spam
This might be a very dumb idea or somebody aleady tried it if it is
not, but it just occured to me. I have an e-mail address at ISP. I
haven't been using it but since the address has been valid for several
years I receive lots of junk mail. In fact, almost all mail I receive
there are junk mail. So how about having an e-mail address that can
be easily harvested but not used for real communication to filter spam
mail in main mail box? One simple observation of those unwanted mail
is that those are sent to many addresses at almost the same time. So
it might be possible to filter out unwanted e-mail by just comparing
mail sent in similar time period. If a mail "matches" one in the spam
mail address then it's definitely a garbage.
I know there exists a spam filtering system using kind of a voting system.
However, for the system to work participants have to behave correctly.
But with spam only address, nobody has to behave. In this system, spammers
kind of identify themselves as such.
Hmm, quick check showed that actually there are not many identical unwanted mail
going to my two different mailboxes independently. But it might be still worth
experimenting. Knowing something is 100% spam helps traditional filtering
even when this method doesn't work quite well.
[]
permanent link
Open Source Search Engine
I searched for open source search engine, or rather indexing library for my
intended use.
Bunch of software is listed here.
Other engines and libraries like
Estrainer and libibex are not
listed but it should be a good start. It helps to know the name of existing
libraries to get good search result.
[]
permanent link
Broadband
From now on, I can connect to the Internet using 100M optical fiber connection.
I finally got tired of 32K link early this month and applied for this service.
After waiting for three weeks, it's finally arrived. Woohoo!
[]
permanent link
Dvorak layout
In part of my little campaign to protect my hands, I switched to dvorak layout
about a week ago. I've finished gtypist typing course.
Now my typing speed is about 35 words per minute which is much
slower than before. But I'm getting better at it as it goes. One problem is
that I sometimes still have to think a little to input rarely used characters.
This disturbs my thinking process. The bigger problem is that I have to
adjust to XEmacs key binding again. This will take more time than just
typing. I hope this investment will pay off in the long run.
[]
permanent link
Linux and USB Audio
I bought DENON D-MA3 audio system on last Wednesday and it arrived yesterday.
Since it comes with USB input and can be used as USB speaker, a natural thing
to do is to make it work from my computers which run the Debian GNU/Linux.
I spent some time looking for information how to do it and finally able to do it.
After you found the solution, it was very simple but it wasn't very obvious.
So here's a note in case some one is trying to do the same on Deiban unstable.
There are only two things you have to do:
- Enable snd-usb-audio module in ALSA
- Modify /etc/usbmgr/usbmgr.conf and add an entry for your device
To do the first, I have to run dpkg-reconfigure alsa-source
to add snd-usb-audio because I had a configuration to compile only snd-intel8x0.
Then just follow the usual course of make-kpkg modules_image
to create alsa-modules package and install it. It doesn't look like you have
to add the following as suggested by some documents I found on the web but YMMV:
alias snd-card-1 snd-usb-audio
alias sound-slot-1 snd-usb-audio
Then I ran lsusb and found that 0x08bb and 0x2702 are
vendor and product IDs of this device. Finally, I added the following to
usbmgr.conf.
vendor 0x8bb product 0x2702 module snd-usb-audio
Using this from XMMS is also very easy. Just go to preferences and choose
Configure in Output Plugin. Then choose the appropriate output device and
you are all set.
[]
permanent link
Vncrec
Recently I was pleasantly surprised to find that there's an excellent article about
vncrec on Linux Gazzete. vncrec is a enhanced version of vncviewier to record
a vnc sesession. I spent just a couple of days to modify vncviewer to record
RFB events to a file and then play the recorded file. The software didn't get
much attention when I first wrote it more than a couple of years ago but
it seems it became a little bit more popular this year.
As always, cool things about open source software is that other people
who think the software cool sometimes contributes features back to the
software. Features like converting the recorded session to series of image
files or directly converting recorded file to mpeg movies are provieded
by others. And those are what made this "I wrote this just because I thought
it's cool" piece of program to a 100 times more useful software.
It never attracted enough people to build a development community
around it but it's nice to know the software is used somewhere in the world.
[]
permanent link
Translation Memory
Wikipedia says that Translation Memory Manager is another name of Translation
Memory. I think TMM is a better name.
There are Open Source translation memory managers.
The first two works on Linux but the ForeignDesk is Windows only.
Unfortunately none of them matches my taste. It is a firm requirement for me
that a tool must be able to be used from XEmacs. I only needed a backend
program that given parallel corpus, perform exact or fuzzy matching on
input string and return the translation if found. tmdaemon 1.1 in OOxlate might
come close but it's not obvious how to use it despite its short length
(463 lines of code). I also think I have to write lots of code to make it
do what I want it to do.
I guess I have to write my own program, then. I don't like re-inventing a wheel
but I believe I've done enough homework. Hope I can come up with something
usable without too much effort.
[]
permanent link
Yes, this is yet another weblog
Just trying to start my weblog here. As always, learning curve is an issue.
I'll probably waste good amount of time to configure blosxom to my taste.
Or I'd just leave it as is. Probably the latter.
The only reason I started using weblog software is to save maintenance cost.
It doesn't make sense if I spend too much time on configuration. One of the
reasons I chose bloxsom is it looks like very easy to configure and it looks
like it will just work out of the box.
[]
permanent link
|