Complete.Org: Mailing Lists: Archives: freeciv-dev: December 2001:
[Freeciv-Dev] Re: freeciv 2.0 spec
Home

[Freeciv-Dev] Re: freeciv 2.0 spec

[Top] [All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index] [Thread Index]
To: Andrew Sutton <ansutton@xxxxxxx>
Cc: freeciv development list <freeciv-dev@xxxxxxxxxxx>
Subject: [Freeciv-Dev] Re: freeciv 2.0 spec
From: Raimar Falke <hawk@xxxxxxxxxxxxxxxxxxxxxxx>
Date: Tue, 4 Dec 2001 17:13:16 +0100
Reply-to: rf13@xxxxxxxxxxxxxxxxxxxxxx

On Tue, Dec 04, 2001 at 08:51:15AM -0500, Andrew Sutton wrote:
> On Tuesday 04 December 2001 03:29 am, Raimar Falke wrote:
> > I want to answer with a question: which open source projects do you
> > know really have a design phase?
> 
> MICO (MICO is CORBA) had a design. it's doing pretty well.

I see design principles (good) and there are usage and sample programs
(nice but can't be transfered to freeciv). The only thing which can be
counted as a design document is the start of the chapter about POA.

> > > there seems to be some concensus that having a requirement/design
> > > phase is so bad that anything written that way isn't good (enough?
> > > raimar?)
> >
> > If you develop a part of freeciv (not as a whole) you do a design. It
> > took me one year to come to the conclusion that the thing which I call
> > agent is _the_ solution (at least to the problem I was working on). I
> > implemented three things which I would now call prototypes (a
> > standalone gui-less civbot in C from scratch, a python civbot with XML
> > and an early agent using the GTK client). It would be nice if I had
> > come to this conclusion earlier. But we both know that this isn't
> > always possible.

> that sounds about right, but why wouldn't there be an architectural
> design. something says how everything fits together? a
> blueprint. you're right though, design/prototype is how things are
> done. i'm not suggesting that everybody give up writing code
> alltogether.

Because people don't like to write these. The maintainer can't force
somebody and nobody can force the maintainers. Also
freeciv_hackers_guide.txt covers _some_ of this. Also note that we
aren't alone here: AFAIK the linux kernel has the same problem. Some
documents in Documentation/ doesn't make a design. There never was
never a design and nobody wants to document the current code.

> > As a whole it is the job of the maintainer to think about design. But
> > this works by accepting or rejecting patches. And by posting something
> > like "it would be nice to have xyz. I would be thankful if somebody
> > can work on this. Nevertheless I will put it on my todo list at place
> > 312." So we know what shortcomings freeciv have.
> 
> and it still would be, except with a documented architecture and highly 
> extensible system, it would hopefully be much easier to tick off the "we 
> need..." entries in the todo list.
> 
> > So what is design? You see a problem, set a goal and thing about
> > solutions. You examine several solutions in your mind and than test
> > one or implement the best one. I think each freeciv developer does
> > this. You may not notice this because you only see the final patch. I
> > agree that it may be bad the author isn't forced to write his
> > motivations/goal down.
> 
> that's all i'm asking for.

What? That the author should write down/document his
motivations/goals?

> except complete rationalization for the server framework.

??

> >  1) it is very hard to find people which work on the same part of
> >  freeciv (the people working on gotohand are an exception here)

> that's the same for corporate development too. i've always gone to great 
> lengths to avoid working on the same stuff as people on my team. it's a pain 
> to sync changes to the same code.

I was speaking about the opposite! To get a design discussion some
people have to at least know the part they want to discuss.

> >  2) you don't see the other people face to face, e.g. the conversation
> >  isn't very interactiv. This is a problem. I would really like to show
> >  some of ideas I have on a board and get a discussion going. I would
> >  really like to have a monthly freeciv meeting to design/set the
> >  direction but this is unlikely to happen.
> 
> i work from my apartment 300 miles away from my company so i don't really see 
> my team members face to face that often. when i do, there usually isn't too 
> much design discussion. it's usually, can't get this to work, can't get that 
> to work. ugh. a monthly meeting would be nice.

IMHO it needs several people to get a good design. It looks to me that
the design is finished and you/your team members just code now.

        Raimar

-- 
 email: rf13@xxxxxxxxxxxxxxxxx
  "Windows is the one true OS. MS invented the GUI. MS invented 
   the 32 bit OS. MS is open and standard. MS loves you. We have 
   always been at war with Oceana."


[Prev in Thread] Current Thread [Next in Thread]