Building low-power servers has typically meant using laptop CPUs. The main issue is that these cost a lot more than comparable desktop CPUs. At the moment, a 25W dual-core mobile CPU (Intel, Socket P) costs around 200 Euros (e.g., an Intel Core 2 Duo P8800 Penryn 2.66GHz 3MB L2 Socket P 25W costs $260 at Newegg and 222 Euros at Mindfactory). This is more than times as much as a similarly speced desktop CPU (e.g., Intel Dual-Core E5400 Wolfdale 2.7GHz 2MB L2 Socket LGA 775 65W costs $70 or 55 Euros, respectively).

This is not the only cost difference: since the processors use different sockets, different motherboards are needed. Currently, Newegg only has 2 Socket P motherboards. They cost $120 and $160, respectively. This is again a factor of 3 price difference relative to similar desktop equipment. Moreover, neither is very impressive--200 pin SO-DIMMs and a single PCI or PCI-E slot plus a mini PCI-E slot.

Another issue is that these "high performance" chips are typically not needed in a home server. In my experience, disk I/O is usually the performance bottleneck.

These three issues--power use, cost and performance--and the need to replace our MythTV & backup server motivated me to look at the Intel Atom offerings. The CPU and main board come bundled together and cost around $80-$100/65-80 Euros. The CPU has a TDP of around 10W. The main drawback is the mini ITX form factor, which means there is at most a single expansion slot. After some hunting, however, I came across the Point Of View MB-D510-MATX, which has 2 PCI slots, a PCI express slot, a mini PCI-E slot and 2 DDR2 memory slots. And, it cost just 80 Euros at PC-Pirat. FTW

I'd never heard of Point of View and there was not much info about the product online (even the manual wasn't online!). I decided to give it a go nevertheless, given the small monetary risk. I'm very pleased with the result and can recommend the product.

Public Service

The board requires 4 slots of space.

I scanned the manual.

GNU Hacker's Meeting Group Photo

GHM Group Photo by Neal H. Walfield is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unported License.

I saw the following sign last night in front of a Giant supermarket here in Baltimore:

PRIVATE PROPERTY

• NO PARKING

• NO SOLICITING

• NO LOITERING

• NO TRESPASSING

• NO DUMPING

• NO PETS

• HACKERS NOT ALLOWED

Hackers not allowed? Huh?

I saw Inglourious Basterds last night. It's a great film telling a neat story. The dialog is excellent and the violence rarely seemed gratuitous (in the sense that it added to the story and the character development, unlike, e.g., in an action film). I highly recommend it.

I submitted a paper to this year's Hot OS, 30 Years of Memory Mismanagement is Enough! Improving the Memory Residency Problem. It motivates better operating system support for adaptive applications in commodity operating systems. This is main idea behind the resource management framework in Viengoos.

I was reading the comScore report linked to by Slashdot regarding who paid for Radiohead's new album. The most interesting aspect of the report for me had nothing to do with the statistics or the interpretation of the data, but the origin of the data:

The results of the study are based on data obtained from comScore's worldwide database of 2 million people who have provided comScore with explicit permission to monitor their online behavior.

2 million people gave explicit permission to a company to spy on their on-line behavior in such an invasive way? I find it hard to believe.

Do our actions and decisions make a difference? That is, can we, who are not the Nelson Mandelas, the Martin Luther Kings, the Richard Stallmans and the others who have voiced their opinions against the establishment and altered their world, spur change. None of these were responsible for change. It was only the people who, in supporting these individuals' ideas, affected the change; it is only that history remembers these individuals' names for they become the figure heads and the embodiment of the resulting movement.

There is an irony here: how is it, that on the one hand, I claim we as individuals who support a movement make a difference but, on the other, the leaders of these movements made essentially none. It is the following: change is not affected by individuals working alone; chane is affected by the solidarity of individuals through their actions. An individual's action is only the cause of a change insofar as it is in solidarity with other actions.

I conclude from this the following: actions and ideas in isolation do not affect change. I, an individual, insofar as I desire change, can only affect change insofar as I support a movement, either as its figurehead or as a supporter and, with others, allow my actions to be guided and guide that movement. Any compromise which I make, provides immediate strength for the establishment. This does not mean that compromises cannot nor should not be made only that a cost benefit analysis must be performed and only if the result works in favor of the change can I be said to have helped affect that change.

On May 25, 2005, Nokia released a press release announcing the Nokia 770 and promising an open development platform and collaboration with the FLOSS community: "[t]he maemo development platform (www.maemo.org) will provide Open Source developers and innovation houses with the tools and opportunities to collaborate with Nokia on future devices and OS releases in the Internet Tablet category." On June 7, 2005, the Linux Journal interviewed Ari Jaaksi, the head of Nokia's open source software operations. He asserted that "[i]t is important that Linux for the 770 is not controlled by any company."

These remarks have proven deceiving: the Maemo platform appears open because it uses many FLOSS components and many FLOSS applications run on it, however, the 770 distribution contains a number of essential closed source components controlled exclusively by Nokia. This prevents a free platform from appearing for two reasons: first, the closed components are not easily replaced with free components; and second, it would cause a split in the community. I contend that the result is that it is more difficult for a free platform to appear for the 770 than it otherwise would be for a similar proprietary device.

Given market pressures, it is understandable that Nokia relicenses a number of closed source value added components that have no free counterparts, e.g., their support for Real Audio and Flash. Yet, this is not the extent of the proprietary software which comprises the platform. Jaaksi notes that they "keep some limited parts of the software that are very close to [their] hardware close (sic). Examples of such components are the boot loader and battery charging." Additionally, "[v]arious application user interfaces that provide the Nokia user experience are closed source." By retaining control of these components, Nokia provides a technical barrier to competing platforms.

Jaaksi provides the key to preventing similar power-grabs in the future:

It is important that certain components are licensed under an open source license, such as LGPL, that allows us to integrate also (sic) proprietary components on our platform.

Thus high level proprietary components can be integrated in a FLOSS-based stack. As long as the stack allows such proprietary components, it will be difficult to prevent such selfish behavior. If libraries that implement new interfaces were instead licenced under the terms of the GPL rather than the LGPL, proprietary vendors would not be able to gain value from the work of the FLOSS community without also contributing their enhancements back to the community. Stallman further explains when not to use the LGPL.

The disadvantages of proprietary software are, however, clear to Jaaksi:

We often requested bug fixes or modifications to the commercial closed components on our platform. If the vendors didn't have the capacity or will to fix the problem on time, we had few options. We could not fix problems ourselves because the companies using closed source didn't want us to access their source code.

Yet, the same critique has been systematically leveraged against Nokia: this past April, May, August here and here, September and October. But a stronger critique can be made as well: because Nokia, not the community, controls the distribution and has made it difficult to create a competing one, they do not have to listen to the community: their interests prevail and, as Murray Cumming has observed (local), community members have a difficult time getting even simple and correct patches integrated when these do not directly forward Nokia's interests. Nokia clearly understands this frustration and wishes to avoid it but they do not want to afford the same independence to those who use their product; they want to impose restrictions on their users which they do not want imposed on themselves.

Nokia's openness is hypocritical. It may, perhaps, be tolerated if their contributions prove, on the whole, to be an asset to the FLOSS community.

Nokia has contributed to the FLOSS community. They have hired developers to not only adapt mainstream components to their environment but to also improve them for the general community, e.g., Matchbox, D-BUS and Scratchbox.

As the device has been largely welcomed by the FLOSS community, unlike, e.g., the Zaurus, they have demonstrated a desire for such a platform. Given the rejection of the Zaurus, this may also suggest that Nokia has provided the leadership and resources required to realize such a platform which the community otherwise could not have. On the other hand, the 770 and n800 are significantly more affordable and more readily available than the Zaurus. Given that there are three non-general purpose platforms which the FLOSS community has developed which have exceeded their respective vendor's distribution, OpenWRT for Linksys' WRT54G, RockBox for a growing number of portable audio players and Slug for Linksys's NSLU2 network storage appliance, it seems plausible that Nokia simply developed the right hardware for the right price at the right time, i.e., they found a market niche.

If the platform is as open as Nokia claims and if it is correct that the community to develop a platform independently exists, then if Nokia abandons the device, it should be possible for the community to continue its maintenance as has happened for so many other devices. This litmus test has recently been started: Nokia announced the end of life for the 770 only a year and a half after its introduction. Yet the question has been asked: given the amount of proprietary dependencies how far will the community be able to continue developing the device? It is unlikely that we will find out, however, as Nokia has provided approximately 500 developers with n800 discount codes to ensure that the community adopts the new device.

Nokia should be thanked for their contributions to the FLOSS community, however, I contend that Nokia's half-openness presents a greater liability to the FLOSS community than a typical closed platform. Like the Sirens, the song of the 770 and n800 has allure: they support free software applications and their platform appears to be based on free software. Yet, they are not free software and the community has little influence over its development. Indeed, continuing development after abandonment appears problematic as Nokia does not support the four freedoms.

Nokia has put on a likeable mask, yet they wish to continue to play by the same rules that proprietary companies play by and retain control of their platform. If Nokia continues to ignore the desires of the community, sufficient momentum will eventually build and their platform will be abandoned and their influence made redundant. If Nokia wishes to remain a member of the FLOSS community in the long run, they should remake themselves as an authentic community member. In particular, they should release their proprietary software under a free software license and provide community members more control over the development of the platform.

According to a BBC article published today, Bush appeals for Congress unity:

Bush called on Democrats "to set aside politics and look to the future".

What does this mean? The Collaborative International Dictionary of English defines Politics as:

1. The science of government; that part of ethics which has
   to do with the regulation and government of a nation or
   state, the preservation of its safety, peace, and
   prosperity, the defense of its existence and rights
   against foreign control or conquest, the augmentation of
   its strength and resources, and the protection of its
   citizens in their rights, with the preservation and
   improvement of their morals.
   [1913 Webster]

2. The management of a political party; the conduct and
   contests of parties with reference to political measures
   or the administration of public affairs; the advancement
   of candidates to office; in a bad sense, artful or
   dishonest management to secure the success of political
   candidates or parties; political trickery.
   [1913 Webster]

So what Bush seems to be saying is that the Democrats should put aside governing for the sake of the future.

Aldi-Süd is selling a Strommessgerät (electricity monitor) this week. This is something that I have wanted for a while and for 9.90EUR, we invested.

When on and otherwise idle, my ThinkPad X40 consumes 25 Watts. Turning the LCD off, it uses just 18 Watts. Our DSL modem from Deutsche-Telekom requires 15 Watts. That is, it is using approximately as much power as the laptop when the screen is off.

The most interesting results are how much electricity devices use when they are connected but turned-off. My ThinkPad power-brick without the laptop connected uses 6 Watts. The 19" LCD uses 6-9 Watts. The biggest offender was our printer: when an HP OfficeJet 6510 is plugged in but off, it consumes 11 Watts. This is the same amount that it uses when it is on and idle.

I recently started using VoIP. My first experience was with Ekiga using a cheap handset, which I bought at Conrad for 20 Euros. I choose DiamondCard.us as my VoIP gateway, it happened to be directly supported by Ekiga.

The quality of the calls made using the VoIP gateway was lower than that which I was used to over a normal telephone line, however, it was acceptable. The reason for the quality degradation was difficult to diagnose given the large number of variables (handset, connection, provider) so I did not investigate very much.

Isabel then changed telephone and Internet providers--from T-Com to Alice. Although Alice is cheaper, the real deciding factor was the lack of a contract: it is possible to cancel the connection with 4 weeks of notice; to get a good deal with T-Com, you have to sign a 24 month contract.

One of the disadvantages of the Alice is that it is no longer possible to use the cheap prefixes (e.g., those at http://www.billiger-telefonieren.de/). Since I call the US a fair amount, it was pretty important to find some solution. We decided to invest in a Linksys SPA-3102, which allows the transparent use of VoIP over a normal telephone handset in conjunction with a traditional POTS connection.

The Linksys come recommended, but with the caveat that it was hard to set up. The solution that I ended up using was to route all calls over the VoIP gateway: in our case, this turns out to be cheaper for every type of call. I do not want to say that this was trivial to set up, however, it was really quite easy. Developing even a simple dial plan (to route calls or rewrite numbers), however, looks like quite a project.

The quality is really excellent and I conclude that my initial poor experiences were due to the cheap handset. I can highly recommend a similar set up.