Develop Reference

Is phaser capable of large multiplayer games?

Newbie here. I am working with phaser, specifically with the isometric plugin.
I would like to know if it is possible to create games in phaser similar to agar.io, in terms of handling real-time multiple connections, generating a enormous map with about 300 players in it and all this without having too much impact in game performance. I seriously don't know how to handle the multiplayer part (probably sockets, node.js) for it to work really well. And as for generating a really big map I am quite blank too.
Is it possible, in phaser, to create a isometric-type game that handles multiples real time multiplayer and HUGE maps that are generated when the user gets to the edges of the visible "map"? How?
If not, what should I opt for (game engine in js and other applications) in order to achieve what I want?

You're not asking the right question, but you're close!
Your first guess is correct. You wouldn't handle multiplayer with Phaser, you'd use web sockets, or nodejs, or some other backend. So Phaser does not really limit you in what you can create with regards to multiplayer, since none of the networking code has anything to do with Phaser.
The idea of handling a huge map also just depends on how you optimize your graphics, regardless of what platform or framework you're using. For example, if you have huge or infinite maps, you can always just only show what's on screen, or around the edges of the screen, and use object pooling to show the rest of the map as the players move.
For multiplayer in Nodejs, check out Socket.io. It's really easy to use. I've set up a barebones example using it here. And in case you might find it helpful, here's an open source game I made for Ludum Dare in Phaser, with networking (this one is only p2p, so it's only made to handled 2 players connected to each other, but like I said, that's only a limitation of the multiplayer framework I used, in this case peerjs.com, and has nothing to do with Phaser itself, which can take care of all your rendering and game logic needs.)
Hopefully this helped answer some of your questions!

Phaser (at least in its 2.0 version) is not a good candidate for implementation of real time game networking as explained here.
If you're looking for a Javascript Multiplayer game engine you should check out Lance, which was written specifically for this purpose. You can then choose a renderer of your choice (Pixi.js, for example, if you're aiming to implement something like Agar.io. It's the same Renderer Phaser uses)
Regarding PhasedEvolution's comment above - Firebase is a nice tool if you're doing turn based games. It's not up to par for real-time game development as it doesn't allow low level access for any game-critical features that mitigate latency like client-size prediction, bending, interpolation and extrapolation.
Proper disclosure: I'm one of the co-creators of Lance :)

Rendering huge, interactive SVGs in a browser

We have to display huge SVG documents (about 20mb) inside a web application. Users should be able to zoom in and move the image.
Rendering the SVG directly as a DOM object is too slow and the performance is inconsistent. The same applies for painting it on a canvas.
Generally, handling SVG on the client side seems weak. So I thought of implementing a server-side solution for providing the data in small chunks, in a non-vector format. If the user is not interacting with the document, the buffer starts lazy loading higher detailed pieces. My concern with this solution is, that the network traffic could be critical.
We will be rendering 2D DWG / DXF files, which will be converted to SVG.
The AutoCAD API seems really slow. The DWG sample does not work on any of our devices. Also, the application has to run without an internet connection, so we can't use the AutoCAD REST API.
How would you solve this? Are browsers even built for handling huge vector graphics?

When it comes to SVG it depends on the number of nodes, gradients, opacity and blur effects; however, why not use the end-user's graphics accelerator to handle this?
Most modern web browsers are made to support graphics acceleration through WebGL -with which you can build very complex (and "huge") rich graphics in 2D (or 3D) that is handled as fast as your graphics accelerator can handle it; exactly like modern games.
Using a WebGL library is recommended where a lot of work has been done for you already:
PlayCanvas : https://playcanvas.com/ -- you can import other formats, build & script your scenes with a friendly interface. PlayCanvas is well documented.
Three.js : http://threejs.org/ -- an advanced WebGL library, aimed at coders. ThreeJS can also handle many different types of 3D formats and this library is also well documented.
With Three.js you can also render your graphics as SVG, however, using WebGL is recommended for the obvious advantages in speed and quality.
Both of these libraries are very powerful, have an active community and is well supported in modern web browsers; however there are many others you can try.
For more information on the libraries mentioned above, it's best to visit the sites where extensive information and examples are available.

Instead implementing yourself, I would suggest you use the Autodesk Viewer, also available for developers with full REST + JavaScript APIs.
Basically this library will convert your DWG file (2D or 3D) into a JSON stream and adjust the amount of data according to the browser/device capabilities. It uses Three.js, but you don't need to handle the geometry directly (but you can).
Check the Forge Github for samples. I like the Galley better.
You may also run it locally using NodeJS to server it to the browser. The Extract sample does the whole process.

WebGL graphics rendering engine

So, I have thought about designing a WebGL Graphics Engine, which will facilitate designing 3D Interactive Graphics for web. Now, my question is :
WebGL is Javascript API, so in order to design an engine for WebGL graphics, do I need to have a JavaScript compiler or anything ? What I want is a system which will let users see what they are creating (for example, like blender workspace, if you draw up a scene you can see and make changes simultaneously)

You would have to create some kind of engine, or framework that you'll build your system onto.
Creating only framework/engine would take at least 2-3 months, and if you plan creating something really big and advanced, that supports various effects rather than simple rendering primitives, than that might come down to 5-6 months. After that you could start creating your web application. So 6-7 months time for that? That shouldn't be a problem.
I don't know how advanced you are, how many people are you working with, but that seems very plausible and doable. But is worth of it? In a year, many different things will change, maybe new openGL ES version for webGL, changing api, supported browsers (IE recently joined the game),... it's really questionable.
You wouldn't need any kind of JS compiler or anything like it, just knowledge of advanced JS and many different techniques used in 3d, and since you plan building system that is far beyond just-graphic-stuff, then it adds even more to overall complexity and time consumption.
So, to answer your question: yes, it's very doable in a year, but will it pay off?
Similar things already exist in some form:
http://errolschwartz.com/projects/threescene/
http://badassjs.com/post/12885773103/threenodes-js-a-visual-webgl-scene-editor
CopperLicht engine has its own real-time editor
there are more lab-playground-like editors

Is there a way to write javascript interface requirements?

I'm the leader of my development team, consisting of:
Back end developers (PHP and API's /Frameworks, CMS backend, server side technologies)
Front-end developers (JavaScript with APIs, client side technologies)
HTML/CSS Ninjas that work closely with the graphics ppl
other staff related to development process
Lately, projects have many requirements related to user interface, lots of them comming from the graphics / creativity ppl and the user itself. The requirements sound like "when the user hovers the logo, the letters should scatter in horizontal opposite directions, and the logo should fade-in while moving up in a smooth movement animation where the speed decreases while the logo reaches the target". It's my duty to document these requirements to send them to the front-end developers.
I was wondering if there's any way to document such things in a way that's good for everyone. Lately, describing animations and such has been a pain, and the documents are good for nothing.
My enterprise is in a position where the creativity staff and the javascript staff can communicate with each other directly, but we are having trouble monitoring the process, estimating times / effort and filling up metrics.
Can any one give some idea to document such things? I'm sure it's not only happening to us... I'm loooking for an organized / structured way to make a document / whatever that I can give the javascript ppl (along with the HTML /CSS that make up the web page), that they will surely understand without even asking the creativity ppl directly, allowing them to start working immediately without further communications.

How about create primitive mockups of the intended animations, say, in PowerPoint or flash? The creativity team wouldn't need any programming skills to use those, and nothing brings graphical stuff better across then graphics. Even pictures showing the desired start and end state of animation, maybe with supporting arrows etc., would probably be much more helpful than text (or, at least, text itself).

HTML5 Canvas / Flash substitutes?

I'm looking for a good plugin for:
a) rapid rendering of
b) lines, shapes, and imagery on top of a
c) rectangular canvas area who's size can be declared at load time.
It needs to run at 20-30fps without putting a heavy load in the browser. It also needs to be able to interface with JavaScript and the DOM.
Creating my own Flash plugin is the first choice, but I'd like to aim for a free, open-source and/or non-proprietary solution first. HTML5 canvas is out of the question - it renders way too slowly.
Anyone happen to know of anything that does these features? (I'd even be okay with a pre-made Flash plugin that meets the requirements mentioned above)

Your request is still vague. What do the lines and shapes need to do? Sit there looking pretty?
If your lines, shapes, and imagery are going to be fewer than say 5,000 objects total, I'd recommend using SVG and perhaps the Raphael library to go with it. Every SVG object is a DOM object from the get-go, which will save you some associated headaches with trying to use Flash or Canvas.
If you really need a lot more performance or plan on having 50,000 objects on screen, Canvas may be for you.

"Creating my own Flash plugin is the
first choice"
I'm unclear what you mean by "plugin" here - I assume you just mean "flash file" (an SWF?).
I think #WTP is making a good point. You say "rapid rendering" but of what? how complex is it? Flash has very good speeds when it comes to vector graphics, and much faster rendering of optimized bitmap data (the technique of choice is blitting here). It all comes down to optimizations / complexity of graphics. No matter the plugin / tech solution, you will always be able to cripple a machine with inefficient design.
To answer your question, Flash will definitely meet your needs.
I would also venture that Canvas/JS would as well, but apparently you've tried that already? I've seen quite complex scenes running quite rapidly, so that surprises me.
I'll note, also, the upcoming Molehill APIs for Flash. This provides low-level access to the GPU - and will create the potential for breathtaking 2d/3d performance in the browser. But its still in alpha, so... don't hold your breath ;)

I know of only four major players in the "Vector graphics capable" department -- HTML Canvas, Flash, Silverlight, and Java applets. Aside from canvas, all them are proprietary in some way or another. The good news is that all of them have the ability to compile for free in some way or another and they are generally faster than canvas by my understanding. Now, I happen to know Flash so that might color my opinion, but I am fairly certain that it is your best option. It has decent performance and a solid install base. It also runs on Linux and does not raise major security issues.
Look up the Flash Builder (Flex) sdk. There is command line compilation for it.