[dmd-internals] Planning software?

[Brad Roberts]

On 1/18/2012 2:48 PM, Leandro Lucarella wrote:
> Andrei Alexandrescu, el 18 de enero a las 13:14 me escribiste:
>> On 1/18/12 12:26 PM, Leandro Lucarella wrote:
>>> Alex, el 18 de enero a las 18:46 me escribiste:
>>>> It's simple: It usually goes in both. You file a bug/enhancement
>>>> request/whatever on Bugzilla. People then post on Trello when they're
>>>> working on it, adding comments/changes as they make progress.
>>>
>>> So now to know the state or a bug I have to see 2 sites. Great!
>>
>> We have in-house task management software at Facebook that is
>> separate from (and virtually not integrated with) the issue tracking
>> software. We use it to great effect, and based on that experience I
>> can tell it would be beneficial (I'd actually say "basic survival")
>> to use something similar for D.
>>
>> Leandro, your thoughts and ideas are always welcome, but they seem
>> to come straight from the viewpoint that we're good as we are. As a
>> participant to this project, I can tell we're not. Me must get
>> organized in 2012 as an essential matter.
> 
> No Andrei, you're wrong, take probably the biggest opensource project
> ever, Linux, as an example and realize that that's not **necessary** (if
> you think Linux is a completely different piece of software, take Python
> as an example). I've been always *pushing* ways to be more organized in
> the development of D, so I completely agree about the goal, I just feel
> like you might not be moving in the right direction.
> 
> Anyway, I will just shut up because I have no idea about what is Trello
> about and have no information whatsoever about do you plan to use it (is
> there any place where I can find this discussion?). If bugzilla will
> **really** stay the same (I wish it would get much better though, maybe
> that's why you feel like you need another tool), I will not complain
> anymore (same if it get *replaced* by a better alternative for that
> matter).
> 

No, he's not wrong.  There's little value in taking the term 'necessary' to the extreme end of the scale.  That there
are examples of projects/products that can survived with little if any organization and planning (and I posit that
you're wrong about linux itself.. it has a LOT of structure within the organizations that contribute the most man power
(redhat, being a great example)) is irrelevant.

Having roadmaps and planning out what the priorities are _is_ important to the point of being essentially necessary to
have any semblance of professionalism.  The D, and more specifically DMD, development history is a fairly continuous
stream of hoping from one area of excitement to another leaving many of them incompletely implemented.  The addition of
more people contributing to DMD has done wonders for addressing some of the half done stuff (don's work on CTFE is a
perfect example).