On Thu, Mar 6, 2014 at 1:18 PM, Joshua Marsh <jos...@themarshians.com> wrote:
> On Thu, Mar 6, 2014 at 11:37 AM, Levi Pearson <levipear...@gmail.com> wrote:
>
>> My issue is that we tend to get emotionally attached to technologies
>> like PHP that we had success with in the past.
> I don't do most of my daily programming in PHP. I have several existing PHP
> applications I maintain though. Many of them are 10+ years old. It would
> take several man-years to switch to another language. The people who pay
> for me to program don't want to spend the time or money doing that when the
> product works just fine how it is. The occasional enhancement or security
> fix is cheap to do but many of them require PHP to keep up with technology.
> I'm guessing there are far more many people who fit in this camp, than do
> in the zealot camp. For that, I'm glad that Rasmus is still riding the
> gravy-train and someone is paying him to keep working on his pet project.
Maintenance of legacy products is a perfectly reasonable reason to
continue to write PHP code. I was trying to emphasize that there are
fine reasons to write PHP code; in case that wasn't clear, I want to
repeat it here. Legacy code can be worth the maintenance cost.
Writing *new* programs in PHP is just going to create more legacy code
with maintenance costs, though. It digs you deeper in the hole of
bondage to a lousy language that many very good alternatives exist for
now. But if you're stuck with it, look to Facebook for tools to make
it suck a bit less. They've got an efficient JIT-compiling vm,
presumably with less internal brokenness than the official
implementation. They've also got static analysis tools (i.e.
lint-style tools, code visualization, etc.) for PHP as well as other
languages they use. You'd be crazy as a maintenance PHP programmer
not to take advantage of this stuff.
Personally, I'm not entirely convinced that Facebook made the right
call on their investment in PHP. I would guess that, with the
investment they've sunk into PHP tooling by now, they could have
brought a re-implementation of their page generator in another
language up to feature parity. But I have no special insight at all
as to how that actually would have played out; maybe it would have
been disastrous. The upshot is that all the legacy PHP programmers
elsewhere have reasonable tools to work with now to help them avoid
PHP warts.
--Levi
/*
PLUG: http://plug.org, #utah on irc.freenode.net
Unsubscribe: http://plug.org/mailman/options/plug
Don't fear the penguin.
*/
