I’m interested in exploring more ways to improve cross-distribution coordination, specifically as it relates to the Debian Sugar packages. I’m working to get all of the Ubuntu-specific Sugar packages included in Debian, which will probably be a summer-long effort.

wmf-01 “le premier”

2x Dual Core AMD Opteron(tm) Processor 285 @ 2606.342 MHz
2x 250 GB HDDs, 2x slots empty
2 Gigabit Ethernet NICs

This machine worked swimmingly.

wmf-02 “something witty”

2x Dual Core AMD Opteron(tm) Processor 265 @ 1800.000 MHz
2x 250 GB HDDs, 2x slots empty
2 Gigabit Ethernet NICs

This machine was incredibly noisy when turned on.

wmf-03 “lemon”

2x Dual Core AMD Opteron(tm) Processor 265 @ 1800.000 MHz
2x 250 GB HDDs, 2x slots empty
2 Gigabit Ethernet NICs

This machine did not fully POST, and was incredibly noisy when turned on.

Between them, only one of them had working fans. The other two made ungodly noises. We managed to salvage enough fans from the machine that didn’t post so that we now have two working machines cooling-wise.

We hope to install these machines at a Virginia co-lo center after we finish getting all the parts for Ivan Krstić‘s blackrock.

NB: This post has been sitting around in my drafts for a while, and I just got around to posting it now. We’re still waiting on some last-minute parts before putting these serves into production.

(11:28:29 AM) Luke Faraone: don’t you hate it when you think of something cool, only to find that someone else already thought of it?
(11:28:47 AM) Peter Harkins: Depends. Sometimes I then think “Awesome, now I don’t have to spend all that time building it.”
(11:29:34 AM) Luke Faraone: I recently was thinking “it’d be cool to be able to have a smaller ‘little computer’ with a NIC, some RAM, and a low-powered CPU to maintain presence on IRC etc when my computer’s sleeping.” Then I saw http://it.slashdot.org/story/10/06/13/0641228/Microsofts-Sleep-Proxy-Lowers-PC-Energy-Use
(11:29:39 AM) Luke Faraone: … and it’s from MSFT.
(11:30:12 AM) Peter Harkins: cute
(11:30:25 AM) Peter Harkins: There are lots of tiny Linux pc’s out there, though.
(11:30:45 AM) Peter Harkins: I’ve seen a couple the size of a power brick – you plug them in, add ethernet, done.
(11:31:14 AM) Luke Faraone: what’d be really cool is if one could author an API that would allow for desktop applications to request access to run services on the device, and have state magically transfer across them.
(11:31:40 AM) Peter Harkins: I’ve seen people talking about doing that – I wouldn’t be surprised to see it commonly in 5y.
(11:31:49 AM) Peter Harkins: It’s sort of the logical extension of GNU screen.
(11:32:10 AM) Luke Faraone: we have live migration of VMs in the enterprise market.
(11:32:34 AM) Luke Faraone: if the wall wart had hypervisor support, you could just operate each service in a sort of sandbox.

I know that in order to make it work in reality, we’d need support from app developers, but are there any technical reasons this won’t work?

After reading several articles about the alleged spying that was enabled by a Pennsylvania school district via its one-to-one MacBook, and seeing discussion on a variety of mailing lists, I’ve decided to implement my own zero-cost, no-hassle solution to the problem for the OLPC XO-1‘s camera.

This should be able to be adopted in deployments everywhere, by anyone with a piece of paper, or anything else they can slide through the plastic faceplate.

Just say no to fancy addons and factory-added “shutters” or “covers”, make your own!

The GTK+ version of the package is published in Ubuntu 9.10 Karmic. While upstream is continuing to do regular releases of the GTK version, they are focusing on the KDE version and have changed autokey to refer to the KDE version, while renaming the GTK version (formally autokey) to autokey-gtk. To make matters worse, upstream releases both as separate tarballs, and the packages conflict with one another.  (due to technical limitations)

What’s the proper proceedure for handling this in Ubuntu? Should -gtk conflict with autokey, replace it, with  autokey-qt being available as an option, or should I just keep things as they are and have the package “change out from under” users when they upgrade?

This really burns me up!? What kind of craptacular beta is this? Do you kiss your mom with this kind of code? You developers will be the first up against the wall when the revolution comes!?

Chrome Beta: password sync failed with horrible javascripty error message. [via Google]

• The XO-1s were shipped with Open Firmware‘s security enabled. This caused them to have the same anti-theft protection as laptops deployed in the third world, without any of the benefits of a remote killswitch or tracking
• The XO-1s had a version of OFW which would fail to boot when the clock was below a certain value

The above two issues combined with the manufacturing fault was a recipe for disaster. Owners who discovered this in the first 30 days were able to get a RMA and a working laptop, but OLPC lacked the resources to support those outside of this minimal warranty. I’ve been running an OLPC repair center, OLPC DC Repair, (charging only minimal fees for labor and shipping) since mid-2008, and have handled dozens of these “unbricking” problems.

The procedure for repairing the above is straightforward and well documented. However, it can be tedious, especially for those who are unfamiliar with the tools involved. Since I needed the programming practice anyway, I decided to write a rudimentary Python script to automate the process.

Thus enters d6.py. d6.py makes it (hopefully) amazingly simple to unbrick your XO, so that you can get up and running as soon as possible. You can clone the git repo, or download it directly (permalink). To download and run in a single command:

python -c "import urllib2; exec urllib2.urlopen('http://dev.laptop.org/git/activities/olpc-contrib/tree/d6.py').read();

Plug in your OLPC Serial Adapter (or one of the compatible alternatives), and run the script as a user which has access to /dev/ttyUSB0 (or as root, not recommended) or change the path inside the script to something suitable to your system. This script is in the alpha state, is poorly documented, and may not handle all edge cases (read: other people’s systems) well. I’m not responsible if it kills your cat, lights your XO on fire, or makes your wife leave you, but hopefully it’ll be of some use. Expect a GUI shortly.

Limitations:

• Does not handle all error conditions
• No command line params
• Hard-coded path to serial adapter

Therefore, the code does not run on other platforms other than Linux. You might also encounter problems if you’re running it on a system with brltty installed, removing it should fix the conflict.

Feedback is more than welcome in the comments.

The primary purpose of attending school is to learn new ideas, concepts, and methodologies. Ideally one should retain everything one learns in the classroom, so that one may do well in the short term, eg. on an exam, as well in the long term, in college and beyond. Unfortunately, there are a number of constraints on a person’s time and energy, such as after school activities and biological necessities. Therefore it is highly desirable if there are means by which we can increase our information retention with a minimal increase in the amount of time or effort spent learning.

An example of such tool is the Spacing Effect. It essentially states that one has a higher level of retention if one spaces periods of study over a longer period of time, contrasted with “cramming” everything in one go. Even if the total amount of time spent studying is the same, one will remember more if said time is spread across a wider period. While it is very easy to procrastinate, it is important to pace oneself and not “leave it all ’till the last minute” before an exam, essay, or other assessment. While one may still pass the assessment, they will probably not exhibit long-term recall, and will have to mostly relearn the material when it comes time for a final exam or standardized test.

Another principle that affects memory is the Serial Positioning Effect. If one is memorizing a list of items, one is more likely to remember the items that are at the beginning or end of the list, compared with those in the middle. To combat this, one should vary one’s routine. If, rather than studying each subject in the same order each night, one instead studies those same subjects in either a random or constantly different order, one can maximize recall of all items. Also, if one had a list of vocabulary words or something similar to remember, one would benefit by varying the sequence in which they review the items.

Often one needs to remember a list of items in order or in relation to one another. For this, one may use the Peg-word System. In this system, one assigns concrete names to numbers, such as “one is a bun”, “two is a shoe”, etc. Through this method one can construct narratives incorporating the peg-word for the item number as well as the item itself. When one needs to remember the list, they simply mentally run through the narrative. While in school remembering a list in order is not as important as simply remembering the contents of the same, this method could be effectively applied there.

Compared with other forms of memory encoding, Semantic encoding provides superior recall. When information has a semantic meaning, when one is able link a bit of information to previously studied concepts and ideas, one better understands the material to be learned. Moreover, one is more likely to remember such information in the future. Therefore, effort should be made to find meaning in what is to be learned, rather than simply memorizing the words and sentences.

Finally, people have excellent recall for information that they are able to relate to their own lives. The self-reference effect, as it’s called, means that one will be more likely to remember something if it is relevant to them, and specifically if it is directly related to them. For example, if one is asked how well an adjective affects others, they will be less likely to recall the question being asked than if they were asked if that same adjective affects themselves. This could be effectively applied in school by attempting to relate or link concepts with past experiences and personal attributes.

While assisting with the various presentations at the Open Source Center and staffing the Sugar/OLPC booth, I ran into some of the folks from NComputing. Their corporation has some similiar goals with that of OLPC, as both involve low-cost computing for the third world and elsewhere. Providing multi-seat technology, which is similar to thin-clients without the lag and network overhead, they enable multiple displays and mice to function off a single computer. Since they support both Ubuntu and SuSE Linux with their (admittedly closed) hardware, I decided to investigate their technology as a means of enhancing deployment of Sugar.

The method to activate the NComputing software (requires registration to download) is not obvious; one must navigate to the console, select “Serial Numbers” from the side menu, and then right click the empty license list and choose “Manage”. Many users would not be able to do so without the manual; maybe it would be easier to prompt for a license key in the debconf install process?

Unfortunately, I wasn’t able to test their software with Sugar; the most recent version of Ubuntu they support is 8.04, and I can’t even insmod their kernel module on 9.04.

aside: meant to put this out an age ago, just got around to hitting “publish” today. (2009-08-16)

We’re all too smart to use the same password on multiple sites, right?

While most people, I included, cannot say “yes” to the first question (at least not for everything), that alone is not enough. This is because no matter how secure your password is, be it 20 letters long with various dingbats and 中文 characters, there is a weak link in this system. Or rather, two:

Your email is an obvious vulnerability: if someone was able to gain access to that, it would be trivial to reset your password for Facebook, YouTube, Meebo, etc.

What if you have a strong email password, you ask? In that case, we get on to the heart of the matter (which is also the most relevant to all those social networking users out there): secret questions.

Secret Q&As (SQAs) were initially a good idea: provide an alternative in case one has lost access to one’s email, or never set one in the first place. (as with Gmail or Yahoo) It presents an interesting problem, however: while the average netizine is unlikely to know the mother’s maiden name of dogggzlover98382374@hotmail.com, if even your name can be figured from your email address (or the attacker knows you personally), it is trivial to use sites such as Facebook and MySpace to find the answers to SQAs. A rather public example of this vulnerability can be seen when Sarah Palin’s email account was broken into last summer: all of the information needed could be found out using public records.

An example of a possible attack against Facebook in particular:

• Gain access to someone’s profile by either friend-requesting outright them or by masquerading as someone they know (and don’t already despise)
• Look for an email address on the profile or in wall posts.
• Visit their email provider and reset their email password via the information in their profile.
• Now reset their Facebook password. This will send an email to their address, which you already have access to.

This works against any site that uses a email-loop, even if it is well designed to avoid common SQAs. Social networking sites, however, are particularly vulnerable because of the wealth of personal information one shares freely on them.

This is because, as they are part of your personal history and not transactional, SQAs are almost always the same between sites. So, if you’re truly concerned about your information security: use something random for your SQAs and store them in a safe place.