Drop the webclient?

Page: 1 2
ID:2388347
 
Jul 30 2018, 10:15 pm

Poll: Do you think lummox should drop the web client in favor of other projects?

Yes. 42% (15)
No. 54% (19)
I'm not sure. 2% (1)

Login to vote.

I've been chatting it up a bit with some other developers that use the engine, and they don't seem to even use the web client. The developers I knew of that DID use the webclient have left.

The web client uses up a lot of lummox's development time. This time could be used elsewhere. On the other hand, a lot of resources have been dumped into the web client - so it might be wise to keep going with it and hope things improve.

So, what do you think should happen? kill it or keep going with it?



The webclient in a nutshell. (according to me ; opinionated)

Pros
  • you can play BYOND games on any platform that supports a browser.
  • More modern styling on web client games.
  • A few nifty features that only the web client can do.

Cons
  • Performance on the webclient is worse than on the dream seeker client.
  • You have to host a server for the webclient to even work.
  • You have to learn some html, css and javascript to effectively use everything the web client offers.
  • Fewer learning resources on using the web client than the dream seeker. (fixable)
  • Trying to play a game built for DS on the WC doesn't always work without rewriting parts of the game.
  • Taking advantage of web client only features effectively locks you out of DS.
Jul 30 2018, 10:59 pm
Here's the two options that are only acceptable:

1) Drop the webclient and focus what used to be webclient resources/effort into the standalone exe.

2) Keep the webclient because it offers us multi-platform support.

I, personally, find the use of the webclient nifty but I won't be using it anytime soon. Not for any dislike of it, but lack of content on my end. I will eventually use it unless focus on the standalone exe also includes focus on a standalone html5 compile/packager so we an take BYOND games and integrate them into Android/Mac/Consoles.
Jul 31 2018, 5:54 am
A web client is a good idea.

The web client is not currently that good idea.
Jul 31 2018, 12:23 pm
It's nifty, but seems like it will never be halfway as functional as DS (considering how long it's been so far) and it doesn't have hardly any demand.


I honestly think the web client is a waste of time at this point.
Jul 31 2018, 1:05 pm
In response to Ter13
do you think it's "getting" there?

tbh, if lummox somehow managed to kill the "have to connect to a server" requirement. I'd use it.

The problem I have with it is I can't really release a singleplayer web game with it, because I'd have to deal with hosting a server 24/7 for people to be able to play it.
Jul 31 2018, 3:10 pm
In response to Orange55
Orange55 wrote:
do you think it's "getting" there?

No.
Jul 31 2018, 3:55 pm
Having to host and not having singleplayer is a major turn off and probably why not many existing games bother with it
Jul 31 2018, 4:33 pm
That would require creating the server in javascript, and I can't even imagine that being remotely possible.

Creating a client is a lot easier than creating a server in this case, and even that required an entirely new core setup to handle the "thin client".

The original idea was to let people tinker with the web-client's source to make their own thin clients in the language of their choice, and that's still very possible.

Running the interpreter within the browser would probably murder performance even more than just running the client does though, the best you'll get there is something that has to run the actual daemon process behind the scenes and join the game running on that process.

That wouldn't allow you to just load up a game from the website and run it alone though, it would require a download which kind of defeats the purpose.

Summary: The reason it requires joining a server because a browser isn't gonna be interpreting the dmb itself in any realistic capacity.
Jul 31 2018, 5:11 pm
That would require creating the server in javascript, and I can't even imagine that being remotely possible.


It's possible.
Jul 31 2018, 5:13 pm
Probably should have said "plausible" there instead =P
Jul 31 2018, 5:21 pm
In response to Nadrew
Nadrew wrote:
Probably should have said "plausible" there instead =P

It's also extremely plausible.

https://caniuse.com/#feat=filereader

A few other points:

The original idea was to let people tinker with the web-client's source to make their own thin clients in the language of their choice, and that's still very possible.

The web client's code is obfuscated and minified. It is not feasible to tinker with it.

Running the interpreter within the browser would probably murder performance even more than just running the client does though,

The interpreter would be lighter than rendering.

That wouldn't allow you to just load up a game from the website and run it alone though, it would require a download which kind of defeats the purpose.

No it wouldn't. The DMB and RSC could be requested directly from BYOND's domain. This is what browsers do. They fetch required resources and interpret them into a visual context. It'd be no different than your browser grabbing a .swf, only it wouldn't require the flash plugin because the DMB could be loaded and interpreted completely with the FileReader API.

The question is not technical feasibility. Only manpower.
Jul 31 2018, 5:23 pm
While we wait the months or years that would take for Lummox to do, lets get a coffee or something.

(That's what I meant by plausible, Lummox has absolutely zero interest in writing an interpreter for the browser last I asked)
Jul 31 2018, 5:44 pm
It would be more feasible to ditch compiled DM in the first place in favor of a DM->JS transcompiler.

The speed benefits of such an endeavor alone for Dream Seeker would justify the change.

Either way, not happening.

All this because swapping to OpenGL and SDL was too hard.
Aug 1 2018, 12:57 am
In response to Ter13
Ter13 wrote:
All this because swapping to OpenGL and SDL was too hard.

Aug 2 2018, 3:10 am (Edited on Aug 2 2018, 3:21 am)
yes, the web-client as it is unfortunately has no future. it's a dead project that was great in theory, but as long as it runs like hot garbage (and shows no promise of that changing) then it's a waste of resources

severed world timeline in a nutshell on BYOND:
2014 wow, this cool new feature seems like it has a lot of promise. let's make an actual MMORPG in BYOND
later that year: h-haha, hey tom, lummox, are you sure this will perform as well as DM. 'yes, it will'
2015: this is hell
early 2016: ok abandon ship

I'd call this a positive experience because you absolutely do not want to make a multiplayer game of any significant scale on BYOND, webclient or not, and we learned that the hard way. The networking code is best suited for single-player games and text-based ones--as soon as you're multiplayer then there's a noticeable performance hit. To keep up with the times a JS transcompiler definitely sounds like an interesting project.
Aug 2 2018, 7:52 am
I still think the web client has great potential.

Since JavaScript can emulate entire operating systems, it should be possible to run the entire BYOND suite on the web. Of course, that's far easier said than done. A good starting point is Emscripten.

All of the relevant browsers support asm.js and WebAssembly, and anything that runs in those tends to be orders of magnitude faster than it would be if ported normally, so that's definitely something to keep an eye on.
Aug 2 2018, 8:30 am
Focus effort on making programming easier. Make byond a beginner friendly language like these. https://www.slant.co/topics/4801/~2d-non-coding-game-engine
Aug 2 2018, 8:33 am
In response to Cuart
Cuart wrote:
Focus effort on making programming easier. Make byond a beginner friendly language like these. https://www.slant.co/topics/4801/~2d-non-coding-game-engine

No.

There's nothing wrong with BYOND having a programming language instead of a click'n'stick IDE. The problem is a good majority of BYOND developers were BYOND game players for a long period of time before they just jumped into programming without any knowledge or learning, i.e. source ripping. So they expect themselves to fiddle with already-botchy code someone else did and they figure, "Oh, I'm a coder now."
Aug 2 2018, 8:55 am
In response to Cuart
Maximus_Alex2003 wrote:
...

what this guy said.

Aug 2 2018, 11:35 am
I'd rather see BYOND on ios then on a web client. I dream of a day I can make an app or console game with BYOND.

How feasible that is, beats me, probably pretty darn difficult. I'd rather Lummox just focus on what Lummox thinks needs working on, whether that be the webclient, performance, new features, whatever. I like the direction BYOND has taken since Lummox took over and respect whatever choices he makes on where it needs to go in the future.
Page: 1 2