[BW-dev-discussion] Is it important to adopt a new framework before proceeding?
peter.e.lind at gmail.com
Sun Aug 22 14:47:45 CEST 2010
I would offer the following idea:
1) keep the old system in place
2) implement a new framework on the side
3) code every new feature in the new framework. Migrate existing
features to the new framework as needed
This avoids a big-bang strategy (one of the most likely risks of
failure for the project) while easing developers in.
I would go with Zend or CodeIgniter, personally - haven't used CakePHP
enough to say if it's great or not, so could possibly serve the
A few comments on Claudios post, inline:
On 21 August 2010 17:36, claudio pacchiega <pakkio at gmail.com> wrote:
> 1) Zend: I studied its documentation and found it quite difficult to learn
> and very complex in itself. Doing even simple things costs a lot.
There's a minor switch to the Zend frame of mind when you take it up,
but it's pretty quick and once you get settled there's no extra cost.
I'm assuming you're not speaking about performance, as Zend can be
optimized to run as fast as any of the frameworks (and apart from
that, any comparison of frameworks is close to meaningless unless
you're basing it on actual benchmarks done on systems with
> 2) CakePHP is quite interesting because it is the PHP implementation of RoR,
> which is in my opinion one of the best frameworks out there, relatively easy
> to set up
RoR is - to the best of my knowledge - great for getting started
making applications. It's a bad framework to use for actual systems in
production where performance matters. I don't know if the same thing
translates to CakePHP but I would be wary of going for frameworks
based on the RoR way of doing things.
> 3) CodeIgniter appears the more flexible since it puts the lighter
> infrastructure and thus being the most rapid out there.
> see http://avnetlabs.com/php/php-frameworks-revisited-codeigniter-vs-zend-framework for
> a well documented insight on this.
Interesting point: I18N is marked no for CodeIgniter and yes for Zend.
That may be something to consider as well (currently the translations
are one of the bleeding soars of BW - the page after you've logged in
has about 130 database queries for content and translations) as - for
some involved people at least - the translations/internationalization
is the most important aspect of BW/Servas.
Overall, there's incredibly little php-side caching going on in BW,
which is something to look at when implementing a new framework (most
should be done in objects for easy reuse and querying - not a
framework specific thing).
So, yeah, Zend or CI would be my 2cents.
WWW: http://plphp.dk / http://plind.dk
More information about the bw-dev-discussion