[PLUG] Joe Jobbed

Ronald Chmara ron at Opus1.COM
Thu Oct 11 04:49:52 UTC 2007


On Oct 10, 2007, at 8:10 AM, Brent Rieck wrote:
> Carlos Konstanski wrote:
>> I'm no PHP insider, but I believe it.  Fixing 1000 bugs is not
>> uncommon in a release cycle of a large piece of software, and neither
>> is introducing 1000 more.
> Well, I don't know how many bugs the PHP people added but we can hope
> it's not 100x the number of named features added.  And the number of
> bugs fixed appears to be somewhat less than than 1000:
> http://php.net/ChangeLog-5.php#5.2.4

Much of that is pulled from bugs.php.net, which is the normal, public  
facing, bugs server, as well as rough developer logs and CVS  
checkins, and kind of glosses over, uhm, a few things. ;)

So, the PHP development model is kind of a blend between wikipedia,  
and linus' kernel. It's like linus' kernel in that there are *core*  
pieces that very few developers get to patch and modify. It's like  
wikipedia in that there are whole chunks of php's CVS where user abcd  
checks in an updated version, user bcde checks in an update to the  
update, user cdef checks in a revised update, etc., over a matter of  
hours or days.

So, what is publicly listed as something simple like, oh, "Fixed  
array handling bug" can actually be anywhere between 1 and /n/  
patches to the existing codebase.

While it might be more informative to list *every* change to the  
whole of the php codebase, that can be up to 100 different changes.  
Per day. Listing 3,100 patches/reverts/changes each month is not very  
useful to a web developer who just wants to get a current, stable,  
version of PHP for their blog.

-Bop



More information about the PLUG mailing list