NYCPHP Meetup

[Mitch Pirtle]

On Mon, Oct 26, 2009 at 4:29 PM, Ajai Khattri <ajai at bitblit.net> wrote:
> On Mon, 26 Oct 2009, Mitch Pirtle wrote:
>
>> Here's my take, would love your feedback:
>>
>> http://www.mitchitized.com/nerdy-stuff/php-frameworks-explained/
>
> A couple of points:
>
> You say a framework is opinionated (fair comment) but that to me
> implies a stack has specific layers, a naming and directory layout
> convention, etc. Zend doesn't have that so each project is different -
> there is no common frame of reference. Thus I dont call it a framework.
> "Glue" is certainly right, but its really a library of components - how
> you use them, how you structure your app and your own naming conventions
> is how they're often used. If you like gluing things together, that might
> be OK. But it also brings up the issue of long-term maintenance, when you
> have to hand the code over to someone else that has a different idea of
> how things should be structured. While a lot of frameworks have a common
> layout that facilitate easier maintenance by different developers, most
> also allow you to tweak and customize a little, so they are not as rigid
> as seems to be implied.
>
> The other thing is, despite being full stacks, many frameworks can be
> scaled up and there are many options to tweak that. You can switch off
> layers you dont want to use (like maybe you dont want to use an ORM or
> you want to swap in a lighter component). Many frameworks have a lots of
> options for caching functions, templates, etc. You would have to build
> that yourself if you want to glue something together with components.
>
> There are many large sites using frameworks, so your comments about
> scalability and performance might be a bit off. Frameworks require much
> more learning to get the best out of them, this is why picking four
> frameworks and bulding a simple "Hello World" page merely scratches the
> surface and cannot be regarded as a serious testdrive, but these
> unoptmized experiences are sadly often used as benchmarks to pick!

You just made me re-read my own article, as you make it sound I said
things I didn't. And that makes me nervous that other folks would
misunderstand as well.

This article is about the different types of frameworks, and only
cites a few as examples for reference. I have no idea what your
talking about regarding benchmarking frameworks with a typical "Hello
world!" test app. Lastly performance was specifically mentioned as a
challenge with scaffold frameworks, and as a strength with glue
frameworks.

You sure you read the right blog post Ajai?

-- Mitch, checking his website one more time