The future of Nethack

[bhaak]

When developers pick an existing project and start to "improve" it it usually doesn't become better as a game. Often the opposite happens and the original feel of the game is lost. These people have good intentions, but it doesn't mean they are great or even good game designers.

And the DevTeam consists of good game designers? Hardly believable if you look at the design issues that NetHack has and had for the longest time.

NetHack has grown quite organically. It wasn't handed down to the DevTeam by a higher entity. Many popular features have come not from the DevTeam itself but have been suggested by players or have even been programmed directly by patch writers.

The worst option is a group of open source -minded developers who lack a single, focused idea of a good gameplay and just throw in some stuff.

But Slash'Em was never meant to have a single, focused idea of gameplay. Slash'Em was the embodiment of having added any patches that the Slash'Em DevTeam could get a hold of!

[Krice]

And the DevTeam consists of good game designers?

I don't know about that, but at least they got it right most of the time. When random internet people start to improve someone else's project it usually doesn't become a better game, often just worse. When you make the game from the start it has much better chance to become a good game, because it requires more than regular tweakers have.

[miki151]

Isn't that exactly what happened when Hack turned into Nethack? By the way, a developer of one of the main Nethack variants once spoke highly of the Nethack source code. He said it's well written by the standards of 80's and 90's. Of course now we use different conventions, and OO design is used everywhere. But I guess it's a common misconception that the Nethack code is "crap".

[Krice]

But I guess it's a common misconception that the Nethack code is "crap".

I would say better that it's difficult. Same as original Crawl's source code. But the source code is not the main issue here. It can be improved by almost anyone.

[Kevin Granade]

It's really easy to look at a huge chunk of code and declare it broken or terrible, I had the same knee-jerk reaction when I started on Cataclysm, but once I got to know the code a bit I realized it was quite well designed and implemented within certain constraints.  It suffers (working on it, but it definitely is still suffering) from a lack of OO and advanced data structures where appropriate, for example manually encoding flags as bitfields rather than using a super efficient stl set container.  But the enum-based bitfield implementation was quite well done and robust for what it was.

As for joining together teams, it's really hard to comment on project direction from the outside, while it's possible that they simply haven't thought of it, or that some reason they were separate no longer exists, it seems more likely that there's a reason that the projects are heading in their own directions.

[transcendreamer]

hello!
is there any info when UN v5 is going in release?
thanks!

[bhaak]

is there any info when UN v5 is going in release?

Really soon now. This time for real. Last time in May when I actually intended to release the next version, there were some unresolved issues and missing tiles and then some more commit happened by new developers and the ideal time for a release was missed.

[bhaak]

Isn't that exactly what happened when Hack turned into Nethack?

Yes, Hack didn't get any updates for 2 years (Ha! I just say looking at NetHack's last release date) and then Mike Stephenson made a usenet posting about combining bug fixes and patches and collaboratively release a new version.

https://groups.google.com/forum/#!topic/rec.games.hack/3qVaxNFkcVc

By the way, a developer of one of the main Nethack variants once spoke highly of the Nethack source code. He said it's well written by the standards of 80's and 90's. Of course now we use different conventions, and OO design is used everywhere. But I guess it's a common misconception that the Nethack code is "crap".

If it's been a Slash'Em dev, then I wouldn't value that opinion too much. Slash'Em is buggy as hell.

But I agree that it isn't that bad. Most people use NetHack's use of the old K&R method declarations as a sign that it's bad code but this isn't actually a problem.

NetHack is a complex program, probably more than half of its game engine is dedicated to special cases, it has grown organically over the years and for that, it's not that bad in most places.

[bhaak]

And the DevTeam consists of good game designers?

I don't know about that, but at least they got it right most of the time. When random internet people start to improve someone else's project it usually doesn't become a better game, often just worse. When you make the game from the start it has much better chance to become a good game, because it requires more than regular tweakers have.

This is really funny as the NetHack DevTeam were random internet people who took someone else's project and didn't make it from the start.

You probably never played NetHack 1.3d, the first version that the NetHack DevTeam released. I and some other people did for the 2011 Junethack tournament and it's actually not a really good game. If there wouldn't have been constant development, with input from many players, we wouldn't remember NetHack as something special but only as a fork of Hack done by some random tweakers.

I don't have a problem if you don't like NetHack forks. But idolizing the DevTeam does not help improve those forks (or hoping that the DevTeam will release a new version) but clearly name what do you think is wrong with them. This doesn't mean that every criticism will be taken into account. Rec.games.hack and rec.games.roguelike.nethack has thousands of such postings. Some had an impact on development and some didn't.  But this is how development works if you do it in public.

[transcendreamer]

thanks a lot
looking forward for v5
UN is my fav fork

[transcendreamer]

is it possible that sometimes UN will present player a way to see weapon stats (like in ADOM or Angband) my be in indirect way?
for ex. - having found different polearms (bardiche/halberd/...) it's hard to decide which one to use as their stats are hidden from user

like warning message when trying occasionally to bash a moster with bow - there could be a message like: "you feel this weapon is less effective against this monster"

or this would be against nethack core conceptions?

[bhaak]

is it possible that sometimes UN will present player a way to see weapon stats (like in ADOM or Angband) my be in indirect way?
for ex. - having found different polearms (bardiche/halberd/...) it's hard to decide which one to use as their stats are hidden from user

There's been the idea of an info screen for items for a while that would show information like this. I don't think this goes against NetHack's principle and even if it did. You can't nowadays ignore that people just go to nethackwiki and look stuff like this up. Having it ingame would be much more convenient. Although this info screen has not yet been coded.

[transcendreamer]

oh yes, such screen would be very handy
maybe combined with indentify screen ("slash" command) or something like this

you're pushing nethack forward to best roguelike project

[miki151]

By the way, a developer of one of the main Nethack variants once spoke highly of the Nethack source code. He said it's well written by the standards of 80's and 90's. Of course now we use different conventions, and OO design is used everywhere. But I guess it's a common misconception that the Nethack code is "crap".

If it's been a Slash'Em dev, then I wouldn't value that opinion too much. Slash'Em is buggy as hell.

I think it was the Sporkhack dev.

After developing my own roguelike, I actually started to admire Nethack for being so stable despite having so many special behaviors. You could play version 3.4.3 for years and not find a single bug. I think that it tells a lot about the code.

[bhaak]

After developing my own roguelike, I actually started to admire Nethack for being so stable despite having so many special behaviors. You could play version 3.4.3 for years and not find a single bug. I think that it tells a lot about the code.

If you don't find any single bug while playing 3.4.3 for years, it might tell more about your style of playing than about the code 

But seriously, to this day the admin of NAO has to regularly recover games because of crashes and you have to remember that the version on NAO has fixed all known crash bugs that are in 3.4.3. Granted, most of those crashes happen if you take the game to the limit, for example doing excessive pudding farming. I myself found several crash bugs by ramping up the monster generation and looking how stable is the game when creating a monster might fail (not very, but it's easily fixed).

There are lots of bugs in 3.4.3 (no wonder, it's a big game) and you can easily make the game crash. For example when saving a game while riding and wielding Sting, the game crashes while loading the game or - one of my favorites - just throw a potion at iron bars.

The official bug list on nethack.org lists about 500 bugs, on nethackwiki there are some more. Is that much for a program that was released 10 years ago?

You could say no because most of those bugs you can workaround or avoid by not doing some stupid ascension tricks (although NetHack offers itself for doing such thing and it's part of its appeal) and there are very few bugs that can hit you without you doing anything (for example a monster with a potion of acid on the drawbridge falling into the water because another monster zapped a wand of striking that collapsed the drawbridge might lead to a crash).

OTOH you can say that as the last real changes to the core code was in 3.3.0 which means 1999 that this is quite a lot. The release notes for 3.3.1, 3.4.0, and 3.4.1, talked about hundreds of bug fixes (3.4.2 and 3.4.3 "only" about dozens). But of course the real problem here is not that there are bugs but that there are no bugfix releases anymore.

You shouldn't look at your roguelike and feel bad that it has bugs. If you had hundreds or thousands of volunteers that tried to find bugs and because of the seclusive nature of the DevTeam we can't really say how many of those volunteers also wrote patches for fixing bugs. If you had that, your roguelike would also be much more stable. 

We often overlook the fact that NetHack didn't get big and popular just because of the DevTeam but also because there was a great community that helped the DevTeam tremendously.