'Donate' isn't the right word. You just add some resources you choose to the community resource pool.
These resources will be used to achieve better performances/capabilities for the whole Grid.

A permanent node, is a node running for as long as you can, it starts when your computer starts, and stops when huu.... you get the picture. You'll be getting much better results, as your node will be a lot more connected to the grid, and will have opportunities to do some optimizations a temporary node can't.

Very Yes. But the best way to improve results a lot, is to have lot's of people running permanent nodes. Spread the word.

Each node in the Grid gets a unique ID when it first starts. This unfalsifiable ID will be used for identification (see below) when using secured objects (files or chat's digital personalities for instance) of the grid. This ID also provides protection against a bunch of attacks at the Grid.

Anonymity is the inability to link a NodeID to a physical person (read IP address). Identification is the ability, for instance, to know what node 'owns' a shared file. Identification is provided through the use of strong cryptography (digital signatures), and as direct benefits, allow to resist to fake attacks (only the owner of a file can modify records he created) or allow content/nodes blacklisting without compromising anonymity.

First of all, Nodezilla try to provide anonymity, and next Nodezilla will be extended with features that only a secured grid can provide.

See here.

Short answer: they aren't.
Both are targeted at large scale grid networking, but they do not use the same kind of algorithms (this is from my understanding of this document).
Freenet uses statistical routing which try to 'learn' where keys can be found from previous answers and answers latencies, EVL uses deterministic routing where a node knows for sure where a key can be found (or can be sure the key doesn't exist), no learning time, of course, as a drawback the join time for a new node in an EVL topology is longer.
In this respect it's much more closer to Pastry grid topology.

No. Only the Network Agent is required to run, it does all the logic. You need the client to search and start a download, once a download is in the queue you can close the client. All downloads are performed by the Network Agent.

Yes, see Nodezilla4Web.

Simply put, you can't. The grid construction algorithm ensure that every objetcs in the grid can be reached through a route of different nodes, but that doesn't mean (of course) that each node knows each other nodes (this is why we can build anonymous services on top of the EVL Grid). As a direct consequence it's not possible to have an accurate number of participating nodes at a given time.
It's is however possible to compute through statistical mathematics a rough estimate of how many nodes are currently online by examining the content of a routing table. This is currently not implemented.

As explained in the previous topic, a node doesn't know all other nodes. Typically a node doesn't know how many peers are sending data for a given file request, it just receives data through relaying nodes, hence it's not possible to count sources (and if you understand the Architecture Document), it has no sense on an EVL Grid.

These errors can be safely ignored. They are logged for debugging purposes.

TCP Port 4242 (by default, or the one specified in your nodezilla.ini file otherwise) should be open for inbound connections. All ports should be open for outbound connections for nodezilla.exe process as not all Network Agent will listen on the same port.
See more infos here.

If you use NAT or go through a gateway, you need to enable Port Forwarding on your gateway for TCP/4242 to the PC running Nodezilla.
See more infos here.

List of routers that have problems with P2P applications, if you have one of those... either try the solution specified with it (if there is any), update the firmware or get a new one.

Reason not yet verified

• Apple Airport Extreme (802.11g wireless with 128-bit WEP)

Due to too many connections

• SpeedStream 5660 in Router/NAPT configuration. There is no firmware newer than 2.(3).7. Switch to bridged mode and firewall your network to fix it. Otherwise, when it dies just power-cycle the router and continue on.

The following modems/routers have known problems with too many global connections, limiting them to 200 or less should fix the problems:

• D-Link 302G
• D-Link DI-624
• D-Link DSL-G664T
• Linksys BEFSR41V4/BESR41
• Linksys Wireless-B
• Netgear DG632
• Netgear DG834G
• Netgear MR814
• Netgear WGT524
• Netgear Rangemax 802.11n WPN824
• W-Linx MB401-S (and SMC Barricade 7004 BR, which is identical in construction)
• Westell 6100

Due to UPnP

• Most D-Link 5xx and 6xx (can also be caused by too many connections)
• D-Link DI-604
• Dynalink RTA1025W
• TP-LINK TL-R410
• ZyXEL Prestige 660H(W) (Firmware versions PE8+ will fix this)
• Some SpeedStreams

Due to Port Forwarding

• D-Link DI-514 (not port forwarding UDP protocol consistently)

Common fixes for routers/modems freezing/crashing/etc

• Updating the router firmware
• Turning off the UPnP (in the client and in the router) and doing manual port forwarding
• Restricting the number of global connections to 200 or less (may need to go below 100, depending on the router)
• Putting the modem/router into bridge or gateway mode (doesn't apply to modems that aren't routers)
• Getting a new router/modem

Sourced from (with some modifications): Bad routers - AzureusWiki

The following note does NOT apply to WRT54G/GS v5 and up! Use the latest official firmware (1.00.9+) with those, they do not suffer from this specific problem (though they do suffer from different problems).

The default firmware for Linksys (and all replacement firmwares except for the latest DD-WRT and HyperWRT Thibor) have a severe problem where they track old connections for FIVE days, which causes the router to hang when using P2P apps, or any software that generates a lot of connections.

Linksys has yet to address this issue, but there is a fix. If you use alternative firmware, you can put in a start-up script to fix this problem. DD-WRT and HyperWRT support custom start-up scripts.

• HyperWRT Thibor (14+) (54G/GL/GS only, excluding hardware revision V5 of 54G/GS)
• DD-WRT (v23+)(54G/GL/GS excluding hardware revision V5 of 54G)

I am not responsible if you screw up your router, so you do this at your own risk. This page has instructions on recovering a bricked router.

If you experience any strange issues or performance problems, clear your NVRAM, and if all else fails, try changing firmwares (but make sure you clear the NVRAM afterwards!)

You should avoid upgrading over a wireless connection, since the connection may drop out and brick the router.

To upgrade the firmware follow the instructions on the download section of Thibor's page if flashing to HyperWRT Thibor, and make sure you download the correct firmware version for your router!

Neither DD-WRT v23+ nor DD-WRT Thibor14+ require these steps, but older versions of HyperWRT tofu/Thibor do. If you are on an old DD-WRT, you MUST upgrade to v23 SP2 or later.

• The following instuctions for HyperWRT tofu and older Thibors.
• Go to the web interface (default password is admin) by typing http://192.168.1.1 into your browser.
• Click Administration, then Edit Startup. Put the following commands in the box:
  

echo 1 > /proc/sys/net/ipv4/icmp_echo_ignore_broadcasts
echo 1 > /proc/sys/net/ipv4/icmp_ignore_bogus_error_responses
echo "600 1800 120 60 120 120 10 60 30 120" > /proc/sys/net/ipv4/ip_conntrack_tcp_timeouts

• Click Save, close the window, then click Save settings.
• Reboot the router (the Administration page has a Reboot button) and you'll be done.

To upgrade to DD-WRT, follow the instructions on this page. WRT54G v4 Installation Tutorial.

• The following instructions are for DD-WRT v23 or later only.
• Enter the following values at 'Web-Admin -> Administration -> Management -> IP Filter Settings'
• Maximum Ports: 4096
• TCP Timeout (s): 300 (decrease if you have many TCP connections)
• UDP Timeout (s): 300 (decrease if you have many UDP connections)
• Save Settings and then Reboot Router

For a detailed explanation of what these settings do, see Router Slowdown @ DD-WRT Wiki

Remember, I am not responsible if you screw up your router, so it's your problem if something goes wrong. However, if something does go wrong, this page has instructions on recovering a bricked ("dead") router.

If you require further assistance with either firmware, please visit HyperWRT.org (the main site does not have the latest builds however), HyperWRT Forum's Advanced Help Section, DD-WRT.org, and DD-WRT's Wiki.

Yes, through NetDiagnostics an independant utility located here: http://jws.nodezilla.net/dev/jws/netcheck.jnlp . The documentation is here

Click here, select the Nodezilla Client application and then go to the "Application" menu. Select "Create icons". You're done.

This is normal behavior. As, for instance Seti@Home, nodezilla uses your idle CPU power to do the forward error encoding. Even if your CPU is 100% used in your system monitor, you should be able to use your computer as usual with no slow down as only the UNUSED CPU is taken by nodezilla.

This process runs as a Windows service. To stop it, either use the 'Stop Network Agent' shortcut from the start menu.
Or, from the 'Start Menu', select 'Run', and enter 'services.msc', you will see something like the following, select the 'stop' button on the Nodezilla service.

The nodezilla.exe may take up to 2 minutes to terminate cleanly.

This happens mainly on XP SP2 systems. See there for a fix: http://sourceforge.net/forum/message.php?msg_id=3352145

These are not processes but threads of one process. This is normal behavior of Linux systems.

Double check that the directories used by Azureus, and the Nodezilla Network Agent to store files and torrents are writable by BOTH the Nodezilla Network Agent, and Azureus.