The (G)Zdoom Question

General discussion of the port and Doom-related chat.
User avatar
Mobius
Topic author
Forum Regular
Offline
Posts: 515
Joined: Wed Jun 13, 2012 3:43 am
Contact:

The (G)Zdoom Question

Post#1 by Mobius » Sat Apr 08, 2017 4:47 pm

Hello! It's Mobius again with another interesting topic I want people to specifically pay attention to, and I want specific attention from those from Zandronum's developers.

It's come to my attention that Zandronum 3.0 will not be realized until later this year if at all, and much talk circulated on the future of the forums about this port's future. A noticeable activity stagnation chokes Zandronum with player populations only occupying specific niche servers with choice wads and speculation revolves around reasons behind this event, but honestly this is an inevitable outcome of all gaming communities and in particular the universe. We're hitting a "Heat-Death" scenario where activity is sprawled far too thin to make energy out of the forces accumulating together. Day in and out -- I see, the staff of the forums being an unnecessary function because they hardly play and hardly anything happens in the forums itself. This isn't an indication of their competence. This is indicative of the port's need, and the population that aggregated has dispersed to people's favorite mods.

This is alarming.

The main culprit to blame is a lack of Zandronum 3.0. A new update will revitalize Zandronum for a projected number of years before we're back to this very situation, and unfortunately Zandronum will play catch-up with Zdoom until literal heat-death; however, can blame be squarely placed on Zandronum 3.0 or Torr Samaho? No. The problem Torr faces isn't a lack of resources or time from players, developers, or tester but rather the system in which he's been deadlocked into since Skulltag. Zandronum is in perpetual inadequacies with Zdoom with GZdoom and QZdoom over the horizon. This port will forever NEVER be up to date and each year it will worsen as the burden becomes heavier for our main developer. Zandronum will never catch up with Zdoom. The popular ports of yesteryear will get their momentary shine of popularity with multiplayer to be out of date or overshadowed by the latest mod Zandronum WILL NEVER PORT. This isn't fair to Torr Samaho who feverishly slaves over Carnevil's mess doing what can go give us this port. It's time we finally answer the Zandronum question, or rather.. it's time we answer the Zdoom question.

It's time Zandronum finally ports itself to (G)Zdoom.

Zandronum is nothing more than an obsolete Zdoom with a superior, far superior, multiplayer feature. It's time Zandronum finally comes back home, and gives (G)Zdoom that very aspect to make it whole. This will lighten the load on Torr where he can now work with the Torr of (G)Zdoom so they can actually have equal footing in development. One man laboring, toiling, over Doom will eventually burn that person out but imagine if we had 2 of them? 3? They'll accomplish everything they seek out to do and their productivity will increase ten fold. Torr can handle the multiplayer aspect while Graf can do everything else, but he won't cause Torr will probably finish the netcode pretty quickly and they'll both bang out the same issues. Zandronum is a hindrance. It's time we finally admit that perhaps the future isn't separation but unity. What's going to happen when Torr quits or dies? At least in the chances of such an event we'll have Zandronum in another port. It's time we finally move on.
<+Thomas13> Mobius u r inferior, go outside and get beaten up

"Fool you couldn't roll on me if you were a condom" ~ Godlike


The trouble with loyalty to a cause is that the cause will always betray you

User avatar
Sean
Zandronum Tester
Offline
Posts: 614
Joined: Thu Jan 16, 2014 9:09 pm

Re: The (G)Zdoom Question

Post#2 by Sean » Sat Apr 08, 2017 5:08 pm

All I can say is good luck with that. This idea has no doubt come up before, and it hasn't happened yet. What makes you think it'd happen now?
Shame on us
Doomed from the start
May God have mercy on our dirty little hearts
Shame on us
For all we've done
And all we ever were
Just zeros and ones

User avatar
Bloax
Forum Regular
Offline
Posts: 348
Joined: Mon Jun 04, 2012 5:11 pm

Re: The (G)Zdoom Question

Post#3 by Bloax » Sat Apr 08, 2017 5:08 pm

This isn't 2007 where there's a whole bunch of competing ports to a huge audience, the audience has shrunk, and it would only make sense to do a merge of effort as opposed to a merge of ports.
Maybe it is time to say Goodbye Skulltag, who knows.

Wasted efforts are so painful though, uff.

User avatar
Mobius
Topic author
Forum Regular
Offline
Posts: 515
Joined: Wed Jun 13, 2012 3:43 am
Contact:

Re: The (G)Zdoom Question

Post#4 by Mobius » Sat Apr 08, 2017 5:19 pm

Sean wrote:All I can say is good luck with that. This idea has no doubt come up before, and it hasn't happened yet. What makes you think it'd happen now?


I do not believe this was ever publicly announced in such a proposal before, but thank you for the luck.
<+Thomas13> Mobius u r inferior, go outside and get beaten up

"Fool you couldn't roll on me if you were a condom" ~ Godlike


The trouble with loyalty to a cause is that the cause will always betray you

User avatar
Torr Samaho
Lead Developer
Offline
Posts: 1390
Joined: Fri May 25, 2012 6:03 pm

Re: The (G)Zdoom Question

Post#5 by Torr Samaho » Sat Apr 08, 2017 5:31 pm

Mobius wrote:It's come to my attention that Zandronum 3.0 will not be realized until later this year if at all

Most of the known problems with 3.0 are resolved, so I don't think that it will take that much longer to finalize it.

Mobius wrote:It's time Zandronum finally ports itself to (G)Zdoom.

Technically, I already did this a while ago. I kept a version of Zandronum fully in sync with GZDoom in the zdoom-sync branch. I just paused this since ZScript introduced some conceptual problems that need to be ironed out before it makes sense to sync further and I concentrate on finishing on 3.0 at the moment.

Mobius wrote:Zandronum is nothing more than an obsolete Zdoom with a superior, far superior, multiplayer feature. It's time Zandronum finally comes back home, and gives (G)Zdoom that very aspect to make it whole.

While I'd be very happy if we'd merge into a single port, I don't think that this will happen for at least two reasons:
  • I'm sure that Graf has no interest in having (G)ZDoom's p2p multiplayer replaced with Zandronum's c/s approach. The design of GZDoom is not c/s friendly, so our c/s handling creates quite a bit of maintenance overhead and is more like an alien element in the code base than a natural extension of our base port.
  • Zandronum and GZDoom have very different priorities. For Zandronum, stability is extremely important. People are not willing to switch back and fourth between different versions to play their favorite mods. All serves and all clients need to run the same version and we can't continuously release new versions without scaring off the users. GZDoom prioritizes adding features and doesn't worry about stability too much. If something is broken, they just make a quick fix and tell people to use the current repo version instead of the last stable version.

User avatar
Mobius
Topic author
Forum Regular
Offline
Posts: 515
Joined: Wed Jun 13, 2012 3:43 am
Contact:

Re: The (G)Zdoom Question

Post#6 by Mobius » Sat Apr 08, 2017 5:45 pm

Torr Samaho wrote:Technically, I already did this a while ago. I kept a version of Zandronum fully in sync with GZDoom in the zdoom-sync branch. I just paused this since ZScript introduced some conceptual problems that need to be ironed out before it makes sense to sync further and I concentrate on finishing on 3.0 at the moment.


Which means that the eventually of bringing them together is closer, but I want to address this in the rest of the post below. I just want to clarify that my propsal was to retire Zandronum completely in favor of an entire community migration. I should have been a bit more precise.

Torr Samaho wrote:I'm sure that Graf has no interest in having (G)ZDoom's p2p multiplayer replaced with Zandronum's c/s approach. The design of GZDoom is not c/s friendly, so our c/s handling creates quite a bit of maintenance overhead and is more like an alien element in the code base than a natural extension of our base port.


Is there a way to convince him otherwise? I recall an argument involving the netcodes of Zdoom with Zandronum and though the former was poorly constructed it proved that Zandronum's stability supersedes what Zdoom had to offer. How difficult would porting c/s be to (G)Zdoom's infrastructure especially for the incentive of broadening the port's overall player base.

Torr Samaho wrote:Zandronum and GZDoom have very different priorities. For Zandronum, stability is extremely important. People are not willing to switch back and fourth between different versions to play their favorite mods. All serves and all clients need to run the same version and we can't continuously release new versions without scaring off the users. GZDoom prioritizes adding features and doesn't worry about stability too much. If something is broken, they just make a quick fix and tell people to use the current repo version instead of the last stable version.[/list]


I suppose priorities are based around the aims of the ports with (G)Zdoom being offline they don't have to worry too much about compatibility.
<+Thomas13> Mobius u r inferior, go outside and get beaten up

"Fool you couldn't roll on me if you were a condom" ~ Godlike


The trouble with loyalty to a cause is that the cause will always betray you

User avatar
Torr Samaho
Lead Developer
Offline
Posts: 1390
Joined: Fri May 25, 2012 6:03 pm

Re: The (G)Zdoom Question

Post#7 by Torr Samaho » Sat Apr 08, 2017 6:07 pm

Mobius wrote:
Torr Samaho wrote:I'm sure that Graf has no interest in having (G)ZDoom's p2p multiplayer replaced with Zandronum's c/s approach. The design of GZDoom is not c/s friendly, so our c/s handling creates quite a bit of maintenance overhead and is more like an alien element in the code base than a natural extension of our base port.

Is there a way to convince him otherwise? I recall an argument involving the netcodes of Zdoom with Zandronum and though the former was poorly constructed it proved that Zandronum's stability supersedes what Zdoom had to offer. How difficult would porting c/s be to (G)Zdoom's infrastructure especially for the incentive of broadening the port's overall player base.

It is not a matter of porting. I fully ported over Zandronum to GZDoom in the zdoom-sync branch and kept it in sync till about a few months ago. The problem is that the underlying design of our c/s approach still goes back to Skulltag (perhaps even to CSDoom, I never checked this) and does not fit nicely with the design of the ZDoom engine from a programmers perspective. I'm just speculating on Graf's opinion, but I'd say this design is simply unacceptable to him. I'd guess you'd have to redesign both parts of the ZDoom engine and our c/s approach to have Graf consider integrating c/s into GZDoom. I'd be interested in thinking into this direction, but unless there is also interest from GZDoom, this is not going to happen.

Mobius wrote:I suppose priorities are based around the aims of the ports with (G)Zdoom being offline they don't have to worry too much about compatibility.

Absolutely. These priorities naturally come with the intended purpose of the port.

User avatar
Combinebobnt
SNS Team
Offline
Posts: 1604
Joined: Mon Jun 04, 2012 3:37 am
Contact:

Re: The (G)Zdoom Question

Post#8 by Combinebobnt » Sat Apr 08, 2017 6:32 pm

mobius I think you'd need to talk to the big boy himself graf zahl for your idea. good luck and god's speed (turbosphere?)

User avatar
Fused
Forum Staff
Offline
Posts: 484
Joined: Sat Nov 09, 2013 9:47 am
Contact:

Re: The (G)Zdoom Question

Post#9 by Fused » Sat Apr 08, 2017 7:07 pm

I think this is the dream of many around here.
To be honest, when 3.0 is stable and released I have no problem with waiting a very long time in order to take this approach at all.
#zamapping
00:21:50 <@TheMisterCat> my name is
00:21:50 <@TheMisterCat> who
00:21:53 <@TheMisterCat> my name is
00:21:54 <@TheMisterCat> what
00:21:56 <@TheMisterCat> my name is
00:21:59 <@TheMisterCat> wiki wiki
00:22:03 <@TheMisterCat> read the fucking wiki


ZCC || ZH anti-RQ patch || MultiHud
Zombie Horde: the Extinction || DoomZ : Overpoch
Image

User avatar
HTG
Forum Staff
Offline
Posts: 282
Joined: Sat Jun 02, 2012 11:53 pm

Re: The (G)Zdoom Question

Post#10 by HTG » Sat Apr 08, 2017 11:15 pm

As i said earlier in irc, if we could actually come to an agreement for this to happen it would be great. Looks like we need to work something out with graf zahl however.

User avatar
Zakken
Forum Regular
Offline
Posts: 502
Joined: Sat Jun 02, 2012 2:32 am

Re: The (G)Zdoom Question

Post#11 by Zakken » Sun Apr 09, 2017 6:55 am

I've always been an advocate for port unison over division, so I'm hoping that a compromise can be reached. Things can only stay as they are for so long.
* Determination.

User avatar
Catastrophe
ZanStuff Reviewer
Offline
Posts: 2196
Joined: Sat Jun 02, 2012 2:44 am

Re: The (G)Zdoom Question

Post#12 by Catastrophe » Sun Apr 09, 2017 9:29 am

I feel the whole "download from the latest repo" scenario can be avoided if Zandronum had an automatic updater similar to GZDoomBuilder so its less of a hassle for users.

User avatar
ARGENTVM
FNF Team
Offline
Posts: 349
Joined: Thu Jun 13, 2013 4:08 am

Re: The (G)Zdoom Question

Post#13 by ARGENTVM » Sun Apr 09, 2017 5:26 pm

Fused wrote:To be honest, when 3.0 is stable and released I have no problem with waiting a very long time in order to take this approach at all.


I don't think time is on our side if the player population stagnates to a point of no return; the low player population has been evident through a recent trend in FNF events as events that previously would get at least a 8 or more players for at least 90-120 minutes fail to bring in even a few (too bad that server stats site is down; I'd love to see a graph of the long-term statistics again). I do not think that it helps that there is not really much of a community, given all of the groups of people in Zandronum who stick to one thing and one thing only without any association with other mods, servers, or even the forums. A relative lack of new content is also evident as few mods, or even quality DM map sets for the matter, are produced for Zandronum or any other Multiplayer port. It's not really surprising in the long run, but does not point in a good direction given all of the subdivisions that already exist.

Sure, new releases stir interest, especially when a lot of ZDooM/GZDooM mods become compatible, but if few people are around by the time that happens and are aware of its release, then any resurgence will be nil when compared to previous ones.
[01:05]<+capodecima> konar eat those cheat chinese soups with noodles and drink 2 cents beers you cant except a lot from him
[01:03]<capodecima> i dont say any more word without my loyer jenova

Projects: Skirmish One-on-One | DBAB Remastered |
Zandronum IRC: #Sterling | #AoE | #MSPD | #A3 |

User avatar
Rachael
New User
Offline
Posts: 9
Joined: Mon Apr 10, 2017 12:06 am

Re: The (G)Zdoom Question

Post#14 by Rachael » Mon Apr 10, 2017 12:15 am

Yay my first post on Zandy forums.

So I figured I should give somewhat of an official response to this.

Torr Samaho wrote:
  • I'm sure that Graf has no interest in having (G)ZDoom's p2p multiplayer replaced with Zandronum's c/s approach. The design of GZDoom is not c/s friendly, so our c/s handling creates quite a bit of maintenance overhead and is more like an alien element in the code base than a natural extension of our base port.
  • Zandronum and GZDoom have very different priorities. For Zandronum, stability is extremely important. People are not willing to switch back and fourth between different versions to play their favorite mods. All serves and all clients need to run the same version and we can't continuously release new versions without scaring off the users. GZDoom prioritizes adding features and doesn't worry about stability too much. If something is broken, they just make a quick fix and tell people to use the current repo version instead of the last stable version.


On point A) here - Graf actually does want to have some sort of C/s networking in GZDoom, just not Skulltag's. More to the point - he just wants something that's not as invasive - he wants a more passive network code. GZDoom's old deterministic model is currently holding back ZScript development, and has accepted the idea that either I or someone else create a new multiplayer code so that it can move on. This isn't 1993 anymore, and that kind of networking model doesn't work for the kinds of things that GZDoom is capable of. And besides, use any CVar that is not sync'd and it's game over.

On point B) - GZDoom does worry about stability - but it worries about it a lot more during the last month before its release. Take a gander at this topic - GZDoom releases do tend to be quite stable. This is a point where I actually have to be a little bit more critical of Zandronum - 3.0 has been in the pipe for how long now? The biggest thing holding it back is simply the fact that it's been beta for so long. Sometimes you just have to let things go wild and let point releases clean up the mess. If everything mobius posted here is to be believed, then that would have solved the stagnation problem a long time ago. (I know you'll disagree with me Torr, but I just had to put that out there)

--

EDIT: By the way, if you want my personal stance on this whole issue, I have no problem with GZDoom and Zandronum merging. But merging the communities together would be a problem. It's not that either community is bad, per se, every community has its bad apples (we've had a few at ZDoom), but the goals of each community is so different. There is some overlap, and for sure we can handle that, but there is also some separation as well, which definitely would not go quite as gracefully. I think some of our rules may be more strict, and likewise might be the case for you guys too.

So from a technical standpoint, yes I think this really is a good idea (if done correctly). From a managerial standpoint, it would be better in the long run if we tried to encourage everyone from both communities to register on both sites so that they get used to each other and so that any possible transition could happen gradually.

User avatar
ibm5155
Addicted to Zandronum
Offline
Posts: 1437
Joined: Tue Jun 05, 2012 9:32 pm
Contact:

Re: The (G)Zdoom Question

Post#15 by ibm5155 » Mon Apr 10, 2017 3:54 pm

I belive the problem isn't the port itself but more about what players are playing.
I'm just going to get an example of myself, I use to open doomseker each day, and if the only servers with more than 3 players that I can find are only megaman and complex doom, I just quit doomseker and go do other thing.

That happens with many mods, people tend to join a server if they like that mod And if there's a good amount of players playing it (ie, joining a server with 1 or 2 players in who dun it is no fun, but 4 or more is nice).

I don't see new ports as a problem, since everything that's in the latest gzdoom can be done in zandronum 2.0 (well, maybe 99% of his features), the new code features aren't making new things being true, but in most cases making the code to be much less complex and smaller (I'm gonna take the example of my monster ravishing, I ported it from a code that required the latest gzdoom in 2014 to zandronum 1.0, while the gzdoom code was nice and 100% decorate, to make the same code in zan I was forced to do alot more decorate lines of code AND acs scripts for doing some stuff that couldn't be done at the moment in decorate from zan 1.0 [the code ended in ehm, 200 lines of decorate and 200 lines of acs).

Zscript for now isn't a killing feature, the documentation is not even compared to what we have for now and also many of his features are in beta stage, I belive for now there may be ehm, 10 guys that knows zscript, and ehm from that number hmm 3 or 5 that knows well (but in the future it'll be a better than decorate).

And my last point of view: There aren't many new mods that are as innovative as before, in a 2 years period we have like new mods called: Jump Maze, zdoomwars, Ghouls vs Humans, All Out War, Who Dun It, Armageddon, LOS,... And those were the days where we had the highest number of players playing online, but from 2015 - 2017 what new mods we had there were unique? I can only remember prop hunt and maybe doomz, the rest is just trying to maintain alive with updates those old mods from the skulltag days and others are just dead.

Resume: what's killing zandronum isn't the code, but yes the new unique mods that aren't coming anymore (zandronum 3.0 and 4.0 may help with that, but not if people doesn't make mods for it)
Projects
Cursed Maze: DONE, V2.0
Zombie Horde - ZM09 map update: [3/15/13]
Need help with English? Then you've come to the right place!

<this post is proof of "Decline">

User avatar
Rachael
New User
Offline
Posts: 9
Joined: Mon Apr 10, 2017 12:06 am

Re: The (G)Zdoom Question

Post#16 by Rachael » Mon Apr 10, 2017 4:03 pm

A common complaint I see on the ZDoom forums is modders using features that are not currently available in Zandronum 2.x. In fact, one of the common jokes is "if you want multiplayer support, then wait about 5 years for Zandronum to pick up the features this mod uses." So I disagree there, ibm5155.

Don't get me wrong, I get where Torr is coming from concerning stability, but there's an extreme to extreme and I think Torr is too far to the end of one of those extremes while vastly overstating GZDoom's tendency to the other. The fact remains true for GZDoom's dev builds, but in my experience Zandronum's betas aren't always a whole lot better anyway - and for both, they really need to be taken as just that - betas.

If GZDoom were really that unstable no one would make mods for it - period.

User avatar
fr blood
Frequent Poster Miles card holder
Offline
Posts: 897
Joined: Wed Mar 06, 2013 4:04 pm

Re: The (G)Zdoom Question

Post#17 by fr blood » Mon Apr 10, 2017 4:51 pm

Ibm why someone would want to make a mod for on an out-dated port(here Zandronum) when he can have a much more lagger amount of possibilities with another one(GZdoom)?
The only reason would be of course to make it playable online but then here comes another problem that would for sure make him leave, the famous "only online bug" which can be very frustrating.

So in the end the decision here is quit simple.

User avatar
ibm5155
Addicted to Zandronum
Offline
Posts: 1437
Joined: Tue Jun 05, 2012 9:32 pm
Contact:

Re: The (G)Zdoom Question

Post#18 by ibm5155 » Mon Apr 10, 2017 5:39 pm

I feel both ports are too different to be mixed, and both can (almost) do the same thing in the end...

It's not a simple problem, the decorate may be outdated, but with what we have we can do almost everything that could be done on the last zdoom 2.8, plus, we have clientside scripts, where you can do things that are impossible to do in zdoom' multiplayer (plus 64 players support).

It doesn't hurt to have both ports in your machine, so if you want to make a single player mod, you can work just fine with gzdoom, but if you plan to make a mod for multiplayer, so you better use zandronum (even so, If I remember with the zscript, zdoom isn't working....

EDIT: If Torr gets the actual build from today and release it as an official release, what's going to happen?
Projects
Cursed Maze: DONE, V2.0
Zombie Horde - ZM09 map update: [3/15/13]
Need help with English? Then you've come to the right place!

<this post is proof of "Decline">

User avatar
Rachael
New User
Offline
Posts: 9
Joined: Mon Apr 10, 2017 12:06 am

Re: The (G)Zdoom Question

Post#19 by Rachael » Mon Apr 10, 2017 6:57 pm

ibm5155 wrote:I feel both ports are too different to be mixed, and both can (almost) do the same thing in the end...

They really aren't. Not only are they related by ancestry, but they share quite a few features both downstream and upstream. What's different are the communities and the focus of the ports. Other than that, they are virtually identical, outside of Zandronum using older code and having a working client-server model multiplayer implementation.

ibm5155 wrote:It's not a simple problem, the decorate may be outdated, but with what we have we can do almost everything that could be done on the last zdoom 2.8, plus, we have clientside scripts, where you can do things that are impossible to do in zdoom' multiplayer (plus 64 players support).

A lot can happen in a year (or slightly more).

ibm5155 wrote:It doesn't hurt to have both ports in your machine, so if you want to make a single player mod, you can work just fine with gzdoom, but if you plan to make a mod for multiplayer, so you better use zandronum (even so, If I remember with the zscript, zdoom isn't working....

No, it doesn't hurt, but wouldn't it be nice to have a single port that can do it all? And for the reasons I mentioned previously, the deterministic netgame model of ZDoom has now started holding the port back. We can either throw hacks at the thing and hope some of the things stick without causing too much trouble - or we can do a more sensible thing and try to modernize it a bit by tracking and syncing at least the visual parts of the game simulation.

ibm5155 wrote:EDIT: If Torr gets the actual build from today and release it as an official release, what's going to happen?

The immediate stuff should be obvious, even to someone like me who never comes here. Long-term? I have no idea. What exactly are you trying to figure out by asking this, anyway?

User avatar
Torr Samaho
Lead Developer
Offline
Posts: 1390
Joined: Fri May 25, 2012 6:03 pm

Re: The (G)Zdoom Question

Post#20 by Torr Samaho » Mon Apr 10, 2017 7:23 pm

Rachael wrote:On point A) here - Graf actually does want to have some sort of C/s networking in GZDoom, just not Skulltag's. More to the point - he just wants something that's not as invasive - he wants a more passive network code. GZDoom's old deterministic model is currently holding back ZScript development, and has accepted the idea that either I or someone else create a new multiplayer code so that it can move on. This isn't 1993 anymore, and that kind of networking model doesn't work for the kinds of things that GZDoom is capable of. And besides, use any CVar that is not sync'd and it's game over.

Do you already have specific plans in this direction? I think if we'd work on this together from both sides, we could transition Zandronum's c/s approach to something much less invasive.

Rachael wrote:Don't get me wrong, I get where Torr is coming from concerning stability, but there's an extreme to extreme and I think Torr is too far to the end of one of those extremes while vastly overstating GZDoom's tendency to the other. The fact remains true for GZDoom's dev builds, but in my experience Zandronum's betas aren't always a whole lot better anyway - and for both, they really need to be taken as just that - betas.

I have to admit that Zandronum's focus is too much on the stability side (3.0 is long overdue...), but I still think that GZDoom does not put enough focus on stability for what a multiplayer focused port needs. Don't get me wrong, for GZDoom intended purposes its development model is absolutely fine, success proves it right. I wasn't trying to say that it's not. My main point was that we have different priorities, which result from the different requirements. Meeting somewhere in the middle could be possible.


Return to “Zandronum Talk”



Who is online

Users browsing this forum: No registered users and 1 guest