that.

This will seem a bit off-topic, but I get so tired hearing about
Yet Another Magic Bullet that isn't.

When will we get tired of all the hype surrounding these sham
revolutions, and take a cue from more mature engineering
disciplines? The real productivity killer is not in the initial
design or development or testing - it's the long-term
maintenance costs, and the near non-existence of industry-wide
(black-box) code reuse!

The two factors that have always plagued sw development, and
are still with us today, are these: Dependencies (brittle code
structures), and Assumptions (unexpected runtime states).
When code evolves from a snappy demo to a living product, it's
usually problems of these two types that begin cropping up.
Without slaying these twin demons, software development will
continue to be more of a craft than an engineering discipline.

Dependency Injection frameworks are just beginning to address
the former, but state assumptions are still The Elephant In The
Room. Until these are solved, code maintenance will continue to
be hell, black-box reuse will be practiced by only a few hermits,
and the revolution will not be televised.

And no, Ruby, with or without Rails, is not the revolution. Just a
faster way to generate ephemeral code.
Posted by: Robert Kohlenberger   Posted on: 11/01/05 You are currently: a Guest | Members login | Terms of Use

Alert moderator to an offensive message

Subscribe to this discussion via Email or RSS

I brought this up  Roger Ramjet | 10/31/05
PAPPL?  jpratch | 10/31/05
Its just another trick  Roger Ramjet | 10/31/05
I don't mind, with some reservations  Mark Miller | 10/31/05
Debugging tools for PHP  eheimer | 11/01/05
So what's wrong with Fortran?  Gravitas@... | 10/31/05
Like what?  Not average Joe | 10/31/05
Like handcuffing your code to specific hardware  Robert Kohlenberger | 11/01/05
Right on the Money RK  tbbrickster_z | 11/02/05
Productivity writing assembler?  cgraham_z | 11/03/05
Agree that Database Procedures are Evil  daver_z | 10/31/05
Maybe for small applications  Yensi717 | 10/31/05
OOD  Gravitas@... | 10/31/05
I think  Yensi717 | 10/31/05
Spot on right  Justin James | 10/31/05
Yeah, well...  emofine | 10/31/05
Varying opinions  tero_t_vaananen@... | 11/01/05
Right on.  zztong | 11/01/05
Multiple database vendors  Yensi717 | 11/01/05
Not a pain to edit  Chad_z | 11/01/05
Great framework  bkatz | 10/31/05
Looked like a pain to me...  Justin James | 10/31/05
For Unix deployments you might consider...  Mark Miller | 10/31/05
Message has been deleted.  khakman | 10/31/05
I've heard of Ruby  CobraA1 | 10/31/05
You mean like Smalltalk?  wkharold | 10/31/05
Funny  John Carroll ZDNet Moderator | 10/31/05
true  jimk_z | 10/31/05
true  jimk_z | 10/31/05
Ruby is OO  tero_t_vaananen@... | 11/01/05
Just what are people spouting about?  jacec | 10/31/05
Are there other benefits?  DougOfCBS ZDNet Moderator | 11/01/05
This is not serious development  SQLDBAJames | 05/29/06
Not everything is a nail  rarsa | 11/03/05

What do you think?

SponsoredWhite Papers, Webcasts, and Downloads

advertisement
advertisement

SmartPlanet

Click Here