« October 2004 | Main | December 2004 »

November 19, 2004

darcs on Mac OS X

In case you want to try darcs with Mac OS X, these links are usefull:

Have the appropriate amount of fun.

Update: fixed the URL's, Markdown was messing them up.

November 17, 2004

skype for mac

I installed the skype client for Mac some weeks ago, as soon as it was released. I've been letting it running in the background since then, and I used it 5 times at most.

I was interested because I would like to have a cross platform voice solution. My wife used Windows XP, and I would like to talk to her with VoIP as I'm away from home three to four days each week.

My problem with skype is that you can never be sure the kind of performance you'll get. Most of the time, my skype connection goes to the US before reaching the other side. We are both in the same provider, in Portugal, so this seems a bit stupid.

I know that the peer-to-peer design of skype is the root of this problems, but that makes skype inadequate for my purpose. I have tried the AIM 5.5 client (I use iChat AV everyday) but couldn't get it to work. I think I'll have to try harder.

In terms of voice quality, if you get a good peer, it's very good. But you cannot trust it to perform always at that level, due to it's design. In that respect, the iChat AV/AIM client is better. The quality is not as good, but it's reliable (when tested between iChat's...). And it supports video also, allowing me to put my iSight camera to good use.

I also tried skype out with some friends in england. They where having a conference call: 2 guys with skype, calling Pedro via mobile phone, and me via fixed phone. It did work, the quality was acceptable, but it was very confusing.

Let's wait and see if the final version is better, but the peer-to-peer design doesn't seem a good one to me.

Next powerbook upgrade

So my trusted old (30 months today) powerbook is still kicking. I never had a laptop that lasted this long. I'm a 15'' TiBook 800Mhz owner, with 1Gb RAM and 40Gb disk.

I've been thinking my requirements for my next laptop, what are the things I value the most. This will allow me to choose my next one in the next 6 months or so. I'm expecting some annoucements from Apple in January, or the WWDC, in June at the latest.

The list is in order of importance, most important things to me at the top: * screen real-estate: pixels, pixels, pixels... * battery life; * processor type.

Let me start with the things that are not on the list. Airport Extreme and built-in bluetooth are a must-have. There is no way I'm ever getting another laptop without them. My current TiBook has Airport (the "normal" kind), and it rocks (although TiBooks suck in terms of signal strength). And I'm tired of plugin the bluetooth dongle. Also, the backlight keyboard is at most a nice-to-have. I can keep on using my USB-powered thingie for light.

The current minimum disk drive is 40Gb and I'm happy with that. If it's bigger, great. If not, that's enough for me, for now. I always carry a external 40Gb Firelite, and I have a 200Gb Maxtor OneTouch at home for weekly backups.

I did not list the superdrive also. In two and a half years, I think I burned 10 CDs at most, and some of them where not for me. I don't use DVD or CD for backups, and my line of work does not require a constant use of a burner. A superdrive would be cool to make some home projects with the kids photos, or something like that, but it's not clear to me if the price difference is really worth it.

Now let's see the list and each point in turn.

The first point is a tricky issue for me: I think that the 17" PowerBook is very very big, but I haven't ruled out the possibility of buying one. I still have to check if it fits my current backpack. It's a great backpack and changing it simple to carry the 17" is not an option. I prefer to keep the backpack and buy a 15". But I can safely say that 15" 1280x854 is the minimum resolution for me.

Battery life is more important than raw CPU power. Apart from games and the occasional compile, I've been working with a two-generations old 800Mhz G4 and I can keep on doing it. I would prefer to have a 1.5Ghz G4 with a 6/7 hour battery life than a 2/3 hour G5. I think that if the difference is 1 or 2 hours between the two, I would probably go with the fastest one, but more than that, I don't think so.

My most used applications are: Terminal.app, Mail.app, Safari, iTunes and NetNewsWire. I also use Quicksilver, iCal and AdiumX a lot. Everything else is used much much less. With regard to games, I play a bit of Massive Assault, and I would like to play Doom3. I also have Apache, MySQL and Jabberd running in the background. I usually have 5 to 9 applications going, so multi-tasking is a must. If you ignore Doom3, I could buy a fast G4 and be happy. Even better, a dual-core G4 would be great. The only reason to buy a G5 is Doom3 right now, and unless they appear in the first half of 2005, I'm not going to wait: if Apple presents a dual-core G4, I'm sold. You can always buy a cheap PC or a iMac to run Doom3.

The CPU question is tricky only if you think that you'll keep your laptop for the next 2 to 3 years. In that case, buying the fastest CPU you can afford makes a lot of sense, and waiting for the G5 might be a good idea. Yet, I don't expect my application profile to change in the next years. My laptop is my main machine. I work in a lot of different places and the laptop must be able to do it all. So a dual-core G4 is the most likely step.

Of course, it all depends on Apple. If they launch dual-core G4 and a G5 at the same time, I really don't know :)...

So, I think I can summarize like this: I'll probably buy a 15 or 17" dual-core G4, or a G5 if they appear in the first half of 2005.

Team Virus

With all the infected zombie PCs out there, and with most virus being written having the support of plugins, I wonder why is it taking so long to create a plugin that would install the SETI client, or a cancer research distributed client, and started putting those computers to a good use?

Go, Team Virus!

Voiding your iPod waranty

I just throw away my first generation iPod warranty. I installed a NewerTech battery, the 2100mha kind.

So far it's looking good. I'll have to test it in the next couple of days, but the battery came with a little charge and I could listen to my music again.

This is still a 10Gb model, but it's enough for me for now. And it prevents me from buying a newer model.

The instructions seem straight forward, but it took me some tries to open the damn thing. In the end is a very simple procedure, and it takes 5 minutes tops to change it. The non-scratching tools that came with the battery work fine and do the trick.

If you have a aging iPod, and don't want to spend hundreds of Euro in a new one, this is a worthy upgrade, and this will give me another 2 years on my 30 months old iPod, purchased May 26th 2002.

If you need one, I ordered mine from ipodworld.co.uk. Recommended. Me and Rui tried to order this from US, but it was complicated.

November 13, 2004

Evaluating source control systems

Which source control system do you use? In a open-source world the usual suspect is CVS. For some years now, CVS has been showing it's age, so a bunch of replacements have appeared: Subversion, Arch, Darcs, to name just a few.

I still use CVS daily, with large projects, and I know it well. Well enough to get around most of the limitations it has. But I have decided to switch systems some time ago. The question is: which one?

Subversion has a lot of momentum right now. Perl is now maintained in Subversion, Mono switched today, and most of the Jabber stuff is also Subversion. The switch from CVS is simple, they have similar command line interfaces.

Arch is very well desingned but the learning curve is very steep. I tried a coupple of time already, and I always give up because I can remmember wich command I need. I'll have to look into some of the arch frontends. It's very different from CVS, but the offline capabilities are extremely good.

Darcs seems very good also, but I haven't had time to test it. That's my next step.

The three things I look in these systems are: * simple things should be simple: commit changes, push changes to remote repository, pull changes from remote repository, see differences between my copy and remote rep, see my changes since last commit, add, mode, delete files; * complex things should be possible: keep several branches of a project, merge branches. * They need to support offline operation: you must be able to commit, update, diff your work to a local repository and then sync when you get to the office.

There are also some nice-to-have things, like commit triggers. I use commit triggers to generate RSS feeds of commits.

While evaluating these systems, I developed a small test to see if it's good enough for me: * create two branches of a project, called B1 and B2; * modify file X in B1; * in B2, rename file X to Y; * merge B1 into B2.

If after these simple steps, your Y file does not have the modifications made in B1, the source control system is flawed, and I just can't use it. If you are serious about source control, this is just not acceptable.

The interesting part is that it seems Subversion fails this simple test. I think I'm doing something wrong, I'm trying to see what, but I can't get Subversion to do the right thing. This would remove it from my list, and I would like to use Subversion (JabberStudio uses it a lot now).

I'm going to try Darcs now, re-check Subversion, but if those two fail, I'll have to find a decent frontend to Arch.

Some references: * CVS * Subversion * Arch * Darcs * Version Control System Comparison * Thoughts on Version Control * Some notes about source control systems * JabberStudio

November 12, 2004

RPC and Perl

During the last couple of years, the idea of a simple to use RPC system for Perl has been on my mind. I've done several designs, mostly focused in object oriented perl, but never did one I really like.

It's been a challenge that I pick up whenever I can.

Today, Ricardo pointed me to mod_perlservice, a very simple RPC system for Perl. I'm still looking into it, it seems to be function-oriented, not object-oriented, but it also seems extremely simple to use. As a added bonus, it has clients in C and Flash out of the box.

I'll have to do something with it to test it out, don't know what yet.

If you want to follow the project, you can subscribe the freshmeat entry for this project.

November 11, 2004

Good old hunter's package

Thanks to The Register for making me laugh at the end of a loooong day at work:

Agents searched Dugan's home and found an AK-47, a loaded rifle, a loaded shotgun, two handguns and 1800 rounds of ammunition. In Texas, that's known as "the weekend hunter's package."

See the whole article.

November 09, 2004

Loose yourself

First, let me just say this, as an advice: give your credit card to a trusted friend before entering any of these sites. Trust me.

Done that? Good.

Now enjoy: * Boys Toys * I want one of those * Gadgets

It's really nice to buy some christmas presents, though.

Jumping into the fire

Well, this is what I call a bad move, like jumping from the pan into the fire.

You just can't win.

November 08, 2004

Color Blind

Color me beautifull.

It has a happy ending: the perl guy gets rich and famous.

November 03, 2004

I need a G5

So finally a reason to buy a G5, at least a good one.

It seems I can wait until February 2005, but it's a fact of life now.