Older blog entries for rcastro (starting at number 20)

It took me a long time to decide, but I ended up quitting my job (will work until the 19th) and will work on some projects. Basically I want to work on technical projects, namely Linux Kernel and JBoss application server, and on entrepreneurship with some projects I have in mind. I am very eager to start this new stage in my life.

Gaim

This weekend I worked on a patch for Gaim 1.0.2 to make the away menu (ie, status menu) aware of the current status of each protocol. Without this patch, I was unable to know what was my current status in Yahoo or ICQ. I don't think it is going to be accepted (because I worked on the stable branch), but I tried and it was a good experience. It is available in my home page and in the patch section of Gaim project page. If the Gaim developers tell me that I could help in the development branch (what I volunteered for), I will port these changes to it.

Apache

In my last days at Convergys, today I started installed Apache Jetspeed 2 Enterprise Portal, but didn't have time to finish. I had the chance to install Maven (I am curious about it) and Tomcat (never installed it without an application server, such as JBoss. Besides all these things, I would like to create a first web service next week in order to learn about Apache Axis.

Linux Kernel

Now it is time to work on some Linux kernel ideas. When they are mature enough, I will publish about them here.

In these last days, I followed a discussion thread between Marcelo Tosatti and Andrea Arcangeli about out-of-memory killer in Linux Kernel. Very interesting and very intriguing. I intend to do some work on this as soon as I start work intensively on the Linux kernel.

Skype?

I talked to Livio, a friend of mine, about Skype and he called my attention to some interesting points about it, that I will share with you. Have you ever checked its license agreement?

**********************************************************************
"Skype End User License Agreement"
[...]
Article 4. Permission to Utilize

 
 4.1 Permission to utilize Your computer. In order to receive the
  benefits provided by the Skype Software, you hereby grant permission
  for the Skype Software to utilize the processor and bandwidth of
  Your computer for the limited purpose of facilitating the
  communication between You and other Skype Software users.

 
 4.2 Protection of Your computer (resources). You understand that the
  Skype Software will use its commercially reasonable efforts to
  protect the privacy and integrity of Your computer resources and
  Your communication, however, You acknowledge and agree that Skype
  cannot give any warranties in this respect.
[...]
*********************************************************************

Besides that, Damien Sandras, from Gnome Meeting, has insightful thoughts about it, that he published in the project web site:

    People are often asking me what I think of Skype. What I think is
    pretty clear...

 
     The main problem is not that the program is not Open Source, the
    problem is that Skype is locking users into a proprietary
    protocol. Would you imagine the Internet with a proprietary
    equivalent to the HTTP protocol that only a given client could
    browse? That's what happens with Skype. Skype also has a great
    marketing force, some people even think that Skype has a superior
    audio quality. How could Skype have a superior quality when it is
    using the same codec (iLBC) than software like GnomeMeeting while
    introducing more latency by making calls go through a 3rd party?
    The only real advantage of Skype is that it is easily going
    through any type of NAT, using a 3rd user to proxy the call. But
    the day when the Linux kernel NAT will natively support H.323 or
    SIP, Skype will have lost its only advantage... Skype is hype...

More than one year later, I finally add an entry to this diary. I am back to the Linux kernel world and I am very glad about that. I will manage to have time to try to solve some challenges in this world and I am already working on some new ideas. So far, the port of compressed caching to Linux 2.6 is not on the top of my list because I want to start working on some tasks that won't take that much of my time. Then, when I notice I will have enough time and knowledge of Linux 2.6, I will get back to compressed caching.

Today I subscribed to Linux kernel and Linux-MM mailing lists, downloaded 2.6.8 and already read a paper about the new idea I mentioned above. Now it is time to read the code and understand some recent discussions I read in the mailing list archives.

The paper I published last year about compressed caching is avaliable on my site.

Skype

Today I tested Skype in Linux and it worked perfectly. I was amazed with its sound quality and with the sound quality from a mic that came with a Sound Blaster Live! card that I bought in Canada.

GnomeMeeting

Check the text from the GnomeMeeting site about Skype:

"The main problem is not that the program is not Open Source, the problem is that Skype is locking users into a proprietary protocol. Would you imagine the Internet with a proprietary equivalent to the HTTP protocol that only a given client could browse? That's what happens with Skype. Skype also has a great marketing force, some people even think that Skype has a superior audio quality. How could Skype have a superior quality when it is using the same codec (iLBC) than software like GnomeMeeting while introducing more latency by making calls go through a 3rd party? The only real advantage of Skype is that it is easily going through any type of NAT, using a 3rd user to proxy the call. But the day when the Linux kernel NAT will natively support H.323 or SIP, Skype will have lost its only advantage... Skype is hype..."

This text was mentioned by my friend Livio.

Packet-Writing

Tried quickly to add support for packet writing, but the patch to add this feature hasn't been integrated yet, and there is only for 2.6.0-test1.

Here you can find the kernel patches. Basically, you add the packet writing support in block devices section, and udf support in filesystems section, then you use cdrwtool and udftools to handle you CDRW.

2.6 VM

Yesterday I studied mm/vmscan.c for a while and it doesn't look too hard to port my code to it, in particular because I had already studied rmap implementation for a bit, but it was only a quick glance. Now I am going to take a look at it briefly again, since I am very tired today. Hopefully tomorrow or Thursday night I can starting adding my code to this kernel tree.

Looking forward to 2.7: compressed caching?

I was very glad to know that CC has been mentioned when talking about 2.7 in Linux Weekly News. Check it out:

http://lwn.net/Articles/37815/

I am working in my spare time to make compressed caching a possible candidate for integration.

My paper for the 15th Symposium on Computer Architecture and High Performance Computing (SBAC - PAD 2003) was accepted, and I sent the final version this Sunday. I am very happy that it got in.

Now I am finally studying Linux 2.6.0 VM (now checking vmscan.c, which is the main part of the VM for CC purposes). I intend to port my code to this version, although it may take a while if many things turned out to be too different from 2.4. First, let me make compressed caching working for swap pages, and later I will try to make page cache support work.

Besides that, I bought a Sony DSC-V1 digital camera. Very nice camera, I am enjoying my freetime in the weekends to take a lot of pictures with it.

On Thursday 10th, date of my last post, my grandmother died at 9h10 PM in the hospital. I didn't work on Friday due to the funeral and burial, and mainly because I was (and still am) very sad.

Since Thursday, it is very cold here in Sao Paulo. We have a dry winter, so, even though the mininum temperatures are not that low, we end up suffering a lot with them. And our houses are not prepared for winter (no heater and window sealing, for example), so it is a tough time. Fortunately we don't have many days like these ones in a year.

Now I am trying to clear out my mind to move on. First task, I have to scan many old photos I found out at my gradma's room.

Very interesting the GRAMPS program to build a family tree. It is a very good alternative to Windows program, although there are some points to improve. I build one from some data I had already collected, and made it available at here. I still want to get data like missing birth dates and, maybe, pictures from most living people.

New performance statistics are available on compressed caching website. They include:

• Linux kernel compilation
• MUMmer (scientific program for genome alignment)
• Open Source Database Benchmark
• piGIMP
• Matlab (fractal dimension computation)
• httperf

I will add some info about problems with Sort, PostMark and Contest that made us implement some features like clean page suspension, or even made us realize that there were some scheduling effect when using compressed caching. Probably by the weekend I will have it done on the site.

Compressed caching website updated (homepage). It has links to a paper in English that I submitted to a famous conference on operating systems and my master's dissertation, but in Portuguese. Now I can't delay publishing latest performance results from compressed caching, because it's the last step before getting back to real programming on the kernel code :-) Wait for news.

My personal website has been updated too, but only minor updates, like showing that I finished my master's degree and my current job as development analyst at Convergys. Most of website is in Portuguese, but you can download my curriculum vitae in English if you feel like knowing a little bit of my life.

Yuppie, I am finally free to work on the project. I finished the final version of my dissertation and now my spare time will be actually mostly spent working on the compressed caching project :-))

Now it is time to update the performance statistics available on the project web site and to make my dissertation available on my web site. Oh, and I can't forget to upload of my papers to the web site, it may be worthwhile for people interested in more information.

On Friday I submitted the paper to the Symposium on Computer Architecture and High Performance Computing (SBAC). Now fingers crossed (notifications by July 28th).

I checked the types of available diplomas I can get, and the free one is done with regular paper. There is a special version, which is done with sheep's skin, but it is expensive. I am still thinking which one I will pick.

Nothing very interesting so far. I couldn't work on any project, but I am almost done with what is hindering from that.

The guilty: master dissertation. But it is almost done. I waiting for some comments from my advisor, but I don't think there will be much work to do (if actually any). One reason to not have much work is that I have to deliver the final version of it this week.

Today I also had to fill out a form from a federal foundation (CAPES), which I intend to deliver on Tuesday. But the most important feeling this time: freedom by the end of this week! :-)

SBAC paper is done. Now waiting for a review, but I don't think I will spend much more time on it. Thanks God!

Only the minor changes must be implemented for the final version of my dissertation. This time I have to fix the errors found and pointed out by other member of the committee. Well, I actually just tried to fix these errors and in 15 minutes it was done. Very quick. Only some suggestions left from the first committee member and we are done with this dissertation. Next week I will be finally free from my latex programming job, and will be able to code again.

Tomorrow I guess I will work on XSL stuff at my job. In particular, Xalan and Saxon implementations of XSLT specifications. I hope it will be fun.