I read an interesting article in The Economist that describes the relationship between sleep and retaining information and new memories. It basically adds new momentum to believers of the power-snooze, something I practice on busy weekends.
The research was presented in this article from the Feb 25th edition.
The basic premise is that putting things into permanent memory requires a series of steps. These steps are performed during different phases of the first 90 to 100 minutes of stage 2 non-rem sleep. Normally this happens once a day when we sleep, however a Siesta enables an early execution of the process dramatically improving memory performance each day.
The memory management process for the human brain described in the article can be summarized as follows:
Store the memory in temporary storage (This is the act of actually experiencing what it is we are going to remember)
Move the memory into long term storage, creating space for new temporary memories
Associate the memory with other memories in long term storage, improving our ability to reflect back on the memory and associate it with other memories
The last two steps occur in the first 90 minutes when we sleep. The interesting observation or conclusion is that sleep actually does two important things.
Allows you to remember new factual things by clearing up temporary memory stroage space
Allows you to use your new memories more effectively by linking them in with other memories
By sleeping mid-day we actually create more room for new memories in the afternoon and establish permanent context for those memories captured during the morning.
If this is all true as suggested by the study, then it would seem pretty silly not to try to sleep after an important learning session, and possibly during breaks between research and writing any sort of document that is going to build upon your established knowledge and experience.
Immediately I think back to the number of times that lack of sleep (often due to celebration and “unloading”) has immediately followed a key learning (exams, sports championships, project go-lives, weddings, conferences, presentations) and the impact this may have had on my retention of experiences to date. There would seem to be a lifestyle correlation between this study, and one that I recently read suggesting that our ability to learn and retain information (outside of language related things) improves mid-life.
A second observation I made is that the behaviour is not unlike certain aspects of computer operating system memory management which has the concept of short term storage (RAM) and long term storage (Disk), and a process that manages availability of short term storage by writing out to long term storage during idle cycles to prevent the risk of data loss or poor performance from running out of short term memory.
The bottom line is that this is just one more reason to get more sleep. Perhaps if you have a snooze immediately following reading this article, you will actually remember why sleep is soo important. (Why they do not suggest this in the The Economist, I am not sure).
A smart professor might actually find a way to insert a siesta in between an important lecture and a following related lab session to ensure maximum benefit from this study.
Recently I received a forward citing a recent “John Hopkins update” on cancer. It turns out it is a Hoax (per the link below), in that it supposedly comes from Johns Hopkins, a well respected US health institution and it also contains non-factual statements mixed with some facts.
I encourage reading the hoax and the response which debunks the non-factual statements. If you have only time for one, stick to the facts in the response. There is lots of good information on cancer and links to related information. Links are at the bottom of the post.
Below are the two interesting things I learned about cancer after reading the Johns Hopkins response,
On Testing and Diagnosis
Diaz and other Kimmel Cancer Center researchers are working on new tests that detect abnormal DNA shed by cancer cells into blood and body fluids and have the ability to find cancers before they cause any symptoms.
Approaches like this could lead to a broad-based screening test for cancer.
Tests like these also are being used to detect cancer recurrences and malignant cells left behind following surgery, and can find cancers that are not detectable under the microscope or in x-rays.
Other researchers are studying cancer stem cells. They are stealth cells that make up just a tiny fraction of a tumor. While small in number, investigators believe they may be the cells that drive certain cancers and lead to cancer recurrence. Therapies that target these cells are now being tested in clinical trials.
A team of our breast cancer researchers has developed a method that could make it possible to detect breast cancer from the DNA contained in a single drop of blood.
On Immune Response
The immune system simply does not recognize cancer. In its complexity, the cancer cell has learned to disguise itself to the immune system as a normal cell. By deciphering the methods cancer cells use to make them invisible to the immune system, Elizabeth Jaffee and team have developed cancer vaccines that have successfully triggered immune reactions against prostate cancer, pancreatic cancer, leukemia, and multiple myeloma.
I had checked out the version of Novell’s moonlight linked through Microsoft from CTV’s site, in order to try to get some Olympic coverage on my PC. I obviously did not look hard enough. I stumbled upon the pre-release version just in time for the Super Sunday gold medal rematch series!
I am using the Karmic 9.10 Firefox 3.6 (You can see my APT sources selected in the post on 64bit Flash just prior to this post). It should work fine on 3.5 (per other posts) but did not on 3.7alpha if you are there. I had some trouble on an Intel i915 adapter, but it worked fine with fglrx on my ATI HD 3400 adapter.
Good to know Karmic users have something working out of the box using this Firefox moonlight add-on.
Enjoy the Olympics.. CTV just moved up in my personal ratings, though I hope they move to something Flash based, or more widely supported using stable platforms.
There is now a Ubuntu packaged 64bit version of Flash. This is something that any user running Ubuntu Linux on current hardware (P4/Athlon64 or newer) should be pleased to hear (just skip on over this post if you are running a 32bit OS on your 64bit hardware).
For quite some time there have been manual ways to go about installing cutting edge 64bit flash on Linux Desktop systems. Personally, rather than pollute my system with non-packaged cruft, I prefer to rely entirely on packaged goods when it comes to my applications. Knowing that a packaged version of 64bit Flash has been sitting in Debian Squeeze repos since 2007, and seen a lot of *exercise*, is in support of the stability now present in the 64bit version.
The process is easy if you have Ubuntu Tweak (I strongly recommend this for managing your PPAs and other external APT sources).
1. Select the 64bit Flash Plugin APT Source (this adds it to your /etc/apt/sources.list) and inserts the apt gpg key into your key store.
2. Select the 64bit Flash Plugin Application (this executes an apt-get install flashplugin64-installer) and inserts the apt gpg key into your key store.
Personally, I have experienced the need to *reload* pages with flash for quite some time and become used to the process. Any regular 64bit linux user running the 32bit plugin should be familiar with these issues. It is most annoying when showing my 21 month-old son clips of animals and trucks and such on You Tube, as he does not have the patience for the npviewer.bin crash that results in *waiting* and *re-loading* various YouTube pages. Now with the 64bit version, we can go right from garbage trucks to giraffes without any fear of a big delay.
As a side note, if you are running on an older Pentium 4 you may need to employ a workaround due to a missing LAHF instruction in your hardware. (I have a 2.9Mhz P4 with this issue) You will experience a crash on every flash page, and if you launch firefox from a terminal window the output will state: illegal instruction. If this applies to you, read this post to verify, and follow the instructions there.
Nearly a week into the Olympics, CTV unveiled their Olympic widget. The widget is put together by ClearSpring and can be embedded just about anywhere easily, including any Wordpress-based blog.
The process is automated if you have a Wordpress.com site by navigating the widget self-help wizard. For those using Wordpress on their own domain on a non-Wordpress server, this process breaks down when it presents a form for input of a wordpress.com username and account. Fear not, the process for any wordpress site, wordpress.com or otherwise homegrown is right here.
At the bottome right of the widget, click the “Grab It!” link on the widget
At the bottom right of the widget window that pops up, click “Copy Code”
In the Wordpress administrative panel navigate to Appearance and then to Widgets
Create a new Text widget by dragging it into your site layout and paste the CTV/ClearSpring widget code copied in step 2 into the text box for this new Wordpress text widget
Adjust the height and width parameters in the code to suit your placement if required
Great work by ClearSpring. The delay in CTV’s unveiling of the widget could be aligned with some *new* resources managing the web presence at CTV, otherwise I would have expected this long before now. Likely these new thinkers are getting a lot of praise and are clearly a side step from some of the less successful technology choices for the CTV site.
Specifically, it is refreshing to see some best practices in play after a bit of disappointment in CTV’s online implementation of streaming Olympic content (based on the SilverLight platform) which unfortunately meant no coverage for us here at the house and any of my peers and engineering team at the office (who also use an Ubuntu OS desktop platform and who probably share the same frustration that the engineers at that other media company that backed Avatar have with their streaming content from CTV).
It is worth noting that CBC streaming video works just fine, everywhere. If only they had negotiated some ability to rebroadcast online for CTV; If so they would have benefited from being the only provider for a growing subset of the Canadian online viewing market, Ubuntu OS users.
To CTV’s credit, the coverage and availability of content is great; It is just unfortunate that being on the Internet in Canada is not the only prerequisite. If Novell’s moonlight version gets a little bit better (Currently results in Unhandled Exception: System.ExecutionEngineException: SIGILL) then there may be hope.. but probably not in time for the Olympics.
Looks like embedding 3GP video from the Wordpress Blackberry application (or any other source) is fairly trivial once you have completed a few configuration steps. One option is to modify the supported MIME types right in the code. Rather than change code I prefer to stay within the boundaries of the administration panels to ensure the the system can ever-green with package and plugin updates. For myself, running multi-site and supporting a few other users unofficially, this is especially important _ I do not want to spend time upgrading/migrating/configuring if I don’t have to. I have learned this from my experiences easily supporting Debian-based infrastructure through upgrades for decades now.
I found these instructions useful after a bit of googling; Since I didn’t hit the information I wanted on my first or second search, I thought it best to post my learnings.
1) Download the MIME type configuration plugin
2) Add a mime type for 3GP, or any other files that you want to upload that current result in the message “File type does not meet security guidelines.”
Below is a screenshot from the PJW mime configuration plug-in I used to add .3GP.
Some registered users may have received email notifications with a new temporary password. This is because until just a few minutes ago, many emails from this site were being blocked by our upstream ISP (Rogers) due to some issues with the headers in those emails. Likely this has prevented many users from registering and submitting comments .. but it is all working great now.
If you are running wordress on a debian-based system, and your ISP does implement good controls around outbound email (like ours), feel free to reach out to me here on ways in which you can work to get your specific configuration working the way you want it to. In particular, among the big residential ISPs (Bell, Rogers, Cogeco) there are differences in their outbound email policies, and I support blogs running on each of these. Rogers provides a more technically sound solution in the residential arena, however it requires stricter local configuration to work properly and their customer service does not have the technical knowledge to easily support what their engineers have in place. For us, there are a number of configuration factors in play, namely:
Correct configuration of our MX records in DNS
Alignment of MTA daemon configured hostname to those MX records
SPF records that contain all our sending hosts, including ISP SMTP servers
Registration of valid senders on mail servers, including upstream ISP SMTP servers
Proper configuration to allow authentication with upsteam ISP SMTP servers
Proper re-writing on the local webserver for all outbound email (corrected today)
Notably outbound mail on Bell generally works even if you mess up or do not complete #2, #4 and #5. I had been testing using my local email domain, bypassing some ISP checks related to #6, until today. Just a few years ago email worked fine ignoring all these steps and I expect that some day soon all ISPs will require all steps, as it no doubt enables more spam senders.
For debian-based hosting setup as a satellite with an ISP smarthost, I have found exim4-daemon-light the easiest to configure to manage everything outside the DNS configuration for successful Wordpress email integration with any of these ISPs. I may post a more comprehensive HOWTO based on our setup. Until I do that, reel free to register here, and peek at the headers of our emails _ nothing was done to any Wordpress configuration; It all happens in the backend MTA, noting that we also run Wordpress in a multi-site configuration.
In anticipation of the big Superbowl XLIV game this Sunday, there is a lot of banter in the social channels. It might not be an original idea, but I found this DC vs Marvel mock match-up hilarious. Although I seem to be familiar with more of the DC team players thanks to childhood cartoons, the Marvel line-up is easily associated with recent motion pictures. Personally, I can think of a number of entertaining on-field match-ups that would certainly include The Blob being handled by Supergirl at the line of scrimmage. If Vince McMahon had these teams, he probably would have succeeded with the XFL. The real question is how the DC team would stand up to this other Heros team that has been called creepy by some.
At the January Toronto Product Manager Association meeting I had an opportunity to listen to Alfred Tan, a Senior Director in Sales at Facebook, lead a discussion on Marketing in Social Media. I was a great presentation; Delivered on a Mac using a Rogers Rocket Stick, it had lots of really interesting slides and a wealth of stimulating media and data supporting successful product marketing initiatives within Facebook.
Personally, there were two thoughts that rang bells in my mind listening to Alfred’s presentation; The potency and viral nature that opinion brings to a brand exposed in social media and the potential for new product-related business intelligence and personalization of the product message based on data relationships originating in social networks.
I found strong validation for the idea that there can be risk to any large active brand that is trying to ignore engaging resources in social media. If you choose to ignore Facebook as a CPG marketing manager or retail marketing manager, and then discover that a social community is impacting your brand message (due to opinion or legit product/customer issues), it may be too late to do anything but try to hire some of the trusted messengers right out of the influencing community. At this stage of the economic cycle I would expect this is something that more and more skilled and unemployed marketers may be beginning to realise in order to leverage themselves back into the industry with their new ideas. Put another way, there appears to be incentive to create impact to a corporate brand in order to bring awareness to the personal brand. I expect (and know) most CPG brand managers have Twintern(s) keeping an eye on things in this domain.
The other thought, most strongly stimulated by this presentation, was validation that the federation of identities between established Identity stores like Facebook, Google, etc. provides a powerful mechanism through which businesses can flush out new successful products and guide potential new customers, among other things. The keynote here for me is that the identity allows for both powerful business intelligence (useful to both the provider and consumer of the identity) and personalization of product messages based on knowledge gained from knowing the identity of a user (browsing, chating, sharing, using). Sounds like nothing new, except that the community has a way of proofing the identity through the relationships established with other users. There is a lot of self-policing and opinion that helps ensure presented data about a user is somewhat accurate. This makes any identity associated with a user and one or more communities much more valuable as a driver for BI or Personalization type activities.
Internally, Facebook is already delivering this BI and Personalization to its business customers; Some great eye opening examples for both products and services were in Alfred’s presentation. Starbucks, Coca-Cola/Energy Brands, BMW to name a few. Facebook does probably the best job today IMHO of softening identity relationships for public consumption using terms like “Friends” and “Fans”.
Two years ago experts knew that 75% of people look to social networks when considering product purchases as shown below. The impact of the identity will no doubt create an even wider gap between the average measures of “Cost Quality” (a McKinsey term from below) as it relates to the effectiveness of retailers with and without identities covered by their go-to-market strategy.
Externally, once other organizations finish getting a handle on their employee identities (something I am actively engaged with on a day-to-day basis with our products) they will begin to look outward and leverage the power of B2C and C2B identity relationships… with C representing the Social Community where their customers reside. There are certainly not many technology constraints here, although it may not be perceived that way by everybody.
The benefits, for example at a retail web site, include having individual consumer expectations more closely met with each interaction as long as that individual has a compatible identity. Logging in and collection of preference data are no longer part of the consumer experience. The Identity brings with it relationships between customers (in other social communities), relationships to feedback (direct and in other communities), knowledge about the customer (location, age, preferences) and drives each interaction, personalizing it and improving the experience for each subsequent potential customer or the return of an existing customer. For instance, imagine driving the customer directly to the two brands or models he or she is interested in without having them guide themselves through a retail site on their own.
Businesses that have established customer identity stores will be the first to lead this next wave IMHO. Where will they learn about you?.. Just think of where you log in today; where you are part of loyalty programs today; where you participate in an online community today.
There is no need for consumers to worry about which Identity store will be the end-all (i.e. rush to OpenID, Google, Yahoo or Facebook now) and no need for retailers to fear being at the mercy of an identity provider for an information service fee. To help understand this point, think about the evolution of plastic in the wallet. Generally people prefer to minimize the number of change cards in their wallet. That does not stop them from shopping anywhere, and carrying either Interac, Visa or Mastercard. This same motivation will drive people to a single, portable, preferred identity (but not necessarily all from the same source). It will also push Internet sites to share their identities and any public meta data they have (anything you would know about your “friends” on Facebook), and allow them to be used anywhere. The providers that put up the first road-blocks on information sharing will likely find the end of the road, whereby they are bought “on the block”, re-thought and re-exposed in a more compatible, profitable manner.
A few nights ago I watched Avatar for the second time in 3-D, in an effort to qualify the difference between regular 3D and IMAX 3D. The difference was astounding, both in the picture quality due to the 10x resolution jump IMAX frames provide, the super-sound and the fact that many scenes seemed to have far more depth with detail. One aspect in particular, people seem to agree with, is that the scenes with the holographic computer consoles had obvious clarity in both the foreground and background compared to the normal 3-D where computer consoles in the background just become a blurry mess. You could read the writing on truck tires, the side of weapons, and other small places that gave you a real appreciation for the level of detail.
The real highlight for me was not related to IMAX at all. It was when a colleague of mine shared with me the news that the super-computing power required for the animation, that supposedly delayed the creation of Avatar for James Cameron for more than a decade, came from a multi-core Ubuntu system. Awesome. My online source Jordan Hall is here.
The real facts of this mega-Debian-SMP system at Weta Digital are apparently:
35,000 cores
over 4000 Hewlett Packard Blade servers
each minute of rendering equates to approximately 17.28GB of data
10Gbit SAN
nVidia Graphics
Below is an impressive photo of the water cooling required for the system. It actually looks more like an HVAC room given the size of the pipes, and lack of any obvious computer related equipment.
I know of some other facts that may have a relationship at work here.
“Toy Story” (1995) was the first feature-length computer-animated film, so successful it pretty much pushed hand-drawn animation to the sideline. It is also the source for all Debian distribution names. (Potato, Woody, Etch, Lenny)
Weta Digital were involved in “District 9″ and “i,Robot” (which I enjoyed) among others.
If 
