Sign in with
Sign up | Sign in
Your question

[Crawl] Interesting development on the Yahoo group Crawl-Dev

Last response: in Video Games
Share
Anonymous
July 11, 2005 9:34:21 PM

Archived from groups: rec.games.roguelike.misc (More info?)

Hi all,

Over on the Yahoo group crawl-dev (http://groups.yahoo.com/crawl-dev/)
there's a Crawl fan who's a developer who has lost patience with the
delays and lack of a clear timeline on a new release and has announced
intent to fork the project based on the last official release, that is,
ordinary b26. His vision is first a bugfix release, and only then any
eventual feature changes/additions. He says he only has a few hours a
week for the project, and is looking for collaborators.

The post in question is about two days old.

I haven't yet commented over there, but I really should. As surprising
as it may sound in light of my many squabbles with Brent, I'm mildly
against it, as it seems there will be a lot of knowledge and changes
gone to waste if Brents' efforts (based on his statements on the subject
over the months) since b26 are thrown out.

For those who don't want to take the trouble to sign up to Yahoo Groups
(I did it so long ago I don't remember how much hassle it is and how
much blood I had to sign my name in to sign up), here's the text of his
post:

---

--- In crawl-dev@yahoogroups.com, "d.brodale" <yahoo@b...> wrote:
> What exactly is the reason why development should be closeted off?

It shouldn't. I agree.

Therefore, since the Crawl license permits it, I'm going to fork the
beta 26 codebase into a new project.

The project's name is to be determined, as are its scope and features.
My instinct is to declare that the first release will focus entirely on
bugfixes from beta 26, new feature work only to be decided on later. The
project will be in source control and read-only access will
be available to the world, read/write access will be available to
trusted developers. The bar to becoming a 'trusted developer' will be
pretty low.

If you are interested in participating, please send mail to "crawl -at-
tgr _dot_ com" and indicate your interest level. At the present time,
please ONLY volunteer if you are a software developer. If you just have
a lot of cool ideas for the next version, then I'd say
keep it in this Yahoo group instead. In your mail, please specifically
mention the following:

(1) whether you know how to use some source control system (cvs, svn,
perforce)
(2) to what extent you've looked at the Crawl source code, and to what
extent.
(3) what fixes, if any, have you made to crawl beta 26 that haven't yet
hit any release (feel free to attach context diffs if you like)
(4) what platforms you primarily intend to develop for.
(5) How much time a month you can devote to the project.

Let me be perfectly frank: I _don't_ have a metric buttload of time to
devote to an open source project. But I am willing to get this set up
and rolling and spend a few hours a week integrating patches, fixing
bugs, and rolling releases, at least until someone else can
take over. The current state of Dungeon Crawl (utterly moribund) isn't
really good enough to keep it alive.

-peterb

---

Erik
July 11, 2005 10:56:05 PM

Archived from groups: rec.games.roguelike.misc (More info?)

Erik Piper wrote:
> Hi all,
>
> Over on the Yahoo group crawl-dev (http://groups.yahoo.com/crawl-dev/)

I think this should be http://groups.yahoo.com/group/crawl-dev.

> there's a Crawl fan who's a developer who has lost patience with the
> delays and lack of a clear timeline on a new release and has announced
> intent to fork the project based on the last official release, that is,
> ordinary b26. His vision is first a bugfix release, and only then any
> eventual feature changes/additions. He says he only has a few hours a
> week for the project, and is looking for collaborators.
>
> The post in question is about two days old.
>
> I haven't yet commented over there, but I really should. As surprising
> as it may sound in light of my many squabbles with Brent, I'm mildly
> against it, as it seems there will be a lot of knowledge and changes
> gone to waste if Brents' efforts (based on his statements on the subject
> over the months) since b26 are thrown out.
>
> For those who don't want to take the trouble to sign up to Yahoo Groups
> (I did it so long ago I don't remember how much hassle it is and how
> much blood I had to sign my name in to sign up), here's the text of his
> post:

Turns out people don't need to sign in to Yahoo groups to read the
discussion, anyway.

It's a very interesting thread IMO and well worth the read.

A.
Anonymous
July 12, 2005 2:59:01 PM

Archived from groups: rec.games.roguelike.misc (More info?)

In <dau3ht$9io$1@domitilla.aioe.org> Erik Piper wrote:

> I haven't yet commented over there, but I really should. As surprising
> as it may sound in light of my many squabbles with Brent, I'm mildly
> against it, as it seems there will be a lot of knowledge and changes
> gone to waste if Brents' efforts (based on his statements on the
> subject over the months) since b26 are thrown out.

I somewhat agree, but OTOH seeing some kind of development is preferably
to vague promises that a new release will see the light of day sometime,
although unknown when.

I haven't looked at the Crawl source (I don't know if I dare, when
people speak of the source here it sounds like Lovecraftian Secrets Man
Was Not Meant To Know; YANI - why no Cthulhu-inspired branch?), but
fixing the most obvious bugs (steam resistance, mutation rate, and some
other things - I suspect you keep track of those better than I do) and
release a new version shouldn't be too much work. Preferably integrating
Darschan's patch.

Personally, I'd recommend a rewrite to make Crawl easier to extend and
tweak. As it is now it seems Crawl is beginning to go against limits in
the underlying architechture, where small changes ripple through the
whole code. Probably a very large undertaking, but it might be worth it
in the long run in better maintainability and expandability.

--
Johan
Anonymous
July 12, 2005 5:59:36 PM

Archived from groups: rec.games.roguelike.misc (More info?)

Johan Strandell wrote:
> In <dau3ht$9io$1@domitilla.aioe.org> Erik Piper wrote:
>
>
>>I haven't yet commented over there, but I really should. As surprising
>>as it may sound in light of my many squabbles with Brent, I'm mildly
>>against it, as it seems there will be a lot of knowledge and changes
>>gone to waste if Brents' efforts (based on his statements on the
>>subject over the months) since b26 are thrown out.
>
> I somewhat agree, but OTOH seeing some kind of development is preferably
> to vague promises that a new release will see the light of day sometime,
> although unknown when.

I do understand; that's why I said "mildly". :-)

Incidentally, Brent has in the meantime replied to the post containing
the call for forking, but his reply is a bit of a non-sequitur, i.e. no
statement on the call for forking itself, which is a bit surprising
considering the seriousness of such an event if it occurs. Also, no
delivery dates mentioned, as usual. ;-)

> I haven't looked at the Crawl source (I don't know if I dare, when
> people speak of the source here it sounds like Lovecraftian Secrets Man
> Was Not Meant To Know; YANI - why no Cthulhu-inspired branch?), but
> fixing the most obvious bugs (steam resistance, mutation rate, and some
> other things - I suspect you keep track of those better than I do) and
> release a new version shouldn't be too much work. Preferably integrating
> Darschan's patch.

Certain parts of the code are quite readable even by a dummy like me,
and certain errors and their most obvious fixes or the gist of those
fixes are also clear (e.g. all it takes to fix the steam resistance bug
is to remove a 'not' sign that shouldn't be there); some parts of the
code are OTOH far beyond my ken, but that doesn't necessarily mean
they're incomprehensible for someone competent; like I said, I'm a dummy.

As far as I can tell, Crawl's code's reputation as being unreadable
throughout is based partially on the past state of the code, which was
worse than its present state.

> Personally, I'd recommend a rewrite to make Crawl easier to extend and
> tweak. As it is now it seems Crawl is beginning to go against limits in
> the underlying architechture, where small changes ripple through the
> whole code. Probably a very large undertaking, but it might be worth it
> in the long run in better maintainability and expandability.

There was an attempt to rewrite Crawl from the ground up fully in C++
and presumably fully as modern, normal, decently readable code, with the
moniker Crawl 5, a while back. It withered and died due to the project
head's inactivity. Brent's efforts seem to be in part somewhat in the
same vein, but without the language shift. It's clear he's doing a lot;
the problem seems to be that he's doing too much, or rather, he's not
willing to publicly present buggy code at all, eliminating the
possibility for collaboration.

e.

--
bang your head to your favorite song
it's very mechanical
- Satomi, Deerhoof
July 12, 2005 6:10:57 PM

Archived from groups: rec.games.roguelike.misc (More info?)

Erik Piper wrote:

> Over on the Yahoo group crawl-dev (http://groups.yahoo.com/crawl-dev/)
> there's a Crawl fan who's a developer who has lost patience with the
> delays and lack of a clear timeline on a new release and has announced
> intent to fork the project based on the last official release, that is,
> ordinary b26. His vision is first a bugfix release, and only then any
> eventual feature changes/additions. He says he only has a few hours a
> week for the project, and is looking for collaborators.
>
> The post in question is about two days old.
>
> I haven't yet commented over there, but I really should. As surprising
> as it may sound in light of my many squabbles with Brent, I'm mildly
> against it, as it seems there will be a lot of knowledge and changes
> gone to waste if Brents' efforts (based on his statements on the subject
> over the months) since b26 are thrown out.

I agree with you about promising Brent. The real problem with Brent is that
he apparently lives in parallel time universe, where 2 years since the last
release is like two weeks (ratio 50:1 and still counting). I say: let's
give Crawl to the people, before all of us will die of old age or Sims 2
overdose.

--
Greetings,
Loonie
---------------------------------
De profondis clamo ad te, Domine.
www.crawl.webpark.pl
Anonymous
July 13, 2005 2:24:55 AM

Archived from groups: rec.games.roguelike.misc (More info?)

In <db0bbb$d0g$1@domitilla.aioe.org> Erik Piper wrote:
> Johan Strandell wrote:

> As far as I can tell, Crawl's code's reputation as being unreadable
> throughout is based partially on the past state of the code, which was
> worse than its present state.

That's what everybody says. One of these days I should really take a
look. (Actually, the real reason I haven't looked at it yet was because
I couldn't find an OS X expander for the source file.)

>> Personally, I'd recommend a rewrite to make Crawl easier to extend
>> and tweak. As it is now it seems Crawl is beginning to go against
>> limits in the underlying architechture, where small changes ripple
>> through the whole code. Probably a very large undertaking, but it
>> might be worth it in the long run in better maintainability and
>> expandability.
>
> There was an attempt to rewrite Crawl from the ground up fully in C++
> and presumably fully as modern, normal, decently readable code, with
> the moniker Crawl 5, a while back. It withered and died due to the
> project head's inactivity. Brent's efforts seem to be in part
> somewhat in the same vein, but without the language shift. It's clear
> he's doing a lot; the problem seems to be that he's doing too much,
> or rather, he's not willing to publicly present buggy code at all,
> eliminating the possibility for collaboration.

The crawl source seems to have a will of its own: it resists being made
into something saner! :) 

A project of this size seems to need more than one person taking
responsibility for it. If I was in Peter's shoes I'd try to get Brent on
board, but go on with the fork anyway if he wasn't willing. Fixes in
gameplay (balance, game mechanisms) shouldn't be impossible to port, and
IMO that's the important parts. No one plays a game because the game
engine is well-written; it's the actual gameplay that matters.

I've toyed with the idea of porting Crawl to Python, but to be truthful
it's not something I'll do in the foreseeable future.

--
Johan
Anonymous
July 13, 2005 3:14:48 PM

Archived from groups: rec.games.roguelike.misc (More info?)

Erik Piper <erikNOSPAM@sky.cz> writes:

> There was an attempt to rewrite Crawl from the ground up fully in
> C++ and presumably fully as modern, normal, decently readable code,
> with the moniker Crawl 5, a while back. It withered and died due to
> the project head's inactivity.

It was always likely to wither and die. Rewriting Crawl (or any major
roguelike) from the ground-up is far from easy, so much so that it's
safe to assume that any such effort is doomed before it begins.

--
Darshan Shaligram <scintilla@gmail.com> Deus vult
Anonymous
July 15, 2005 5:30:21 PM

Archived from groups: rec.games.roguelike.misc (More info?)

In <1121416107.521983.49990@g43g2000cwa.googlegroups.com> great.
throwdini wrote:
>> That's what everybody says. One of these days I should really
>> take a look. (Actually, the real reason I haven't looked at it yet
>> was because I couldn't find an OS X expander for the source file.)
>
> Um... it's just a bzip2'd tarball. Although I now notice that my
> install
> of OS X associates the .tbz2 extension with Microsoft Excel for some
> strange reason.

Stuffit Expander didn't recognize the extension, and me neither.

> Possibly rename the .tbz2 file extension to .tar.bz2 (or just use the
> ..zip source archive instead) and it should expand on a double-click --
> or in the worst case, drop into Terminal.app and extract the archive
> on the commandline.

Will try that. Thanks a lot!

(And in line with your comment re: Linux installation - someone should
fix the Crawl OS X distribution. I had to hunt down a curses package
separately to get it to run.)

--
Johan
!