The evolution of Winamp, and the possible demise of Wasabi
 

Reference this post. I'll dispense with the pleasantries. Winamps 2 and 3 will be merged, and we're trying to work out what plugin/component system will be optimal. Wasabi will not be discontinued, but Winamp won't be based on it, apart from the skinning stuff.

• Would you be prepared to switch back to a more Winamp2-style plugin system?
• If there was to be a new component system, what would you change?
• What would be most language-independent?

No flaming. Just good, clear, concise argument. Go.

IF they used the new .wac component system they have devised for the wasabi engine, and just had backwards-compatibilty for 2.xx plugins that would be the best

but, seeing how everything else is going I doubt they'll put anything in that will require any kind of change for the developers/users of 2.xx

Well, I've spoken to Justin, and everything is still to be decided at this point. Due to the way the Winamp3 component system works, I don't think it would be that hard to do.

The Winamp 2-style plugin system has the advantage that it's very easy to port to other languages, in comparison to the Wasabi SDK which is difficult due to it's sheer size. Still, I haven't looked at the new winamp 2 apis either.

Re: The evolution of Winamp, and the possible demise of Wasabi
 

• Would you be prepared to switch back to a more Winamp2-style plugin system?
  
  No.
  I very much like the Wasabi-styled component system and how Winamp3 is entirely built up by components (as in: everything is components).
  My suggestion: about the same as Pyr0x suggestion, Use the Wasabi-styled components system but merge that with the Wa2-style system = best of both..
  I can't see any reason to NOT do this..
  
• If there was to be a new component system, what would you change?
  
  I think the most important thing, is that it should be possible to manage the components (every component, including the player itself with all it's stuff) from Maki.
  Wa3 doesn't do this too good at the moment.
  One other very important thing, is to make components support desktopalpha and the ability to use them in skins like you can do in Sonique (have for example AVS as background with antialiased free-form borders).
  
• What would be most language-independent?
  
  I really can't answer that, but I believe it would be the Wa2/Wasabi hybrid (merge the plugin systems).

----

I really think that one of Wa3's component systems best advantages is that everything is a component, which pretty much means you can add or remove ANYTHING in the installation. So if you don't want a certain part of the player, jsut remove it..
Winamp2 cannot do this as everything is baked into one core.

-Plague

Well, i'd rather have a fast, and stable mp3 player with the choice to not install things like Video or CD Burning, and I don't care much for having every single thing as a component (things like the Playlist and Media Library should simply be mandatory).

I don't think that it has to be completely cross-platform, or component based just for a good mp3 player. Keep in mind, that most users want a quick, and simple experience, and most don't care to have complete control over something that will just play music. If it's going to add so much time to the development, and so much more CPU and mem usage, then I think it would not be worth the time, and that Nullsoft would be going further than they need to.

I'd love to see the color controls up and running.

To answer your questions:

1. I'd like to see support for both current component types, and to see them develop independetly.

2. I'd like it to be easy for anyone to make components. I'd like to see the functionality of MAKI extended somewhat so that more useful scripts could be made with it (plus some DECENT documentation, not just *ADD LIST*).

3. I'm not sure I understand that question.


As a user I'd like to see something extremely fast and reliable. A color behind DTA where DTA is turned off or not avalible, to minimize the appearence of jaggies. The option of multiple colums in the playlist.


As a skinner I'd like to see some documentation on each and every XML tag and attribute. I'd like to see the gamma controls accessible from maki. Maybe if you really feel like 'busting your bones' a skinning IDE.




A lot of these are more general requests than anything, but if theres going to be a new winamp I want to put my two cents in.

I disagree, personally I never use the Media library, and the playlist isn't a must, the load button can be used depending on how you organise your files.

this is a quote, but im not saying from who cos i dont wanna get them in shit if its wrong..

"if this new 'winamp4/5' goes ahead, all winamp2 and 3 plugins/components will still work"

See here, kthxbye :). I shall try and make an official, properly-informed announcement at the same time as the official one is made, detailing the exact technicalities of the thing. Till then, unsticking this.

I would like to see that the XML language packs remain surported because they are much easier to write...

[edit]
Just to clearify my statement above: There are 116 Locales for Winamp3 today, but only 97 for Winamp 2! Everybody here knows that Winamp 2 is aviable for years and Winamp 3 is' t. So the only possible conclusion is that the XML Locale is the better (simpler) System!

If desired i try to deliver a adjusted statistic that try to take multible packs of the same Language in account.
[/edit]
[edit nr="2"]
Please excuse my bad english, maybe it's the best reason to follow my request. ;)

I'm from europe so it's not realy a big problem, but what about unicode?[/edit]

Re: The evolution of Winamp, and the possible demise of Wasabi
 

Okay, let me see.

As for the first question, I think the winamp3 component system (wasabi) gave very fat and practical tools into our hands.. Spending weeks of coding to make a quite-a-good looking UI for your stuff sucks, just compare these pictures (http://www.lyricsamp.org/lyricsamp2/...-GetLyrics.jpg vs http://www.lyricsamp.org/images/Lyri...references.jpg)..

btw, yeah, there was a lyricsamp for winamp2 but it sucked, I was too young to code it well :)

back to our subject, never was as easy as creating the UI for LyricsAMP using the WASABI stuff... That's why we'll really miss it, I think... If you say, the WHOLE skinning system comes to winamp2, and that must mean all control redraw (hopefully all Win32 common controls) through the skin, then it's OK. Still, the idea of wasabi is very close to what I think about future of programming, and how complex systems should be built up.

Talking about the new component system, I really would make a winamp 5 plugin/component really independent to what winamp 5 is made of. I mean, it would be really great if people could choose language, environment and other stuff for the plugin, and that's the answer for question 3 too..
I think COM is the only way to make this work TODAY... I'd wish that nullsoft made nextgen winamp utilize all the great stuff introduced by the .NET platform, but it's still very new...
I had problems with COM, but still, it's not that bad to base on, just look at the Windows Media Player 9 Series model... It still has to be coded and documented cleanly and professionally. Wasabi wasn't. :)

That's it.

Just so it's clear:

Winamp 2.9 and it's decendent Winamp 5 (provisional name) will have an enhanced Winamp 2-style plugin system. However, Winamp 5 will also be able to load Wasabi up in a plugin, to use all Wasabi skins and (hopefully) plugins. Winamp3 will still be developed, it'll be used as the development platform for Wasabi

oh, cool..
so Wasabi will be loaded as a plugin?
Nice solution.

Since Wa3 will still be developed and Wa5 will be a "waypoint" between wa2 and wa3, I'm open for all new ideas...

-Plague

I like evil_oj's idea of installing only the components you need.

To me you can do anything to WA2 as long as it is still fast and not a resource hog.
The coming 2.9 will support old plugins too, right?

ofcourse...
They will ADD stuff, not remove anything

-Plague

the coming 2.9 is entirely 2.x, 99% wasabi free.

I personally think Wasabi is a pain in the ass, mainly because of the lack of documentation (Yes, of course there are samples, but still I need more docs than a list of classes)
Why do you think there isn't an Eonic for WA3 yet ;)

So I would very much appreciate a more winamp 2.x style plugin system.. Without the randomly occuring bugs in the host app, of course.

Just my $0.02..

You perplex me, sir. You actually prefer frontend.h over 5 megabytes of open source code?


--Brennan

i agree with brennan.. and ill add that you are an ass... 5mb of open source we you can do WHAT EVER YOU WANT, or a 2k frontend.h file that limits u to only a few certain features... i know what i choose

Let me elaborate :)

First of all let me say I didn't mean any offence towards Wasabi or you, Brennan. I'm sure it's a kick-ass system. But not for me.

You'll want to know why not I guess..
Well, basically it gives me nothing, except a lot of extra work for what I like to do. Which is writing visualizations. Programming one has become sort of a living hell for me. The first Eonic (my very first winamp plug) was slapped together in 4 days. From scratch that is. I had never ever done any winamp plugin before, or even looked into it. Conclusion: The old system was extremely easy to use. And provides everything I need/want. A few things could have been easier (like getting the title of the song currently playing, such things) but I can live with the way things work now.

When Winamp3 and the SDK first came out I was probably the most exciting person on earth to get started and do a new visualization. I downloaded all I needed, and tried to get started. I looked at all the examples but after that I still thought it was extremely complicated.
Not of course if you want to do simple stuff (basically copying an example and extending it a bit). But what if you want to do just a little bit more than that? Such as writing a plugin using Direct3D (software rendering has been done).
Not that hard, it turned out afterwards. But because there is no proper documentation available, it'd have taken me days just to figure out it actually is possible to get the window handle from a window (Note I'm bad at reading other people's sources, especially if it's 5 meg of sources :D).
Well anyway, I gave up! What I wanted just seemed impossible, or at least to complicated to accomplish. Just didn't have the time to go figure it out. A couple of days ago I got back here on the forums, deciding that I really was gonna get it to work now, and I found this OpenGL example. The programmer of that piece of code is my personal hero. There I found the getHwnd() function that I needed so badly :)

You'll say well there you go. What's so bad about it now then? Of course it got better for me, at least I can use it now. But there's still some things that are majorly bugging me compared to developing plugins for Winamp 2.x, here's some points:
First and most important: I can't link the .wac while Winamp is running. This means that for every little change I make and need to see the result (a thing that happens ALL the time when doing graphic effects), I have to close Winamp, then compile/link, then fire up Winamp again. And WA3 not being the fastest on starting up either, I find this extremely annoying.
Also, Having to compile 5 meg of code into my plugin instead of 1 simple header file. Result: plugin becomes huge. Almost 900k, filled with what I can do in 10k if it were for Winamp 2.x.
Well I can't think of more right now actually.

Anyway, these 2 things are majorly annoying me. I hope you understand this.. I really didn't get anything new that I will actually use, and on top of that, things only have gone backwards for development. I loved the old and easy way. It might not have been the best ever, but it was as good as perfect for me (a programmer of only visualization plugins). I think I'm not the only one thinking like this. Just take a look how many vis plugs there are for winamp3.. 13!! And half of those come from wildtangent so they don't count :)

I hope this explains my $0.02 above a bit..

I kinda know what you mean..
I'm trying to dig into Wasabi, learning how to make Wa3 components, and it's alot more difficult than it was learning maki.
However, I do believe that 5 megs of source code is alot better than one headerfile...
It just takes longer to learn, but it's worth it once the learning part is over..

-Plague

I have to admit snq has a point. Not that wasabi is a pain in the ass, I think I'm probably one of the biggest fans of the engine around, but I must admit I was suprised when wa3 came out that it had no standard way of doing visulizations. I mean, if you don't want to write an avs, what are your options? Do some small simple skin dependant vis like mmd3, or write a wac.

So now every vis that isn't avs is going to have it's own wac window? And an icon in the thinger? It seems like a bit of overkill. Hate to say the s word, but I was honestly expecting a sonique-style in-skin vis system, simply because it makes sense. Or, at the very least a standard vis window that can display avs and custom vis plugins, that maybe have a simpler architecture and a little less flexibility than a custom wac window does.

hell yea, that would rule!!
a Vis wnd that can show avs OR any other vis plugin!!

great idea!!

-Plague

i got a better idea... hard code sumthing in to winamp that will stop vis all together... why do people insist on flooding the database with thses awful things that say "all new for winamp3" when there the same old shit the put on winamp2, just changed the name... WTF!!!!!

hey,
It may be good to have this info in this thread:

How do Winamp3's components talk to the OS? And how do Winamp2's plugins?

Is the real advantage of components in that they can talk between each other? Or can they really access the system in a more open way?

You're full of brilliant ideas, aren't you.

All about Framework
 

I hope you won't mind if I elaborate on what SNQ said, because I kinda had the same feeling as he did when I first downloaded the WA3 SDK. So Brennan, if you read this, please don't take it wrong, that's not my intent.

The fact that you are exposing much more of the wasabi interface that you could ever possiby did with WA2 was indeed a good thing, and we are all grateful for that, there's no question. The thing is I'm not quite sure it was the right thing to do. My feeling is that you wanted to do things the right way so much that you ended up with something way too complicated.

Let me elaborate, you guys did wrote a complete framework, and by complete I mean *complete*, with everything a developper might need, even a string class. You guys had good reasons to do that, mostly because you wanted to make it platform independant (I won't debate whether it was really that usefull or not, that's not my point here).

As a resultant you came up with MANY abstractions layers on the SDK (and I assume in WA3 too), a code design with countless interfaces and redefined functions. And this is where I think that you went too far, by adding too much abstraction, you are not only loosing the begginer C programmer (and even pro programmers who want to have fun with WA3, who really enjoy derivating classes just to get the song title or pause the playback ?), you are also paying it full price in terms of speed and memory consumption.

Not that you didn't do a good job optimizing all that, but the fact remains, your object oriented framework looks like it's been designed as a complete OS framework. Not just a media player, or even a traditional app. People just can't expect to see the performances or memory usage of Winamp 2 on Winamp 3. You might say that when you see all the stuff it brings, it's worth it. But to who ? To the end user ? He usually don't care that his media player was compiled from the same code base for different OSes. He usually uses one OS, one media player, and that's it. He expects something that works and don't use too much ressources. So maybe to plugin developpers ? Most don't care about other OSes, and even if they use all the frameworks platform independant functionnalities (and a LOT don't, and just use Windows file, socket, or string/cstring classes), will they compile it for another platform ? Most won't, and you know it. So to who is it really worth ? Well, and don't take that wrong, again, but I think it was only worth to YOU guys, the WA3 programmers who made a nice framework with loads of functionnalities, programatical beauties, etc.

But nobody benefits of the extra overhead, you could have done the same thing in a much simpler way, with less abstractions and just forgetting other platforms ports. Winamp 3 would have been smaller, using less memory and CPU cycles. And it might have been out sooner in a really debugged way (you can't argue that it wasn't too soon to get the 3 version out in the state it was in, and yet, how long did the development took ? You know that better than ourselves). And please just don't say "If you're not happy, do it yourself", it's far from the topic here.

Again, I want to say that this is not an attack against you the programmers or even Winamp, but I think that you are too invested into wasabi to look objectively (I'm not saying I got a perfect objective judgement) at what you did and where it failed.

Re: All about Framework
 

This is the problem that's being attempted to be solved with Winamp 5 - There are some developers who aren't happy with the 5Mb of code, and just want simple and easy ways to do it, so the Winamp 2-style system is making a comeback. Hopefully Wasabi will improve, and it's certainly neater for writing completely new, very large and intricate components. And eventually you'll have an option to use either plugin system.

On a slightly related note, if anyone wants to document any of the Wasabi SDK, please send the completed files to me and I'll send the patches off to Aus for the next SDK update.

Re: All about Framework
 

Well, I know Darkain loves coding for Wasabi, so that's one.. :D
And I'm a fairly n00b (I know my way around but I'm no expert at all) when it comes to C++ and Wasabi, but I'm slowly getting the hang of it and I KNOW I will come to love it.. ;)
So that's two...

-Plague

Re: Re: All about Framework
 

Tough love :p

I didn't even use to like winamp2.x, I used Sonique. The only reason I use Winamp3 is because it has an SDK that lets me write the plugins that can do the things I want my media player to do. Things that would never be possible on Winamp2.x are not only easy on Winamp3, they're logical and well structured. So that's 3 :)

We could just aswell add Bizzy to this list aswell, cos he likes Wasabi more and more, just like me, and if someone tries to steal his beloved Winamp3, don't be surprised if he comes after that person with a hatchet... :D
So that's 4. ;)

-Plague

I just have one request which I know is echoed by many others.

For the love of God, make any additions like video playing/freeform skinning/cd burning/cd ripping OPTIONAL components/plugins.

I for one will never want an mp3 player that can burn/rip CDs, and I know others who do not want extra skinning.

I think offering these features is a great idea, but please keep them optional :D

they most likely will be.

they are.

Just a quick question and I really apologise if it's been answered already.

I know Winamp3 development will continue but not as the headline product for Nullsoft due to many factors but has the possibility of Winamp3/Wasabi ever re-taking the position as the headline product, once it has matured, been ruled out? (in the distant future once Winamp3 is considerably less buggy, bloated, etc) I guess I'm just interested since in an ideal world I'd love to see Winamp3 eventually take the top spot when it's really reached it's potential.

-----

One other thing... I can deffinately see what SNQ was saying about how Wasabi seemed to offer no real extra functionality to him and much extra complication which didn't seem necessary. I guess it must be hard to build a platform which is super flexible and ready for complex plugins but still enables simple tasks to be done easily. It's a bit of a tough one there =/

I know Winamp5 is a sort of merge of the two but I can't see how you can have everything from both. There has to be some sort of "meet in the middle" type of compromise. Sounds like a good idea to me.

-Dan

hmmm.

i don't use winamp3 at all, because i want winamp to just play my music - nothing else. i don't want it to play videos, rip cds, burn audio collections or organize my audio files (the worst thing i could imagine is auto-tagging!)

this is why i still LOVE winamp2.x - and because its fast, easy to use and resource-friendly.

winamp = audio player
and for everything else there are other programs, which don't play audio, but do the rest.

ps: i don't even use the skinning system, winamp is a tray icon 99% of the time
pps: get rid of official mp3 support, save the licensing costs and become an official partner of xiph.org :=)

Well.... Wow. But i guess nullsoft will do what it deems best for all users. I was a winamp3 disciple from the begining, so i find this odd. But whatever, as long as winamp keeps going.

I guess I don't understand what is so wrong with winamp3 that this needs to be done. I mean sure, make winamp2 plugins work seamlessly with winamp3 and revert to some winamp2 behaviors like the jump feature etc. Great, fine, awesome. BUT WHAT ELSE IS DIFFERENT on a user-interface level that this needs to happen???

Ok so it doesn't load in .002 seconds maybe that could be fixed by limiting wasabi a little as has been suggested. Maybe you could have some option to load wasabi core components into memory at startup so winamp would start faster (hey- IE does it and you don't even get a choice unless you uninstall it)

I think the only major thing is plugins and features.
So implement all winamp2 features in winamp 3, and make winamp3 fully support winamp2 plugins/ dsp etc. Maybe even expand so you could run the audio through multiple plugins say - dsp1->dsp2->wavout->diskwriter.

the biggest reason Winamp3 hasn't started up in .002 seconds is that Wasabi still isn't that optimized..
That stuff takes time.
Wasabi is still in heavy development and the latest wa3 builds startup ALMOST as fast as Winamp2.81 does!

-Plague

Sweet I know a lot of people still complain about load times.

on another note I was thinking, just to throw this out there, that maybe winamp3 could become open source.

I know this has been discussed before, but it may make more sense if the team is going to be more divided further(or will they?). Winamp5 (or whatever) wouldn't have to be os, look at mozilla vs. netscape (hmm, also aol/timewarner....) Maybe it could be limited to include a number of trusted but non-employed coders, so you don't have to set up a completely public cvs etc.

My feeling on this whole 'merger' of the two programs is that because of the fairly small team, there hasn't been enough time to make wasabi/winamp3 as polished as most of the public wants, especially with resources split between wa2 and wa3. Perhaps the incorporation of more coders would allow winamp3 to mature fast enough that the need for the proposed 'winamp5' would be short lived and winamp3 would gain wider acceptance in a timely manner. I really think that winamp3 is the better program and I think it's a shame to pull it from the forefront as the situation seems to be doing.