|
|
Posted by Puzzled on 12/26/07 12:10
On Wed, 26 Dec 2007 02:32:52 -0500,
My Pet Programmer <anthony@mypetprogrammer.com> wrote:
>jcoder said:
>> starting from scratch..perhaps for better control. easy to maintain.
>
>Adding to that:
> - less links in the chain of possible breakpoints (subitem of both above)
> - syntactic choice in naming
> - performance tuning to your own specifications, rather than those of
>an unknown entity.
> - configuration issues
> - unintended consequences - without reading through every single line
>of code in the open source code, you can't know what it does to session,
>other parts of your code/objects, or how it handles SQL queries/Ajax
>calls, etc.
> - lack of bloat - You can spend a lot of time trimming down someone
>else's code for your own ends and removing extraneous bits (which may or
>may not break it), or write your own.
It might only be a personal failing, but any time I've tried to
save time by customising someone else's code, I've failed. Other
people's code always seems to be designed to a non-intuitive
model, badly commented, fragile, and with unnecessary couplings
that are fiendishly difficult to unravel.
I've found that, if I can use someone else's code without
changing it, I'm ahead. But if it wants modifying to meet my
needs, I'm generally better off writing from scratch.
Navigation:
[Reply to this message]
|