[BW-dev-discussion] Is it important to adopt a new framework before proceeding?

[Jean-Yves Hégron]

Hello John, I receive the mail trough BW Dev Team mail box only once ,
perfect.

Not easy to make a choice between A) or B).

The EF is not impossible to learn (I learn it), techically speaking it
does the job well. The biggest problem is (as said Peer) that it is used
in very different ways. If we are to use it, it is really needed to
define what is obsolete and forbidden for new features and what is not.

At my work place "my" programmer team use with cake php, they look to be
happy with it. One weak point with it (but I guess it is the same for
most framework and the price when they try to be database solution
independant) is performance for database. It is simply not efficient and
for a few case I have had to ask few critical programs re-wrote in a
more classical way, but this was on few cases.

Zend looks more popular than Cake, but it is also possible that it has a
better marketting (IBM partner). It is also often said that Cake is
easier to start with (Keep in mind we are not doing such a complex code
that we need the latest tools)


A complete switch will really cost a lot of delay to the project. But I
think that it will be needed to switch to a more supported framework
soon or later.



There is one thing I am sure (since I don't want to see again the bad
choice who was for BW Rox): it is the people who will really program who
are to make the choice.

For myself, EF cost the less effort, but I will learn Zend or Cake if
needed.

JeanYves










Le samedi 21 août 2010 à 10:02 -0400, John Gunther a écrit :
> Hi everyone. If anyone is getting this twice because they're on the 
> BW-dev list, please let me know so I can remove your redundant address. 
> Likewise, if you know of anyone who should see this thread, please send 
> me their address. I assume Matthias and Jean-Yves are getting this 
> through the list.
> 
> While Gary and Thomas are honing their project leadership skills, I'd 
> like to continue the discussion about the existing BW MVC framework and 
> reach a consensus among the technoids as to whether it should stay as is 
> or be replaced now with something else. If we do this now, the actual 
> decision can easily be made at the proper moment.
> 
> My summary of what I've learned so far -- and please correct anything 
> you disagree with or that I misunderstand -- is:
> 1) The existing framework (EF) has evolved from a base product, Platform 
> PT, that currently seems to be undocumented, unsupported, and not 
> comprehensive.
> 2) The EF has had some additional pieces grafted on by BW.
> 3) The EF is not problem free.
> 4) The learning curve for writing Rox code using the EF is very steep 
> and new coders will have to essentially be tutored to come up to speed. 
> Peter's description was, "It is simply horrible. Think Frankensteins 
> monster (it's composed of three different, at times somewhat 
> incompatible, parts)."
> 5) It's not clear who will have the time or energy for such tutoring.
> 6) We need new developers -- not just for some immediate Servas-specific 
> features, but over the long haul to keep the site growing into a 
> comprehensive, flexible, and ever-more-worthy open source hospex 
> platform. The promise of open-source is greatly reduced if it's very 
> difficult to contribute compatible code.
> 7) Adopting a new standard framework now is a significant task. However, 
> it would save the Servas coders from learning the EF, and would lower 
> the learning curve for future coders because they are more likely to 
> know or be able to self-study a documented product than the EF.
> 8) A standard framework product can be expected to evolve without all 
> the effort being provided by BW.
> 9) It will require effort, but not be an enormous job, to leave the 
> existing code largely unchanged, while requiring new code to use a newly 
> adopted framework.
> 10) I use framework principles in my work but have never used a 
> framework product, but I'm ready to do so now.
> 11) Peter has suggested CodeIgniter, CakePHP, and Zend as the most 
> feasible choices. Having checked them out, I suggest that Zend is 
> slightly preferred, largely because it is so closely associated with the 
> people who wrote and advance the core PHP technology.
> 
> So the question is:
> Do we:
> a) keep the EF, or
> b) adopt a new product?
> 
> If (a), how are the Servas coders going to learn it?
> If (b), which product and who can head up the transition analysis?
> 
> Can the all current and likely technical contributors reach a broad 
> consensus here?
> 
> John
> _______________________________________________
> bw-dev-discussion mailing list
> bw-dev-discussion at bewelcome.org
> http://bewelcome.org/mailman/listinfo/bw-dev-discussion
>