I have been doing a lot of coding for Nucleus lately. It's mostly plugin work and some core hack. I'm planning to contribute to this project since it's a cool software that I'm sure can helps make better. Also, I want some experience on PHP/MySQL and web application development.

On the other hand, finally able to turn on FancyURL for the site! However, now I have to visit all functions to make sure all plugin works with FancyURL (many are broken). Also, posts are need update to point to the right URL.....

Update:
- NP_PDF has the wrong URL and permlink and formating is screwed up
- NP_Print has // in the URL
- Some links have blog/blog or blog/blog/blog It actually just how FancyURL works....
- NP_WIE's spy mode has wrong member URL
- NP_RSSItem didn't work
- RSS 2.0 skin still using the non fancy URL format Fixed
- PageLinkList doesn't work Fixed, thanks to Radek
- Need to test FancyURL off

So, my friend Stephen working for Hutchison Wireless in Hong Kong gets a 3 UMTS phone. He invited me to join the 3.hk on-line videoconferencing service. However, it only supports MS Netmeeting.....

It's just that H.323 (or Netmeeting's implementation?) uses dynamic ports for connection. So, you can only get it to work by putting your PC on the DMZ. Of course, I can't do that since I don't want to put a machine on the net unprotected (yes, there is personal firewall... but....) and I'm running a web server which is behind the firewall happily using port forwarding. To make it more complicated, my new housemate Martin might want to use Netmeeting as well.

I would think standard like H.323 have take into account for working in different enviroment, but did it? Maybe I shoudl take a look into it.

Did I say it's suck not able to get it to work?

Update: So, I saw some posts on the Internet regarding who's to blame. here...

It basically said H.323 is developed standard before NAT is widely used and it is working for many routers.... right! But NAT is deployed in millions of home around the world now and what to blame here? I just hope they fix this in the future.

NetMeeting h.323 and a NAT router

I just installed Nucleus 3.0RC and moved my setup from 2.5CVS. It goes relative smooth, but some skin, namely RSS* somehow no longer works.... going to fixed it later.

The steps I did it:
1 - download, install in www directory
2 - copy config.php from old setup
3 - copy all plugins from old setup
4 - copy all media from old setup
5 - copy all custom skins fro, old setup
6 - remove install.*
7 - chmod ugo-w config.php

Update:
- Also copied additemform.template for NP_NotifyMe

Still not working:
- RSS2 feed (a bug in skin??) It's a bug in Rss2 skin...

I spend the time last night and this morning sorted out the blog to make it XHTML 1.0 compliant. It's cool!

The CDMA Wireless Data system I worked on for the past 2 years is finally into testing phase. To test the system with real traffic, we deployed 100+ PDA phones to help generate the traffic needed. I'm lucky to receive a Kyocera 7135. It's basically a Palm PDA combined with CDMA 1xRTT phone. It loaded with email and web browser function. I installed many useful toys and programs like SSH/telnet/FTP/NNTP clients and VNC.

It's a novel expereince to be able to read email from work while shoppng gocery, or wake up in the morning by email notification regarding a cancelled meeting. I can't wait to try using VNC to check on the progress of my loadbuild while walking to Tim H. for a donut.

I shall be a slave of this device.

Alright, the Asus TV FM card finally arrived.

I installed it on my server, recomplied the kernel with some patches, download and install the Philips saa7134 driver. The driver is able to load now, but there seem to be some configuration problem that xawtv did got any channel..... hum, need more debug.

My plan is to get the server to schedule, record and compress TV shows, stream TV real-time over the network.

Cool, eh?

Damn, if I can get it works with digital cable that will be perfect.

[update]
So, I upgraded to Fedora core 2, which has the 2.6 kernel that include the driver. I was able to load it, but yet to try recording with it......
[/update]

I just upgraded to 2.5 CVS, where a lot of bugs were fixed. It includes the prev/next links and search functions.

It's cool!

So, it's been at least 10 years since the first spam appeared on the Internet.....

Yes, this topic has been beated to death so many times by now. But I thought I can used a summary on battling the lunchie meat.....

Email is transport on the Internet using a protocol called SMTP. Security is not a great concern in the design of SMTP. In fact, one used to able to telnet to the SMTP port (port 25) and send a email spoofing as someone else. Sendmail and etc fixed this problem now, but it demostrates the lack of authenication scheme in SMTP to verify the origination of a message.....

Than they open up the Internet for business, and spamming becomes an "industry" by itself. There just a "market" for it and cost is low. It invaded Usenet, violated the WWW, brutalized SMTP, and causing a lot of grief on the cyberspace.

In the early days, users were fighting back by complaining to the sys admin and most of the time it will be taking care of. This usually mean the offending party is warned and blocked off/suspended. But as spamming become overwhelming, most users get tried of it and stop complaining. It is futile to report 100+ spams on your mailbox everyday!

On the other hand, spammers become more and more evasive, they spoof with non-existing email address, use badly configured domain to relay their spam, and now even using virus/worm/trojan/p2p to create a distributed network of spam zombies to avoid policing from their ISP and lower cost.

As everyone is getting tried of the situation, many solutions are developed or proposed to battle it.
- Client-side mail filter that drops spam as they are downloaded from the ISP. Many difference algorithms are developed. (e.g. Bayesian, rules-based)
- Server-side email gateway software/managed service that inspect mails before deliver it to the recipient.
- White/Black list that only accept known good originators and reject bad guys.
- Regulation effort to outlaw spamming on the Internet.
- Challenge/response scheme, a challenge is sent to sender if a mail is received on a protected box. No mail delivery to user until correct response received.
- Fix SMTP with sender authenication (e.g. CallerID for email, SPF)
Taxing the sender in some form.

We know none of these technologies are perfect. The mail filters work in various degree of success. The problem is that spam sometimes still slip though or false positive happens when real messages got dropped. It often need to combine with other techniques to improve the effectiveness. Also, the filter works on the recipient side in a passive manner. It does not cut down on the wasted bandwidth on the Internet.

The white/black lists need quite a bit of management, and it often blocked innocent server/domain. Furthermore, the solution required intensive processing resource.

The legistration path only work in a regional basis. ie. It won't works on all the spam from Taiwan I got on Yahoo mail. Also, policy-maker often not addressing the issue.

The root of the problem is that spammers are sending mails spooling as someone else. And there is no infrastructue in the MTA to prevent it.

To me, the CallerID/SPF is a good direction if it got deployed widely on the Internet. It helps MTA to detect whether a message is sent from a valid user/domain, and block all senders with unknown identity. I believed a lot of spams these days are sent by domain spoofing, which this technology should take care a big part of them.

More links:
- Spam, Spam, Spam, Spam, Spam, The FTC and Spam
- The Economics of spam
- Comparing spam filters
- Filters that Fight Back
- Spam: The Phenomenon (A very good summary!)
- 82% spam on the net - more bad news