Don't listen to them.Isn't Pulse just a replacement for ALSA+z, z being JACK, ESD, etc?
Pulse versus Jack:
PulseAudio is a daemon. Pulse is a audio API that includes transporting audio. Same thing with Jackd. Jackd is a audio server and Jack is the audio API that does transport stuff, too. So this "transport versus audio daemon" is just a bogus comparison.
The difference between Pulse and Jack are their goals. Jack is designed for audio production and Pulse is designed for desktop audio. The goals of both are both with audio, but completely different aspects of it.
Jack is designed for low latency and high performance audio. It is designed for low latency because if your doing recordings from multiple inputs or if your doing DJ stuff or any sort of live recording or production you do not want to have high variations in managing those multiple streams.
Also it is designed to facilitate audio workflow. So that you can route MIDI and PCM (uncompressed) audio between multiple applications and build more complex audio applications.
I own a midi controller. It looks like a piano keyboard, but it does not generate any sound. Instead it generates MIDI signals and connects over USB. So I connect my MIDI controller to the computer via USB. I route the MIDI signals from the USB input and through to Jackd. Then from Jackd to a software synthesizer. A software synth is a program that takes the MIDI signals and uses them to generate audio noises. Like I can generate piano noises, or drum noises, or barking dogs or whatever. And it'll vary the pitch and rate and volume based on my MIDI signals. So from the software synth I take the PCM audio output and route it through "Alsa modular Synth". AMS is a another synth program, but I can generate all sorts of chains of midi or PCM audio processing in there. A bit like a programming language, except its all graphical. I use AMS to add some reverb, echo, and some stereo seperation. That way I can make the piano noises sound like they are from a big piano in a hall, which sounds cool. Then from AMS I route the PCM back through Jackd and then out through my sound card.
USB-MIDI -(MIDI)-> Jackd -(MIDI)-> Software Synth -(PCM)-> Jackd -(PCM)-> AMS -(PCM)-> Jackd -(PCM)-> Sound card -(PCM)-> Digital receiver -(Analog audio)-> Speakers.
Pretty complicated, right?
Well that sort of thing makes it very cool for audio production. But that same nature makes it kinda lousy for desktop.
Low latency is a trade off. You are trading off the ability to get something done quickly with the ability to get something done efficiently.
Meaning higher CPU rates, higher battery usage, etc etc. Plus it is not a easy concept for people to understand and so on and so forth.
Pulse Audio on the other hand is designed for efficiency and to be desktop friendly. Newer versions of PA do a good job of helping you configure your hardware easily.. you can just go in and select "mic 1" or "mic 2" for external and internal mic, or select digital out or surround sound or whatever. Prior to that you had to manually configure it using something like Alsamixer, which is very very low-level hardware mixers and such and is confusing and misleading to use (since most hardware sucks). Also with PA it can lower your battery usage and other things, IF it is correctly configured (which some distros, like Ubuntu, have had a hard time figuring out how to do).
Two very different goals, but have a similar approach. Transport layer and Audio daemon.
Like your display is managed by a server, so now your audio is.
And, in addition, like X Windows pulseaudio can provide network transparency.
At home when I use Pulse, for example, I will often watch movies or listen to music on my laptop since the laptop is convenient and easy to carry around with me. However I want to use my nice stereo instead of uncomfortable headphones and I don't want to have a very long audio cable I have to string around. So instead I have my desktop hooked up to my home audio and I play music and watch videos by sending audio from my laptop to my desktop.
Another thing PA is good for is handing things that are sorta-sound cards. Like USB audio and Bluetooth audio. With modern versions of PA I can just plug in a USB audio device or connect to my Bluetooth headphones and have everything "Just Work". Jack is not designed to really handle issues like that.
Pulse versus ESD:
Pulse is a drop-in replacement for ESD. It should be compatible with it so programs written for ESD will still work with PulseAudio. However Pulse is superior to ESD in pretty much every category.
Pulse versus Alsa:
This is more complicated.
Alsa, although it was the best they could think of doing at the time, is badly designed and confusing. Alsa tried to be too much. It takes care of very low level hardware drivers and basic configuration stuff... AND it tried to be a high level API for audio applications. So it is confusing for developers. The Alsa API is big and covers too much ground and is not really well documented.
Nowadays we know that like a display server (X, Wayland, OS X's Aqua or whatever) we need a audio server. Back in the bad-old-days with the old Linux OSS drivers every thing was accessed like it was a file... you wrote PCM to a file to play audio and you read PCM from a file to read from audio input. Very simple, very easy to program for. However Audio is much to complicated for such a simple way to access stuff. You have transport issues, timing sync issues, latency, multiple audio streams going in and out, etc etc. Modern versions of OSS and Alsa try to provide for those extra stuff, but really a audio daemon type setup is a much simpler and cleaner design.
So Pulse is meant to be a replacement for the portion of Alsa that is designed for providing high level APIs for audio applications. However it is NOT a replacement for the low level features for Alsa.
So low-level programming, low-level configuration, and audio drivers are going to always be Alsa. High level aspects should be taken care of by Pulse.
PulseAudio vs SDL/OpenAL/Gstreamer/libcanberra/Phonon/etc
Those are all high-level audio programming APIs. All of those are the actually what you want to program for if your a normal application developer. (If you are doing audio production stuff you will probably want to program in Alsa and/or Jack.)
Like people want to make a big deal of PA versus Alsa versus OSS etc etc. But it is actually very simple for application designers.
Say your programming a video game. If you program for Alsa you can only support Linux. If you program for OSS you only supporting FreeBSD and Solaris (and a tiny number of Linux users that install modern OSS). If you program for PA you can only support Linux dsektops running PA.
However if you program your video game audio with SDL you can support Linux/FreeBSD/Solaris/Windows/Playstation/AmigaOS/Mac OS X/etc/etc. SDL is higher level so you don't have to give a shit about what low level stuff people are using. This is the same for Gstreamer and OpenAL and a whole bunch of them. They each have their own plusses and minusses and different design goals and support lots of different platforms so it can be a bit of a hassle, but usually it is not too difficult to figure it out.
It is like the difference between programming for X Windows and programming for QT. If you program for X Windows display server then you can only support things like Linux, Solaris, or the BSD systems... but if you program for QT then you are programming for all the platforms QT supports.
It's very relevant to the idea of PC gaming. It means you can just buy an Athlon X2 240 and an HD4670; a gaming computer built like that will cost you maybe $400 with a hard drive. Yes, outside of the sale that Valve has going on right now, your games will probably cost more. But it'll also last just as long with the same hardware as any console, and can be upgraded to the latest generation, and we've gotten to the point where you *need* a PC, so might as well buy a $50/$60 HD4670 that doesn't require more than ~50W/60W on load (and estimates say something like 3W-10W idle) and make your desktop a gaming computer.
Yes, but most people are buying laptops nowadays. Desktops may be cheaper, but laptops are a hell of a lot more convenient and not much more expensive. Notice how in the previous threads people kept on harping about MMOs being very popular on the PC? Guess what most MMORPGs have in common... they do not require expensive computers. I can happily play WOW on a computer running (*gasp*) Intel graphics (which is the most common chipsets used nowadays and Intel sales are quite a bit more then Nvidia and ATI _combined_). Steam is a good example of something that is still popular on the PC, but the Wii/PS3/Xbox360 all have online sales and games, too.
So on and so forth.
PC will always have better graphics, but most people don't really care and to have those better graphics it is going to take a lot more then a 400 dollar PC.
(in before laptops. I really hate the trend towards using laptops instead of desktops. if it's not being brought around, it should be kept under the desk.)
Laptops kick ass in a huge way. Desks are for being stuck at work. Who wants to do facebook and go youtube-ing at a desk if your at home? Desks at home are for doing taxes or balancing your check book.