Some development opinions :0

Okay.  This may be somewhat a rant...so, you have been warned. 

I've noticed a lot of people get obsessed with performance.  This is kind of okay, provided your code is working.  I know people (we won't name names!) that will just fret on performance before the method they want to write even exists.  Perhaps this is just my point-of-view, but isn't it better to get X piece of code taking the proper input(s) and delivering the proper output(s), first?  You can get hung up on performance and refactor to your heart's content - but get it working first, please.

Second.  Don't go crazy typing before you think about anything.  Comment the process first, don't type any code.  When you're done, make sure the logic makes sense.  Why spend time coding something only to realize it doesn't even make sense after you just typed 30 lines of code?

Perhaps I've bene a bit more on-edge about this stuff since I've been reading the pragmatic programmer lately. 

Have a nice day :)

Print | posted on Wednesday, July 12, 2006 5:25 PM

Feedback

# re: Some development opinions :0

left by Tim Hibbard at 7/12/2006 7:12 PM Gravatar
I disagree. Performance and scalability should always be on the forefront of a developers mind. A piece of code should be perfect in the developers mind before they even open up the IDE. If you need a scratchpad than use notepad.exe, not the raw code.

# re: Some development opinions :0

left by Sanjay at 7/12/2006 8:35 PM Gravatar
Hmm...Perhaps I should have swapped number one and number two in my initial post :)

Nevertheless, there are numerous instances when I've encountered someone hung up on a performance of say, a query, when said query isn't bringing back the correct results as is.

For overall application design/architecture, I agree performance should be [one of the] priorities, naturally. Furthermore, I am in no way suggesting that a non-performant (or less performant) version of any code should be <i>rolled out</i> What am I saying, is that performance isn't relevant until you have a method/function...[a relatively small piece of] code that <i>works</i> Besides, where else would a benchmark be derived from? Naturally, this is barring instances where it's simply best practice to perform a task in a certain way...Of course, if you are aware of a best practice, chances are you will implement it if possible.
Cheers
Title  
Name
Email (never displayed)
Url
Comments   
Please add 8 and 5 and type the answer here: