Comments on: Programming: you’re doing it wrong

By: Encouraging Open Source Contribution | BrandonSavage.net

Encouraging Open Source Contribution | BrandonSavage.net — Wed, 07 Apr 2010 11:08:00 +0000

[...] Tabini pointed out that architecture doesn’t matter as much as having code that works. This is a fine point, and often times is true. But if you ask the average PHP developer if they [...]

By: Oliver

Oliver — Thu, 25 Mar 2010 03:12:10 +0000

Its a bit sad to see that people fully agree with your post, as I doubt you mean it exactly the way its written. ;-)
The best architecture won’t help if the product’s not working. On the other hand, if one writes code like, sadly, many PHP developers do, you will waste heaps of time on maintenance, because you break stuff all over, again and again.

I’d rather not use copy&paste and filesystem wide search and replace, but good architecture. ;-)

By: Kinanaskomitin

Kinanaskomitin — Thu, 18 Mar 2010 03:40:56 +0000

Happy 20th anniversary!
I’ll have my 10th anniversary this October, it feels great! And I agree with you! Keep posting!

By: rjw

rjw — Tue, 16 Mar 2010 19:40:25 +0000

Excellent posts. And John Kary’s points resonate also. Having read both posts, I have to agree with what is raised on both. While I am a big advocate of minimal, simplistic designs to get your release out in a quality driven manner, I can’t help but agree with Marco in the area of ego winning over pragmatism (if I interpret him correctly).

I have seen projects suffer in the past simply down to, imo, ego – on both sides of the fence. Ego encapsulated in the programmer who thinks half way through the project that we must start again because we should have implemented pattern x at the beginning, to the other end of the pole where senior programmers defend their borked designs and hence everyone suffers the maintenance headache / cost – and end up ‘hating their job’ … as I have come to do in the past.

Know your subject, remember the deadline, don’t leave a mess that you wouldn’t want to revisit.

By: I May Be Doing it Wrong, But I Have My Sanity | John Kary

I May Be Doing it Wrong, But I Have My Sanity | John Kary — Tue, 16 Mar 2010 15:15:48 +0000

[...] Tabini recently wrote on the php|architect blog that “programmers are doing it wrong.” He argues that many [...]

By: fsilber

fsilber — Mon, 15 Mar 2010 15:24:26 +0000

Many of the new things Martin Fowler advocates are to facilitate the creation of automated unit tests. Lots of programs in production don’t have unit tests; they work, but they are more rigid and resistant to change as a result.

By: AwesomeBob

AwesomeBob — Mon, 15 Mar 2010 07:05:33 +0000

I don’t agree with everything. I think that code should be written to meet a few criteria besides just “working” to get the job done.

1. Make it easy to maintain. I hate searching through poorly written apps to extend/fix components that are failing or need to be improved because some brains upstairs want things to perform differently.

2. Write so people can read it. Don’t save yourself some time by writing unorganized hard to read code so someone else needs to spend longer to figure out what the code is suppose to do.

3. Reasarch the right way to do things, HTTP_POST_VARS isn’t used anymore, use $_POST and don’t forget to sanitize your input data with actual functions like mysqli_real_escape_string not add_slashes or your own custom function.

It does take longer to work with poorly written code after it’s written than to just take some extra time to write it correctly.

By: Michele Mauro

Michele Mauro — Fri, 12 Mar 2010 18:58:16 +0000

Fowler wrote lots of clever and interesting stuff, and he is right most of the time. The problem you talk about is probably more related to the fact that Software Engineering is not Engineering at all (despite what my boss and my own Phd say :-) ), and that software is such a vast and endless space that different methods and opposite philosophies have still enough space to be true, in their own neighbourhood. Except that a neighbourhood can be a multimillion industry with hundreds of thousand of professionals. And there are many of them industries (just think embedded vs. web programming).
So, it’s not (only) that programming attracts egomaniacs: every field has them. Software Development is so big a jungle that many of them can shout at the same time, don’t hear one another, and be right at the same time.
By the way, don’t say His Holyness name in vain another time. Read at least twice “UML distilled” and you’ll be forgiven. :-)

By: Paul W. Homer

Paul W. Homer — Fri, 12 Mar 2010 15:20:23 +0000

What makes that older code great is that is it simple and straight-forward. Coders sometimes get caught up in these crazy uber-complex ideas that ultimately aren’t added enough to offset their complexity. Others swing too far the other way, and just make the code as brutally specific as possible. Neither approach works; neither is right.

Still, just because someone’s managed to get something working doesn’t mean it is good. You really have to factor in the lifetime of the code, and it’s success or failure in making its user’s lives easier. There are great examples of really crappy code that is awful to use, yet extremely popular. The world would actually be a better place if these were replaced, but we cling to them out of habit.

By: Chris

Chris — Fri, 12 Mar 2010 14:26:43 +0000

You won me at:
“When I open a PHP file inside the WordPress codebase, I feel like I just boarded a time machine that’s taken me back to the way we used to write PHP code six or seven years ago—and don’t even get me started on what I think of Drupal.”

Well said.