[freeciv-data] Re: State of the data patches

[Raimar Falke <rf13@xxxxxxxxxxxxxxxxx>]

On Tue, Jan 21, 2003 at 03:45:34AM -0500, Eric S. Raymond wrote:
> 
> Raimar Falke <rf13@xxxxxxxxxxxxxxxxx>:
> > > There's a lesson here.  Tying the data patch to the GUI stuff complicated
> > > my life way too much.  The data files are designed so they can be enriched
> > > without breaking the server's startup parse; we should bear that in mind 
> > > and decouple the two.
> > 
> > That way we add information in the data files which is never tested
> > and used.
> 
> But that's not true.  Your patch tests and uses the class/legend data,
> for example.  I expect to develop new fields (if any) in conjunction with
> game-logic programmers like you.

Good.

> >         If a new kind of data is added we needed the network code
> > and the GTK1 client code. GTK2 and Xaw would also be nice to have. You
> > provided the new data, I wrote the code and now I want people opinions
> > on the GUI code.
> 
> I like the look of the tabs.  Very nice, exactly as I would have
> chosen.  I also like the way the legend displays in a box beneath the
> flags.  

Good.

> But there is some minor problem; tabs other than "All" are blank.
> This is with your patch version 3.  Once that's fixed, I think your
> patch will be ready to be merged.

I didn't finish the patch because I wanted feedback. So far I got your
and Vasco's response.

> > > Here's how I would prefer to handle it, if Raimer doesn't object:
> > > 
> > >1. Raimer, please drop the data-file stuff out of your 
> > >nation-classification
> > > patch entirely so we won't step on each other.  I think it would be better
> > > if I maintained the class/legend stuff.  I wrote it, after all.
> > 
> > No problem.
> 
> Thanks.
> 
> > > 2. I will take the responsibility to do a rollup patch that will bring 
> > > the data files fully up to date, *with* class/legend, and *with* Barsoom.
> > 
> > See other mail about the Barsoom. Also merging these two are bad. I
> > want to have two separate commits for them.
> 
> You mean separate commits for the other updates and Barsoom?  OK,
> if that's how you want it, I'll leave Barsoom out of the rollup patch.

I will also strip the other changes from the class patch. It this ok?

> > > 3. From now on, the GUI maintainers get to play catchup with the data
> > > files at their own pace.
> > 
> > This won't work. Also will you be happy if it turns out that nobody
> > has produced the network and GUI code in 6 months?
> 
> But it has already worked.  *You* have a patch that uses the
> class/nation data, right now.  And it's not like I'm going to add new
> fields at random; that would be senseless.  Before I add any new fields, 
> I'll be sure to have a clear idea how they'll be used, one I've discussed
> with game-logic developers.

What you wrote above in "3." sounded another way but the new
"clarification" is good.

> > > 4. Informally, I'm willing to 'own' the nation files.  Good role for
> > > me, as improving them is something fun that I can fit into a busy
> > > schedule, doing lots of small steps, with little risk that I'll
> > > break anything.
> > 
> > That is very good.
> 
> I thought you might like it.  They show signs of not being maintained
> well, and have not been edited as a unit.  For example, most use
> native-language forms of city names (which I think is the right
> thing), but a few use Anglicized ones. This should be fixed.  And there
> are all kinds of little historical nits; one I spotted tonight is that
> Baghdad is listed as a Persian city when it should be an Arab one (it
> was built by the Caliph al-Mansur, 921-1001).

Yes almost nobody is interested in this. People are more interested in
sexy things like adding maps with topologies which are hard to display
in 2D ;)

        Raimar

-- 
 email: rf13@xxxxxxxxxxxxxxxxx
 "SIGDANGER - The System is likely to crash soon"