Re: Jamaica roadmap brainstorming

"[email protected]" <[email protected]> Sat, 09 May 2009 11:35:45 +0200
Newsgroups gmane.comp.cms.xaraya.devel
Organization Xaraya
Message-ID <[email protected]>
Sorry for the delay in responding

Marcel van der Boom wrote:
> Some quick remarks/feedback:
> 
> - perhaps we should define terms like track,version,scenario etc.
> - although the numbers are for reference only, i bet many will interpret 
> them as release numbers, to avoid confusion, best re-letter them to 
> their track maybe?
> - is the idea to model this in trac.xaraya.com again or something else?
I like the trac visuals, and once a real roadmap is decided on I'll 
probably stick the milestones and such in there
> - do the numbers as they are now represent the order of 
> *implementation*? (as opposed to releases) ?  for example 2.5 seems to 
> be the first one to be done to me where as 2.1 sounds like something 
> ongoing (unless we know what 'completed' means)

No, they don't represent order of release/implementation. This is just a 
brainstorming, as the name suggests, so perhaps I should have just 
listed the functionality points to avoid confusion.

I agree that 2.5 is more fundamental, and we are in the happy position 
of already having a lot of the groundwork done. Nonetheless I want to be 
pragmatical, given the scarcity of resources. 2.5 could be triggered by 
several needs downstream, for instance the drive to integrate a js 
framework, the desire to create an "official" solution for multisite 
installations, the need to find a common home for an increasing number 
of dataproperties, and so on. However, until these become acute on the 
horizon, it makes little sense to spend a lot of time on designing the 
ideal file system. This is especially true when there are other easier 
wins out there. I try to look at the functionalities like a PERT 
diagram. Once something winds up on the critical path, it needs to be 
attacked and resolved.

But all this is theory. The first step is the list of functionalities. 
Once that is more or less decided on, we have a context within which we 
can juggle needs, resources and priorities.

> - functionality wise, no (additional) feedback.
> - perhaps a few words on tools
> - what is the development process? i can imagine it's hard for people to 
> 'dive in', especially core development. 

Can I be blunt? My dream would be to have anyone new to 2x and wanting 
to contribute to the core to begin by writing a number (say 10) of unit 
tests. I'm happy to show folks examples on how to do it, and I will 
publish the tests we already have as soon as there is a permanent 
official 2x demo site. (hello, PMC?)
The mere act of doing so will familiarize folks with the workings and 
standard practice of the code and make it easier to contribute.

Would it be possible to create a
> 'quickies list' which people can opt in to do?

See my next post.

Marc