[theforum] weo 2b: read me 1st

[Martin Paul Burns]

Dave wrote on 08/11/2004 21:35:00:

> Martin Burns wrote:

> > Given that the current model is a bespoke system, to replicate it
> > really *will* take a ground up rewrite if it's in anything other than
> > the current technology.
>
> Oh, I see where we're disconnecting, here.  See my post above in reply
> to Madhu.  Replicating WEO on Linux need not necessitate rewriting... I
> have a copy of ColdFusion for Linux, and experience porting stuff to it!

Ah, now. Do you have a spare license that evolt can use? Because otherwise
we're somewhat stuck. Paying for a CF/Linux license we're only going to use
for a short while is a non-starter.

> Redacting is in the cards, to be sure.
>
> A. porting WEO exactly from ColdFusion/Win32 to ColdFusion/Linux is a
> very straightforward task

Actually, to be accurate, it's *back* from CF/Win32 to CF/Linux.

> B. rebuilding it from the ground up to be exactly the same, but in PHP,
> Perl or Python is also a very well defined goal, although it would
> presumably take much longer
>
> C. but just the *specification* phase for completely new system, with
> new rules and workflows, even based on an existing CMS, will surely take
> longer still than either A or B because, until the spec is done, we
> cannot even estimate when the system will be completed (...a wise man
> once said, in his article on evolt) :-)

Or alternatively:

D. Build something that's similar to the current system, based on the
available
   functionality of $tool. At its heart, the current system is *extremely*
simple:
   it's the simplest of workflows, with knowledge of users. While you might
lose a
   few bits and pieces like ratings if there isn't an extant module to do
it, you'd
   get enough to be able to launch with. To do this with pretty much any
half-way mature
   CMS would be cake. The majority of the time spent would be in wiring in
the presentation
   layer and migrating the data, not in coding functionality.

I think this is what John's proposing, and it makes a *lot* of sense to me.
I don't know
Drupal well enough, but certainly it could be done in Plone in a week. I
would assume the
level of effort for Drupal will be of a similar order of magnitude.

(Actually, I produced such a thing as a somewhat crufty PoC in Plone in the
current design in under a week while learning Plone at the same time, and
Seb Potter did a much better PoC with the 'new' design in a couple of
evenings, with the addition of some bits of functionality and  migration of
data)

Cheers
Martin