Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Developmental Disillusionment

2 views
Skip to first unread message

Gamer_2k4

unread,
Jun 21, 2006, 11:23:46 AM6/21/06
to
I've been working on my roguelike on and off for about five years now,
and I've realized that I haven't made much progress in the past two
years. I am always thinking of new ideas that the current engine won't
accomodate, so I am constantly rewriting my code. I wrote three
versions in QBasic, two in Java, and at least five in C++ (in that
order). These aren't slight modifications. This is taking my core
code (input and output basically, along with a few algorithms) and
rewriting the display, interface, content, etc.

For some time now I've had a great storyline and am prepared to create
the game and three sequels, but I can never figure out which version I
want which feature implemented in. Due to the nature of the story
(which is excellent), the first game must be a dungeonhack, the next
will have a static overworld, the third will have several static
overworlds, and the last will have a large completely random world.

I've designed cultures and races for my game, have thought of detailed
quests and skills, and designed several (and I think original) weapon
and armor styles. My issue right now is determining what ideas go in
which version and when. IMHO, some of my ideas are really great, so I
want to implement them immediately, but for story and gameplay
purposes, they'd be better left out until later versions.

The end result of all this confusion is that I never get anything
accomplished. I have a great engine for one game, then I change my
expectations and requirements. Also, I play too much of one game, and
I like its style, so I remodify my RPG to be somewhat like it. Then I
find another one I like, and its rewriting time once more.

I guess my question is this: has anyone else experienced problems of
this magnitude? Did you beat it? Do I need to spend more time
preplanning, or less time creating? Is there a "best way" to handle
this?

Gamer_2k4

Krice

unread,
Jun 21, 2006, 12:10:57 PM6/21/06
to
Gamer_2k4 wrote:
> I guess my question is this: has anyone else experienced problems of
> this magnitude?

Yes, in a way. I began with C and changed to partial C++ so it wasn't
that big problem for me in the language level. At least some
source code pieces or ideas have survived in way or another to this
day.
I never re-programmed my project from the scratch, but I don't know
if that was good or bad.

> Did you beat it?

Not yet.

> Do I need to spend more time preplanning, or less time creating?
> Is there a "best way" to handle this?

Plan and implement something which is going to stay. Stick to
your plans, don't change them if you don't need to.
The same with source code, you can always make it better,
but you never get anything ready that way.
You need to make the decisions about what features you want
in your game. Better make those decisions sooner than later.
Don't fool yourself to think about sequels when you don't have
the first game ready.

Slash

unread,
Jun 21, 2006, 12:17:55 PM6/21/06
to

Gamer_2k4 wrote:
> I've been working on my roguelike on and off for about five years now,

SNIP

> I guess my question is this: has anyone else experienced problems of
> this magnitude?

Yes

> Did you beat it?

I guess so

> Do I need to spend more time
> preplanning, or less time creating?
> Is there a "best way" to handle
> this?

I think there are two ways:
* Devote your life to a vision of the game,dont let it slip or get
corrupted by your own experiences with other games, keep your initial
spec, get enough time and energy to work on it.

* Make a 7DRL, get a list of improvements and get working ;)

>
> Gamer_2k4

Radomir 'The Sheep' Dopieralski

unread,
Jun 21, 2006, 12:18:23 PM6/21/06
to
At 21 Jun 2006 08:23:46 -0700,
Gamer_2k4 wrote:

> I guess my question is this: has anyone else experienced problems of
> this magnitude? Did you beat it? Do I need to spend more time
> preplanning, or less time creating? Is there a "best way" to handle
> this?

All the time :)
7drl seems to be a little helpful.
Just sitting down and writing it for once (and not rewritting anything
until you have a playable game) works too.

--
Radomir `The Sheep' Dopieralski

Elsairon

unread,
Jun 21, 2006, 1:15:26 PM6/21/06
to
"Gamer_2k4" <game...@gmail.com> wrote in message
news:1150903426.6...@y41g2000cwy.googlegroups.com...

> I've been working on my roguelike on and off for about five years now,
> and I've realized that I haven't made much progress in the past two
> years. I am always thinking of new ideas that the current engine won't
> accomodate, so I am constantly rewriting my code.

One suggestion I have which has helped me alot is to make
the code you keep re-doing into libraries or modules and
develop a common interface/messaging system which you
will use for all your program. If you make it generic enough
to handle all cases and be extensible, you can plug and
play only 'sub-sections' of your code, and not need to
re-write everything all the time.

I have a core library of functions I optimzed and have used
for a couple years.

Another option is to design an modifiable UI and display
engine. That way if you want your game to look different,
just use whichever of the info files you create which
define your 'RL Interface of the month.'

> For some time now I've had a great storyline and am prepared to create
> the game and three sequels, but I can never figure out which version I
> want which feature implemented in

You know which order the games will be produced in. I would
suggest getting a stable, playable version of your main game
engine done with all the 'Basic' features that you want to
support in all your future versions. And to do this with your
first 'dungeonhack' style game. Once you have completed it
you will have solved all the basic design issues and can move
on to handling the more extensive features with a solid engine.
already in place.

[snip]

> I want to implement them immediately, but for story and gameplay
> purposes, they'd be better left out until later versions.

Sounds like you already made your decision. Knowing that you
will have future versions is more incentive to get the first
one done. You can always add more on to the next project!

> The end result of all this confusion is that I never get anything
> accomplished.

[snip]

> I guess my question is this: has anyone else experienced problems of
> this magnitude?

yep

> Did you beat it?

I think so. I used to spend all my time doing research
on other RL's, RPG systems, case studies of injuries from
medical reports, temperature and climate variation levels,
animal species, mythology, on and on, collecting all the
data to make the ultimate game... and did no code. So I
had a huge archive of almost anything I need for my game. I
wrote up a long design document.. and realized it would
take me maybe 20 years to code the damn thing.

I had to start from ground zero... make @ move on a map.

The way I broke free of the trap was to start keeping track
of what I actually do. I now have four major ongoing
lists/journals that I keep up to date.

1. A list of what I want in my game (huge.. of maybe 100
things I have about 5 or 6 implemented)
2. A change log. Any noteworthy or significant change or
improvement in my codebase I keep track of.
3. General journal for research, other ideas, notes, etc.
4. An error list. Everytime I make a mistake or error in
my code I keep an ongoing list. This lets me learn
much faster from any errors, and also shows me where
I can make the most improvements.

I get new ideas alot, and now I just quickly write them
down and add them to my list. I usually get ideas alot
faster than I can get them working, and thats fine. Using
this system I get to see what I actually do, and have
a record of it which has saved my game more than once.

> Do I need to spend more time
> preplanning, or less time creating? Is there a "best way" to handle
> this?
>
> Gamer_2k4

If you designed your game engine from the ground up with
a core rule that you needed to be able to expand and
implement features that you did not yet know, what would
be different? If you designed your system in a way that
was scalable and extensible adding features would be easy.

I have not yet managed to do that and I still end up
having to re-write alot of mechanics because I simply
did not plan for new ideas I had. How could I?

I split my code into discrete sections.. user interface,
keyhandling, display fuctions, map functions, ai,
monster functions, inventory, etc. I can pull any out
and plug in another with not too much re-writing.
That way I can keep the gains I have made in other
areas without having to start building the city of
Rome one more time.

Ultimately I think its best to spend more time coding
and testing. That way you can actually have an RL someday.
--
Elsairon


Crypt

unread,
Jun 21, 2006, 4:38:01 PM6/21/06
to
> I guess my question is this: has anyone else experienced problems of
> this magnitude?

For rpg tabletop rule designs , yes, for about 15 years.

For rogue like development, no. I have a fairly precise idea of what i want. I
regularly make deep modifications which are code optimizations, not complete
rewrites.


> Did you beat it? Do I need to spend more time


> preplanning, or less time creating? Is there a "best way" to handle
> this?


Have a clear idea of what you want and remember that wanting everything will
lead nowhere.

djabanete

unread,
Jun 21, 2006, 4:42:16 PM6/21/06
to
On 2006-06-21 17:23:46, "Gamer_2k4" <game...@gmail.com> wrote:

[snip]

these are three steps to helping yourself solve your problems.

A) go for a long walk.

B) drink ice water in a dark room and listen to good music.

C) go to a party and tell everyone that you are an amazingly talented roguelike
game designer.

after those three steps, barriers in your life will be easier to overcome.
really.

good luck

Gamer_2k4

unread,
Jun 21, 2006, 5:44:10 PM6/21/06
to
> these are three steps to helping yourself solve your problems.

> A) go for a long walk.

Yes, that's a great way to get more ideas. Some of my best
inspirations came while walking or biking to or from school.

> B) drink ice water in a dark room and listen to good music.

The good music part works.

> C) go to a party and tell everyone that you are an amazingly talented roguelike
> game designer.

I've tried it before...people aren't as impressed by that as they
should be. "Where's the graphics?" "THAT'S your game?" "What am I
seeing here?" "So what you're saying is that you suck..." "Nerd."
"Geek." "*I* wouldn't play this..." "Do you even HAVE friends?" and so
on and so on...roguelikes aren't popular with normal people. That's
unfortunate, too, because they're missing out on some great times
(alone) (staring at a computer).

> after those three steps, barriers in your life will be easier to overcome.
> really.

> good luck

Thank you. =)

Corremn

unread,
Jun 21, 2006, 8:02:19 PM6/21/06
to

Gamer_2k4 wrote:
> > these are three steps to helping yourself solve your problems.
>
> > A) go for a long walk.
>
> Yes, that's a great way to get more ideas. Some of my best
> inspirations came while walking or biking to or from school.

Hmm, I get too many ideas when I spend time alone.

> > B) drink ice water in a dark room and listen to good music.
>
> The good music part works.

What ever works :-)

> > C) go to a party and tell everyone that you are an amazingly talented roguelike
> > game designer.
>
> I've tried it before...people aren't as impressed by that as they
> should be. "Where's the graphics?" "THAT'S your game?" "What am I
> seeing here?" "So what you're saying is that you suck..." "Nerd."
> "Geek." "*I* wouldn't play this..." "Do you even HAVE friends?" and so
> on and so on...roguelikes aren't popular with normal people. That's
> unfortunate, too, because they're missing out on some great times
> (alone) (staring at a computer).

My friend once anounced to a large party that I write rogue-like
computer games, after the silence finally died down, I just got very
drunk. That helped.

Antoine

unread,
Jun 21, 2006, 8:28:43 PM6/21/06
to

Gamer_2k4 wrote:
> > C) go to a party and tell everyone that you are an amazingly talented roguelike
> > game designer.
>
> I've tried it before...people aren't as impressed by that as they
> should be. "Where's the graphics?" "THAT'S your game?" "What am I
> seeing here?" "So what you're saying is that you suck..." "Nerd."
> "Geek." "*I* wouldn't play this..." "Do you even HAVE friends?" and so
> on and so on...roguelikes aren't popular with normal people.

In my opinion there's no such thing as 'normal people'.

Gamer_2k4, what do *you* think you should do about your Developmental
Disillusionment? Putting aside the frustrations of the moment, what
seems to be the best way forward?

A.

Gerry Quinn

unread,
Jun 22, 2006, 8:13:15 AM6/22/06
to
In article <1150903426.6...@y41g2000cwy.googlegroups.com>,
game...@gmail.com says...

> I've been working on my roguelike on and off for about five years now,
> and I've realized that I haven't made much progress in the past two
> years. I am always thinking of new ideas that the current engine won't
> accomodate, so I am constantly rewriting my code. I wrote three
> versions in QBasic, two in Java, and at least five in C++ (in that
> order). These aren't slight modifications. This is taking my core
> code (input and output basically, along with a few algorithms) and
> rewriting the display, interface, content, etc.
>
> For some time now I've had a great storyline and am prepared to create
> the game and three sequels, but I can never figure out which version I
> want which feature implemented in. Due to the nature of the story
> (which is excellent), the first game must be a dungeonhack, the next
> will have a static overworld, the third will have several static
> overworlds, and the last will have a large completely random world.

I think roguelikes tend to lead one into this trap. All the world can
plausibly be in a roguelike; it is hard to think of an idea and say
"that would just needlessly complicate the game". If you were just
doing a simple arcade game, you would be less likely to keep dreaming
up new stuff.

That said, you seem to have a particularly bad case! I think you just
need to write a game, any game. Call it something different if you
feel it will conflict with your dream for the main project. Am I right
in thinking you are always thinking about elaborate AIs, monster
societies etc.? Unless your hero is a monster anthropologist, they
don't matter to him! Just include ones that want to kill him.

I'm progressing slowly on a roguelike myself for the last month or so.
I've found that with roguelikes you just can't progress without cutting
things out and making decisions, even if those will limit some features
in the long term. Whatever language you use, your brain won't be
capable of dealing with the infinitely flexible roguelike that is an
ideal.

I'm hoping to get a working adventure with a few levels in it sometime
month. I'm not even sure it will have items in it - I am deferring
consideration of them in favour of getting basic combat mechanics and
interface right. (Since my long-term aim is a mouse-controlled
isometric roguelike, interface is very important.)

Normally I wouldn't release an unfinished or partial game - but I think
with roguelikes it may be worthwhile. So my suggestion is to consider
something like this; a mini-roguelike with a few levels, that should be
playable and entertaining.

- Gerry Quinn


Sherm Pendley

unread,
Jun 22, 2006, 11:01:57 AM6/22/06
to
"Gamer_2k4" <game...@gmail.com> writes:

> I guess my question is this: has anyone else experienced problems of
> this magnitude?

Yes.

> Did you beat it?

For some value of "beat", yes. I changed my goals - instead of trying to
ship a finished game, I now write a simple roguelike as an exercise when
learning a new language or library.

I suppose that's more of a dodge than a victory, but it works for me.

sherm--

--
Cocoa programming in Perl: http://camelbones.sourceforge.net
Hire me! My resume: http://www.dot-app.org

Gamer_2k4

unread,
Jun 22, 2006, 11:40:13 AM6/22/06
to
> > Did you beat it?

> For some value of "beat", yes. I changed my goals - instead of trying to
> ship a finished game, I now write a simple roguelike as an exercise when
> learning a new language or library.

> I suppose that's more of a dodge than a victory, but it works for me.

I see. That approach will not work for me, because I'm trying to write
a roguelike that many people will enjoy and play. I'm not writing a
roguelike for fun and I'm not writing one to learn programming
(although that is happening). I'm doing it because I want to share my
ideas and creations with the roguelike community. In short, I'm trying
to be famous (egad - got that one off my chest). I suppose that's why
I'm trying to add in all these features; I want my game to be good.
Luckily, I'll be going on my vacation next week. Hopefully I can
quietly contemplate my game design and content, and hopefully I'll
emerge with a plan, ready to do some coding.

Gamer_2k4

Krice

unread,
Jun 22, 2006, 5:30:25 PM6/22/06
to
Gamer_2k4 wrote:
> I suppose that's why
> I'm trying to add in all these features

Don't do that yet. It's better to start with basic features and keep
the
amount of features low. When you get the basic features work
as they should then it's time to add more features. This is good
because doing everything in smallest possible scale is making
the whole project easier to maintain. It doesn't mean you have to
cut down the features from the final game.
I did the same mistake as you: I planned several features and lots
of items and monsters, but couldn't get them to work reliably,
because everything changed all the time and those changes were
reflected all around the source code. That is the reason why it's
very important to decide each feature and try to build a game
where those features are compatible with each other. It's a lot
harder thing than you would imagine - and it's one of the reasons
why roguelike games are so difficult to make.

Elsairon

unread,
Jun 22, 2006, 6:00:17 PM6/22/06
to

"Krice" <pau...@mbnet.fi> wrote in message
news:1151011825.8...@r2g2000cwb.googlegroups.com...

> Gamer_2k4 wrote:
> > I suppose that's why
> > I'm trying to add in all these features
> Don't do that yet. It's better to start with basic features

[snip]

> I did the same mistake as you: I planned several features and lots
> of items and monsters, but couldn't get them to work reliably,
> because everything changed all the time and those changes were
> reflected all around the source code.

[snip]

I think this is the main obstacle that I have encountered.
After adding features I have had to to refactor almost every
time. Keeping all the features localized as much as possible
helps more than anything else I have found.

However some features can't be implemented without major
changes/additions to the rest of the source. I have not
yet found away around this except deciding not to add
any new features that is incompatible and thats not a
very fun option for me.
--
Elsairon


Balaam

unread,
Jun 22, 2006, 6:41:33 PM6/22/06
to
Have a search around for "Data driven design", it's the kind of thing
that can help with a often changing design.

Antoine

unread,
Jun 22, 2006, 7:36:46 PM6/22/06
to

It's very true. Sometimes you just find that you have designed your
game in such a way that it's very difficult to add the new feature you
want. You then have three options:
A - abandon the new feature
B - abandon the entire project
C - bite the bullet and spend the time to change the code.

A is very often the best option, painful as it is.

The tricky thing is to learn to design your game, NOT such that it will
be able to incorporate any new feature you will think of later (this is
too hard of a goal), but so that when you make a design decision that
will rule out options later, you are aware that you are doing so and
what features you will miss out on.

In Guild, for example, I did realise that my system of 'smart maps'
would later rule out any features involving tunneling or passing
through walls. That was fine; I was happy with that and willing to live
with it. On the other hand, I hadn't realised that my AI and interface
design prevented friendly monsters from ever really being able to work
properly. That was a problem.

A.

Elsairon

unread,
Jun 23, 2006, 10:44:34 AM6/23/06
to
"Balaam" <bal...@gmail.com> wrote in message
news:1151016093.9...@p79g2000cwp.googlegroups.com...

> Have a search around for "Data driven design", it's the kind of thing
> that can help with a often changing design.

I will, thanks. I am test tweaking ai now, combat and
pathfinding. Once I have a handful of basic interesting
ai, I plan to write a simple parser and swap out the
hardcoded values I have now for info-files. I have a
basic idea of what data driven design is, and I'm sure
I can learn a lot more.
--
Elsairon


Elsairon

unread,
Jun 23, 2006, 10:58:25 AM6/23/06
to
"Antoine" <ma...@guildgame.com> wrote in message
news:1151019406.6...@g10g2000cwb.googlegroups.com...

> Elsairon wrote:
> > "Krice" <pau...@mbnet.fi> wrote in message
> > news:1151011825.8...@r2g2000cwb.googlegroups.com...
> > > Gamer_2k4 wrote:
> > > > I suppose that's why
> > > > I'm trying to add in all these features
> > > Don't do that yet. It's better to start with basic features

[snip]

> > However some features can't be implemented without major


> > changes/additions to the rest of the source. I have not
> > yet found away around this except deciding not to add
> > any new features that is incompatible and thats not a
> > very fun option for me.
>
> It's very true. Sometimes you just find that you have designed your
> game in such a way that it's very difficult to add the new feature you
> want. You then have three options:
> A - abandon the new feature
> B - abandon the entire project
> C - bite the bullet and spend the time to change the code.
>
> A is very often the best option, painful as it is.
>
> The tricky thing is to learn to design your game, NOT such that it will
> be able to incorporate any new feature you will think of later (this is
> too hard of a goal), but so that when you make a design decision that
> will rule out options later, you are aware that you are doing so and
> what features you will miss out on.

Yes this is the challange. To think ahead and see what
the consequences of implementing feature a will be on
potential features x,y or z.

> In Guild, for example, I did realise that my system of 'smart maps'
> would later rule out any features involving tunneling or passing
> through walls. That was fine; I was happy with that and willing to live
> with it. On the other hand, I hadn't realised that my AI and interface
> design prevented friendly monsters from ever really being able to work
> properly. That was a problem.
>
> A.

I bet that took a lot of rework. So far I have not had
to do any major changes except with my time/speed system.
My first attempt was so bad I just abandoned that branch
of code and started over. I learned alot from that though.

Now I keep copies of the last working build for each
significant feature I implement. If anything happens and
I need to scrap something, I can just start with clean
code and not have to comb through all my source and hunt
for all the changes/errors I made.
--
Elsairon


Jeff Lait

unread,
Jun 23, 2006, 11:17:52 AM6/23/06
to
Elsairon wrote:
>
> Now I keep copies of the last working build for each
> significant feature I implement. If anything happens and
> I need to scrap something, I can just start with clean
> code and not have to comb through all my source and hunt
> for all the changes/errors I made.

You should seriously consider using some form of version control. That
way you don't have to explicitly keep copies of the last good version,
as you have copies of every previous version.

http://subversion.tigris.org/
http://svnbook.red-bean.com/
--
Jeff Lait
(POWDER: http://www.zincland.com/powder)

Elsairon

unread,
Jun 23, 2006, 12:49:12 PM6/23/06
to

"Jeff Lait" <torespon...@hotmail.com> wrote in message
news:1151075872.5...@u72g2000cwu.googlegroups.com...

> Elsairon wrote:
> >
> > Now I keep copies of the last working build for each
> > significant feature I implement.
[snip]

> You should seriously consider using some form of version control.
[snip]

Thanks Jeff this looks like what I need.
--
Elsairon


Gamer_2k4

unread,
Jun 23, 2006, 12:51:20 PM6/23/06
to
> Have a search around for "Data driven design", it's the kind of thing
> that can help with a often changing design.

>From what I've read (immediately after seeing your comment),
data-driven design is where code is dynamically generated based on
static data (say, in external files). If this is the case, DDD seems
fairly complicated. Am I misinterpreting what DDD is? Even if I am,
could you post some links to helpful DDD literature? I personally
don't like how it sounds, but it could be intriguing to learn about.

Gamer_2k4

Jeff Lait

unread,
Jun 23, 2006, 2:08:55 PM6/23/06
to

There seem to be many different formal definitions of DDD on the 'net.
I don't know what the previous author meant, but I can explain my
interpretation.

Data Driven Design is an attempt to separate code from data and then
put as much of your game as possible in the "data" side of the camp.

Some posts I made about this:
http://groups.google.com/group/rec.games.roguelike.development/msg/1ef7085e662b1008
http://groups.google.com/group/rec.games.roguelike.development/msg/44bc6b22f62ae89c

Here's what someone else says which seems relevant. I'm less extreme
than they are - I would contend one *should* put game-specific hacks in
your code and only migrate them to the data when they stop being
exceptions.
http://gamearchitect.net/Articles/DataDrivenDesign.html

Antoine

unread,
Jun 23, 2006, 6:52:42 PM6/23/06
to

FWIW I use Subversion for all my professional and hobby programming
now. I don't claim that one can't do without version control, but I
certainly wouldn't choose to.

A.

Balaam

unread,
Jun 23, 2006, 8:10:54 PM6/23/06
to
Sorry, I was in a bit of rush when I wrote that but I mean that type of
thing discussed in the following types of links. (Basically what Jeff
said)

http://ai.eecs.umich.edu/soar/Classes/494/talks/Schumaker.pdf
http://gamearchitect.net/Articles/DataDrivenDesign.html
http://gamearchitect.net/Articles/GameObjectRoundtable.html
(They mention Naughty Dog, now subsumed in Sony, not using DDD - it's
only interesting because they were using their own version of Lisp to
program their games - in Lisp the difference between data and code is
much less pronounced and therefore DDD is not something that's really
needed)
http://www.drizzle.com/~scottb/gdc/game-objects.ppt

Happy reading and programming. Don't get too caught up in the perfect
design :D

Krice

unread,
Jun 24, 2006, 3:18:28 AM6/24/06
to
Jeff Lait wrote:
> I would contend one *should* put game-specific hacks in
> your code and only migrate them to the data when they stop being
> exceptions.

This is very good and important advice. Yet better is that you know
before what features are going to be data driven and what features
can be done with special cases. It's not that obvious to everyone
and I think it's one of the things that should be planned as good
as possible.
Data driven approach isn't going to save from bad planning or
bad decisions. There is a good possibility that it will require
more re-factoring than anything else.

Timofei Shatrov

unread,
Jun 24, 2006, 4:54:05 AM6/24/06
to
On 23 Jun 2006 15:52:42 -0700, "Antoine" <ma...@guildgame.com> tried to
confuse everyone with this message:

Isn't subversion, like, too bloated for hobby programming? I use Darcs
(http://darcs.net) and it's awesome. Very easy to install and get going.

--
|Don't believe this - you're not worthless ,gr---------.ru
|It's us against millions and we can't take them all... | ue il |
|But we can take them on! | @ma |
| (A Wilhelm Scream - The Rip) |______________|

Antoine

unread,
Jun 24, 2006, 5:16:53 AM6/24/06
to

Timofei Shatrov wrote:
> On 23 Jun 2006 15:52:42 -0700, "Antoine" <ma...@guildgame.com> tried to
> confuse everyone with this message:
>
> >
> >Elsairon wrote:
> >> "Jeff Lait" <torespon...@hotmail.com> wrote in message
> >> news:1151075872.5...@u72g2000cwu.googlegroups.com...
> >> > Elsairon wrote:
> >> > >
> >> > > Now I keep copies of the last working build for each
> >> > > significant feature I implement.
> >> [snip]
> >> > You should seriously consider using some form of version control.
> >> [snip]
> >> > http://subversion.tigris.org/
> >> > http://svnbook.red-bean.com/
> >> > --
> >> > Jeff Lait
> >> > (POWDER: http://www.zincland.com/powder)
> >>
> >> Thanks Jeff this looks like what I need.
> >
> >FWIW I use Subversion for all my professional and hobby programming
> >now. I don't claim that one can't do without version control, but I
> >certainly wouldn't choose to.
> >
>
> Isn't subversion, like, too bloated for hobby programming? I use Darcs
> (http://darcs.net) and it's awesome. Very easy to install and get going.

How do you mean, bloated? In terms of memory, disk space, features,
...?

I find it OK in these respects but you might have had something else in
mind...

A.

Elsairon

unread,
Jun 24, 2006, 5:59:24 AM6/24/06
to
[various snips]

"Timofei Shatrov" <gr...@mail.ru> wrote in message
news:449cfd12...@news.readfreenews.net...


> On 23 Jun 2006 15:52:42 -0700, "Antoine" <ma...@guildgame.com> tried to
> confuse everyone with this message:
> >Elsairon wrote:
> >> "Jeff Lait" <torespon...@hotmail.com> wrote in message
> >> news:1151075872.5...@u72g2000cwu.googlegroups.com...
> >> > Elsairon wrote:
> >> > >
> >> > > Now I keep copies of the last working build
> >> > >

> >> > You should seriously consider using some form of version control.
> >> >

> >> > http://subversion.tigris.org/
> >> > http://svnbook.red-bean.com/
> >> > --
> >> > Jeff Lait
> >> > (POWDER: http://www.zincland.com/powder)
> >>
> >> Thanks Jeff this looks like what I need.
> >
> >FWIW I use Subversion for all my professional and hobby programming
>

> Isn't subversion, like, too bloated for hobby programming? I use Darcs
> (http://darcs.net) and it's awesome. Very easy to install and get going.
>

I'll check it out.
--
Elsairon


Elsairon

unread,
Jun 24, 2006, 7:13:47 AM6/24/06
to
"Elsairon" <elsa...@verizon.net> wrote in message
news:028ng.1053$Gh.1018@trnddc02...

Darcs keeps only a single un-revert level and I will most
likely use 'reverting' a lot, which would lead to my
projects eventual demise so I decided to use Subversion.

I got subversion running and I can see already how
much faster and robust my source backups are, thanks again.
--
Elsairon


Gerry Quinn

unread,
Jun 25, 2006, 6:45:22 AM6/25/06
to
In article <1151107854.4...@y41g2000cwy.googlegroups.com>,
bal...@gmail.com says...

> http://gamearchitect.net/Articles/GameObjectRoundtable.html

> (They mention Naughty Dog, now subsumed in Sony, not using DDD - it's
> only interesting because they were using their own version of Lisp to
> program their games - in Lisp the difference between data and code is
> much less pronounced and therefore DDD is not something that's really
> needed)

"A developer from Ritual observed that this approach only works because
Naughty Dog has a large number of programmers relative to designers."

A typical roguelike does too (1:1 is high)!

I would comment that there's never a hard division. I have a
MonsterData class whose code mostly looks like:

case SNAKE:
{
m_Name = "Snake";
m_nMaxHits = 10;
m_Speed = 0.8;
// etc. etc.
}
break;

If I alter this to read from a spreadsheet, will my game suddenly
become more data-driven? It might IF the programmer and the designer
weren't the same person.


- Gerry Quinn

Elsairon

unread,
Jun 25, 2006, 7:40:27 AM6/25/06
to
"Gerry Quinn" <ger...@DELETETHISindigo.ie> wrote in message
news:MPG.1f08799ce...@news1.eircom.net...

Good point. Whats the difference between tweaking the class in
the code or tweaking an external file. The both probably take
the same amount of time, perhaps less with the code since
you're already working with it right then.

Unless you have already compiled your code, in which case you
gain the benefit of tweaking without re-compiling. Or maybe
tweaking while the game is running by dynamically loading.
--
Elsairon


Krice

unread,
Jun 25, 2006, 7:58:42 AM6/25/06
to
Gerry Quinn wrote:
> If I alter this to read from a spreadsheet, will my game suddenly
> become more data-driven?

The discussion about the distinction of data and code often becomes
ridiculous. The theory behind data/code is not important. Only thing
that matters is how readable the code is. If you use lots of
switch-cases like that the source code will become less readable
and harder to maintain (if no other reason then only because the
size of the source code gets bigger 1.)

1. Reference: Dungeon Crawl source code

Elsairon

unread,
Jun 25, 2006, 8:53:32 AM6/25/06
to

"Krice" <pau...@mbnet.fi> wrote in message
news:1151236722....@i40g2000cwc.googlegroups.com...

If you seperate all your intensive cases such as monsters/items/spells,
etc into different files its 'not much' different than data files.

For rapid development and prototyping I rarely use external
files unless what I am testing has to do with file-handling
in some way.

Having seperate data files is better(for me) for my final
projects, and I support whoever keeps their data in code =)
--
Elsairon


Gerry Quinn

unread,
Jun 25, 2006, 9:37:28 AM6/25/06
to
In article <1151236722....@i40g2000cwc.googlegroups.com>,
pau...@mbnet.fi says...

This is a special case, because the MonsterData.cpp file is mainly just
one big switch statement that gets called at the start of the game to
fill out a static vector of MonsterData instances in Monster. There is
one MonsterData element for each monster type.

I do it this way so that in theory I can switch to a spreadsheet or
something else to fill out the array. In practice, I probably won't.
The main thing is that the data source, as Elsairon points out, is in a
separate file from other code.

- Gerry Quinn

Gray...@gmail.com

unread,
Jun 25, 2006, 1:03:04 PM6/25/06
to
You might try using task oriented milestones.

Working with milestones keeps everything in perspective. If you can
lock down a couple of milestones worth of features, your in good shape.
If any new ideas come along, they are post those milestones.

I recommend really simple milestones.

If you can carve out a milestone map all the way to your Beta 1
release, then all new features and ideas get dumped into your V2
milestones.

This is the approach I've taken and it seems to be working. Although,
I've also been working on my Dungeon Crawler for several years and only
partially done. I didn't start using Milestones until this year though.
I think they really helped put the amount of work I have in
perspective.

My task manager said I have 18 years worth of tasks to do. :| (4 hour
work weeks don't go very far)

GrayDwarf

Elsairon

unread,
Jun 25, 2006, 2:10:44 PM6/25/06
to

<Gray...@gmail.com> wrote in message
news:1151254984.6...@m73g2000cwd.googlegroups.com...

> You might try using task oriented milestones.

[snip]

> This is the approach I've taken and it seems to be working.

[snip]

> My task manager said I have 18 years worth of tasks to do. :| (4 hour
> work weeks don't go very far)
>
> GrayDwarf

Impressive! That sounds like a very useful tool. What
software (if any) do you use to track your milestones/todo's?
--
Elsairon


Timofei Shatrov

unread,
Jun 25, 2006, 6:04:14 PM6/25/06
to
On Sat, 24 Jun 2006 11:13:47 GMT, "Elsairon" <elsa...@verizon.net>

Darcs keeps every state of your source code, but if Subversion works for
you, that's fine. I actually haven't ever used Subversion: I heard that
it's a souped up CVS, and the last time I used CVS... that was
horrible...

Gray...@gmail.com

unread,
Jun 25, 2006, 6:38:52 PM6/25/06
to
> Impressive! That sounds like a very useful tool. What
> software (if any) do you use to track your milestones/todo's?

I currently use MS Project but it's too big for my tastes. I'm working
on writing my own task manager now. I've spent 5 solid days writing out
the spec including actual UI drawings. It's pretty solid. Anyway, I've
posted a screen of one of my test projects here:
http://graydwarf.blogspot.com

You can find it under C# Game Development: Day 4

I normally don't wrap all my versions into one project though. It's
more like: M1-M5 (Beta 1) M6 (Beta 2) M7 (Release(V1) - depends on the
project size. In this example, it was all in one because the project
was so simple.

I also use CHANGE: and BUG: categories which don't really come into
play in that test project.

GrayDwarf

Elsairon

unread,
Jun 25, 2006, 7:10:33 PM6/25/06
to
"Timofei Shatrov" <gr...@mail.ru> wrote in message
news:449f084c...@news.readfreenews.net...

[massive snip]

> Darcs keeps every state of your source code, but if Subversion works for
> you, that's fine. I actually haven't ever used Subversion: I heard that
> it's a souped up CVS, and the last time I used CVS... that was
> horrible...

Yeah. This is my first time using source control system. If Subversion does
not work out I will definetly try DARCS.
--
Elsairon


Elsairon

unread,
Jun 25, 2006, 7:17:23 PM6/25/06
to
<Gray...@gmail.com> wrote in message
news:1151275131.9...@b68g2000cwa.googlegroups.com...

I checked out some free online taskmanagers.. had about the same luck
you reported. I'll bookmark to your blog. Interesting to see someone
elses methods.
--
Elsairon


Chris Morris

unread,
Jun 26, 2006, 4:44:28 AM6/26/06
to
gr...@mail.ru (Timofei Shatrov) writes:
> On Sat, 24 Jun 2006 11:13:47 GMT, "Elsairon" <elsa...@verizon.net>
> tried to confuse everyone with this message:
> >Darcs keeps only a single un-revert level and I will most
> >likely use 'reverting' a lot, which would lead to my
> >projects eventual demise so I decided to use Subversion.
>
> Darcs keeps every state of your source code, but if Subversion works for
> you, that's fine. I actually haven't ever used Subversion: I heard that
> it's a souped up CVS, and the last time I used CVS... that was
> horrible...

Same here. Another advantage (though not so much for roguelikes) of
darcs is that you don't need to set up a complicated server if you
want to share development with someone else (or just let power users
download development snapshots) - you can just email patches around or
put a repository directory on the web.

--
Chris

konijn_

unread,
Jun 28, 2006, 8:14:13 AM6/28/06
to

After reading this :
http://www.echannelline.com/usa/story.cfm?item=20891
I can only agree even more. I stick usually with CVS though.

Cheers,
T.

gf

unread,
Jun 28, 2006, 9:38:55 AM6/28/06
to
On Wed, 28 Jun 2006 05:14:13 -0700, konijn_ wrote:

> After reading this :
> http://www.echannelline.com/usa/story.cfm?item=20891
> I can only agree even more. I stick usually with CVS though.
>
> Cheers,
> T.
>

I also use Subversion (can't tolerate CVS directory handling) and it has
another advantage: commiting releases to the repository makes you organize
bug-hunting and feature-adding, by commiting only stable/consistent
snapshots.

--
Giorgos

0 new messages