Recently, I’ve been thinking a lot about what it means to do really great front end development (HTML, Javascript, CSS). And, one of the key issues things I keep hitting upon is writing “scaleable” front end code. I use the word scaleable in a couple ways:

1. Writing code well in a way that you could keep adding developers and they could each increase productivity linearly.
2. Writing code that will work for increasingly larger sites.
3. Writing code that is maintainable, testable, and extendable.

I know others have thought about this, some people call it “best practices”, some people call it frameworks, Yahoo even went as far as to create the YUI family of libraries. I think deep down, most people were trying to think about how to “scale” whatever it was they were doing.  This entry is the first in a series of writing jQuery code that “scales”.

Another thing I probably need to address before we get started is “Shu”. It comes from the Japanese phrase “Shuhari“, which I learned from Alistair Cockburn and is common in the Agile circles, and “describes the stages of learning to mastery”. Once you’re aware that there are certain levels of mastery of something, you start asking yourself which level you’re at and what level others are.

JQuery is awesome, and I love it. For along time, I treated javascript as a 2nd class citizen, but I’ve recently been really digging into it and making it “scaleable”.

In Dan Cederholm’s recent book, he introduces some jQuery that I’m lumping into the “Shu” box (ba-da ching!).

$(document).ready(function(){
$("#loc-adv a").click(function(){
$("#map").slideToggle("slow");
return false;
});
});

Now don’t get me wrong, this code is perfectly acceptable, especially for the situation it’s presented in (a small, static site, built by 1-2 developers). I’m all for not complicating things if they don’t need to be. This code is the simplest thing that could possibly work, and that’s awesome. However, we’re talking about scaling javascript up to larger sites, more developers, and maintainable in legacy applications, and there are some things about this code which don’t make it very scaleable.

First, it’s not really re-useable. Sure, you could include it in any page that has the `#loc-adv` a link and a map, and it’s going to work, but what if  you needed multiple maps in one page or had multiple links. All of a sudden, you’re going to need to repeat yourself, and we don’t do that.

Second, if you put this right in your page, as it’s presented in the book, you can’t use it on other pages without repeating yourself. Maintaining this code could be painful if you put it in more than one page.

Third, the only way this code is going to fire is if you click the link. There’s no way to call it from other code, or test it, and you’ll have to attach it to multiple elements if you add another link or some other element that needs to reveal the map.

What if you want to link to the page with the map open? This code won’t account for that.

I’m not going to address all these issues in this post. I’m going to leave that up for the “Ha” and “Ri” posts. Look for them in a couple days.

I am have been reading Steve Arnold’s weblog on search, I have known Steve for over ten years now and he likes to challenge the status-quo and pushing people to see beyond the status-quo.

So it was very interesting to read his post about how Google is challenging Reed Elsevier and Thomson by indexing legal texts:

Google has added the full text of US federal cases and state cases. The coverage of the federal cases, district and appellate, is from 1924 to the present. US state cases cover 1950 to the present. Additional content will be added; for example, I have one source that suggested that the Commonwealth of Virginia Supreme Court will provide Google with CD ROMs of cases back to 1924. Google, according to this source, is talking with other sources of US legal information and may provide access to additional legal information as well.

His thesis is that the incumbents are like Vercingetorix stuck in Alesia (1) and that Google is like Ceasar who built two sets of wall around Alesia, one to keep the Gauls in, and the other to keep any relieving force out.

I like the analogy though it is not quite there, Google is not exactly laying siege and they don’t have to defend themselves. On the other hand the incumbents probably feel very much like the Gauls stuck in Alesia.

I was catching up with a long time friend earlier this week (a much smarter person than me), we were talking about lots of thing and one of those things was how particular species will move fluidly from one niche to another as they evolve. My feeling was that sometimes this happens in a fluid fashion without much struggle, but sometimes it can be quite violent resulting in the decimation of one or the other species (2). I wonder if this is closer to what is happening here. Google in moving in on an established market, though not in an explicitly deliberate fashion, and causing discomfort to the incumbents.

Now that a good portion of the data these incumbents charge for is available for free (it always was available for free, but access was difficult), it will likely force them to change their business model if they are to stay relevant. Steve makes that point very explicitly at he end of his post, for example:

Finally, what will be vulnerable to Google disruption will be difficult to use, expensive, and incomplete services. Maybe Reed Elsevier, Thomson Reuters, and Wolters Kluwer should merge. That will give the present crop of senior managers time to cash out. I don’t see an easy, quick, inexpensive, or painless way to prevent the lessons of Alesia being writ large in tomorrow’s digital headlines.

1 – I learned all about the siege of Alesia at when I was at school.

2 – For example, the introduction of Lionfish in the Caribbean is resulting in major population reduction in some indigenous species on the reefs.

Barbie– the monstrously popular doll of the 20th century!

As a child I owned a cabinet-full of them, but I could never really fall in love with her and I actually preferred to play with my animal toys.  Barbie was too… plastic, she didn’t have enough detail or value.  She’s preteen “corporate,” she’s “american,” barbie is “the woman.”

BJDs, on the other hand, are quickly becoming the new, modern beautiful doll in trend… and I must say that they are frankly a hundred times more stunning and well-made than Barbie ever was.  That, and they cost 25 times as much.  =/  Ball-jointed dolls are customizable, posable, more life-like, and larger.  They are also made by hundreds of artisans in all kinds of different molds!   Maybe I should add that they are Asian – and Asians are known for pushing achievements to the next level.

So out of curiosity, I’ve decided to do a bit of research to compare Barbie proportions with BJD proportions and normal people.  Barbies are approximately on a 1/6 scale, and the standard ball-jointed doll is about 1/3 the size of a real person. Using those proportions, here’s the estimated measurements in inches if the dolls were to be real:

BJD vs. Barbie proportions

“avg american” data taken from this site

Pretty interesting, huh?  I see that Barbie’s chest-waist ratio is more extreme than a BJD’s; though the newer generation is still not exactly realistic in size.

Thoughts?  Introspection on culture, trends?

Sewing a Barbie Pattern to a BJD

This is something I’ve been waiting to do since BEFORE I acquired Senna! =)  My generous grandmother bought me Barbie clothes patterns back in July, and now I’m undertaking the transformation of pattern scaling to make Senna her first real girl clothes.   The blue dress she had before is now too big for her.

Simplicity 4719: these patterns are so small, they could not fit all of the text onto the pieces! XD

Scaling -  I found that Senna’s measurements are approximately twice as big as Barbie’s, so I remade the dress pattern.  The really cool thing about doll clothes is they’re so simple; it is intellectually refreshing to see shapes and curves reduced into basic pieces that are very easy to understand how to fit together. And, the pattern doesn’t even call for lining or a zipper :O

Okay, so I ended up doing a lot more modifications to the pattern than “2x” to make it fit her– partially because I was lazy about getting accurate measurements, seam allowance, etc.  I also ended up lining the bodice with fleece to tone down the NHO, and added fabric in the skirting for modesty with this slightly transparent calico fabric. =)

Although I threw this together too quickly and didn’t even add a fastener in the back, Senna was SO happy to finally have a real dress…..

That she was willing to STAND ON ONE LEG for me for a full ten minutes!! BJD owners out there know that this is incredibly tough for their dolls to perform.

Actually, my original plan with this fabric  was to make a life size dress of similar style.

So… audience participation time! What’s an appropriate price for a custom-made dress of this pattern? I have received feedback that my prices should change – any suggestions on this? I will take what you give me on this one! 

*For some really interesting facts about Barbie’s history, I’d recommend checking out her wikipedia article.

edit: I have received ~5 comments on this, and not one price suggestion for this dress! =/

I had a few hours to spare a couple of days ago and decided to check InnoDB Plugin 1.0’s support for data compression.

In a project I work on from time to time, there is a table which contains three blobs which contains text data. To store the data efficiently I was using the COMPRESS() function in MySQL and doing a “CONVERT(UNCOMPRESS(text) AS utf8)” to uncompress the data and present it as utf8. No problems there, but with the recent move to the InnoDB Plugin 1.0 in MySQL 5.1 there was an opportunity to push that down the stack.

I ran a few benchmarks and it turned out that using 8K pages was the optimal trade-off between space and time. Using 16K pages did not compress the data very well, and using pages smaller than 8K increased the time needed to store the data. I should note that 8K is also the default.

There are some interesting wrinkles in all this, innodb_file_per_table needs to be enabled and I think the innodb_file_format needs to be set to ‘barracuda’ thought I am not sure about that.

Service Oriented Architectures [SOA] play a pivotal role in today’s Enterprise Systems. Service orientation requires loose coupling of services with the underlying technologies. There is a lot Web Oriented Architectures gain by taking the SOA approach, especially with regard to aspects like scaling and being cloud ready.

I dont use SOA here to mean XML web services, but as a pattern for breaking down applications into distributed software components (which we call services). The components being independent units that deliver specific functionality. Data is encapsulated and made accessible only through well defined hardened interfaces. Consumers of the service are provided with a contract for the functionality provided through the interface. Hence, a clear isolation of functionality and data storage is achieved. The services manage all data internally and never provide direct database accesses. Whether access happens as XML/JSON or over REST/XML-RPC/SOAP is irrelevant for SOA. They form the various implementation options.

SOA also provides operational advantages in terms of the break down of the development work. First, small teams are created around problems. Next, the interfaces are decided and agreed upon. The teams then have full freedom in choosing tools, design methodologies, etc., as long as they provide the promised functionality through the interface. Developers in the team own end-to-end responsibilities for the service and are subjected to direct customer feedback. The heightened sense of responsibility also usually translates to better ownership of work. With hardened interfaces, localised issues and reduced integration efforts the net chaos in the system reduces significantly.

The parallel for SOA approach in the programming paradigm would be the Object Oriented Programming approach [OOPS]. Like OOPS, SOA is also vastly misunderstood. Most programmers fail to think Object Oriented-ly (esp the ones coming from the procedural world) and end up building poorly architected systems, while blaming OOPs for its clumsiness. Similar is the case with SOA.

We use SOA extensively in the systems we build at Althea. In my future posts I plan to provide more practical insights into SOA.

Lindis Pass road is exposed to some rocks from bluffs near Camp Creek.

Lindis oct 09 (DAS)

Once again, helicopter sluicing and scaling by abseiling means have successfully permitted us to get rid of a few truck loads of rocks overhanging SH4 near Lindis Pass.

We made in five days the work planned for ten, to the satisfaction of Opus consultants . Thanks to NZTA maintenance manager from Downer EDI Works and the road management team who did a very professional work.

Let’s pretend I wrote a book.  Let’s pretend you are reading it now.  Let’s pretend we were childhood friends.  Let’s pretend that sometimes when you dream you dream of me.  And there, where we find each other, you ask me finally, Who?  And I go on circling and circling.  Or I stoop, and when I light, between two talons I hold my food.  Later, when you wake and try to retell what you dreamed, I am there.  Not as I was but in some new form.  A fly scaling the chandelier, a leaf the wind blew inside the house, a noise of engines roaring far away, a loose hair fallen from your head.  Hold me, as the blessing says, in the hollow of your hand.  For now you will be my god, and I will pray to you.  It is your job to listen or turn away.  Here are the first words.  Here are the last.

For a Cloud Infrastructure, the ability to scale dynamically is a double edged sword.

Lets take the example of a more reactive approach to scaling on the cloud. Here, the utilization of all resources(CPU, RAM) in the infrastructure is monitored continuously and when it goes above a pre-defined threshold level, more resources are added. Let’s consider some failure scenarios that can happen with this approach.

* We know that a bot crawling the site will eat up whatever resources are thrown at it. The bot’s activity triggers the addition of more instances hence further empowering it. This leads to a sort avalanche effect and a mightily pissed off CFO.

* Adding application servers to support increasing demand while there are significant bottlenecks in the database infrastructure, further worsens the overall performance of the system.

* Spikes in traffic result in triggering off multiple instances without significantly improving the end user-experience. Hence giving rise to a bloated cloud infrastructure without sufficient business benefits.

There are many approaches you can take to plan your cloud deployment to get around these scenarios. They will be the subject of some future posts.

One of the most touted features of cloud infrastructures is its ability to automatically scale vertically and horizontally with almost zero impact on the running application/service. It changes the way CFOs look at IT infrastructure funding. Though the feature is paradigm shifting, it needs to be exercised with care.

In the traditional approach of hosting private infrastructures, the peak capacity is estimated and all the necessary infrastructure investments are made upfront. Inspite of this you continue to run the risk of underestimating the peak capacity. In contrast, with the cloud approach you pay only for the resources you actually use and the cloud also scales up to accommodate any overshoots.

The cloud however can make architects lazy and not design systems efficiently. Hence using up more resources than required. It may also lead to the development of systems that can respond to very high demands by adding instances, when meeting such demands may have no business benefits.

To conclude my little foray into number encoding (see the presentation by Jeffrey Dean from Google titled “Challenges in Building Large-Scale Information Retrieval Systems” (video, slides)), here are a few conclusions:

• In terms of raw performance the “Varint Encoding” is much faster than “Byte-Aligned Variable-length Encodings” and I was able to get better numbers than Google got, most likely because I am using a different machine. It would be interesting to know what kind of machine/OS they used for their timings so I could do a direct comparison. My lookup array structure is different (and more compact) than Google’s, assuming I understood Google’s lookup array structure in the presentation.
• The “Byte-Aligned Variable-length Encodings” is faster if you are storing three numbers per posting, namely a document ID, a term position and a field ID. The “Group Varint Encoding” is faster if you are storing four number per posting, namely a document ID, a term position, a field ID and a weight.
• As I described in the last comment in the original post, two bits are used in the header for each varint to indicate its size in bytes, so 0, 1, 2 or 3 indicate whether your varint is 1, 2, 3 or 4 bytes long respectively. However if you store deltas a lot of the numbers you store will be 0, using a byte to store 0 seems wasteful to me. So I changed this so that the two bits indicate the actual number of bytes in the varint, and 0 bytes means 0. This way I don’t actually allocate space unless there is a value other than 0 to store. This saves about 10% in my overall index size, and a lot more if you only take the term postings into account because I store some amount of document metadata in my index. Of course this means that you can’t store a number greater than 16,777,216 which won’t happen unless you are creating huge indices with more than 16,777,216 documents in them or have documents longer that 16,777,216 terms.

Basically it comes down to trade-offs, index compactness vs. decode speed, and looking at speed both in test code (usually a contrived example) and performance on a real data set. I used the Wikipedia data for that along with 200 relatively complex searches designed to read lots of postings lists.

Plenty has been written about the Danger data loss over the weekend (TechCrunch). For me the most interesting commentary came from John C. Dvorak, he got some things right but he also got some things wrong:

Over the past week, users of the T-Mobile Sidekick platform found that all their contacts and other important information was permanently lost, because of server mishaps. If Microsoft had wanted to throw a monkey wrench into cloud computing, it could not have done a better job.

Huh, don’t think so, this was a data loss screw-up, nothing to do with the cloud. If what we are reading is to be believed, a SAN upgrade went wrong and the data was lost, with no backup.

Insert Ellen Feiss ad here…

Seriously though backups are essential because things will go wrong. Note that I use ‘will’ and not ‘may’, and these may or may not be under our control.

The other things I used to tell clients is to do a fire drill on a regular basis, by that I mean taking the backups and making sure they can be restored properly and completely. I had one client who discovered that all their backups were useless when they checked.

Last week was a first for me with the blog. First time I’ve written a whole week of blog posts in advance and pre-scheduled them, because I was off visiting Arb (not that I didn’t have net access, we are civilised folks after all).

I’m not sure how other bloggers organise their writing but I usually note down ideas when I have them and write one up either the evening before or early in the morning. So that was my brief flirtation with being organised, I promise it won’t happen again And if I was a bit slower with replies then that’s why.

So Champions Online and Aion have been out for about a month now. How are people finding them?

Melf has a great Aion review up at Word of Shadow. I prefer reviews where people list both good and bad things about the game, especially when the reviewer basically liked the game, because that means they probably ‘got’ whatever it is supposed to be about and can hopefully explain it to readers. Evizaer also had a look at Aion and gave it a straight no.

Girl Unplugged posts a Champions Online review, again this is a review from someone who likes the game and can explain why.  And Syp has a solid point by point comparison between CO and City of Heroes. I do find it interesting that people who have bought a lifetime subscription are much more likely to take a longterm view of a game – ie. Oh it’s a bit rough now but it’ll be great in a year’s time. If I’d paid $200 up front, I don’t think I’d be too thrilled about having to wait a year for greatness.

And still on the superhero theme, I have a basic disagreement with Muckbeast in the comments on his post about attracting women gamers, about whether the superhero genre is more popular with women than fantasy or sci-fi. (Harry Potter, Lord of the Rings, Twilight … I’m just sayin’.)

Other Stuff to Read: Twelve of the Best

1. Make Your Own Zombie Game – the zombie game experience gives you a chance to throw in $10 and participate in some game design. I’ll be curious to see what they come up with.
2. Wolfshead explains why scaling content should be our future! Why do we have to stick to group and raid sizes that are predetermined? Couldn’t the content just scale, like it does in CoH or Diablo?
3. tor.com is a blog that deserves a spot on any geek’s newsreader. This month is Steampunk month and they post a Steampunk 101 guide with an incredible shot at the bottom of a steampunked up laptop that has instantly become my object of desire.
4. Jaye at Journeys with Jaye explains why his exercise bike is an MMO and the scary thing is … he isn’t entirely kidding. Edited to add: Mea Culpa and sorry Jaye for getting the gender wrong, that should read SHE.
5. Jormundgard tries to psychoanalyse Garrosh Hellscream and explains why he’s disappointed with how that character has been developed.
6. Andrew Doull finds Puzzlequest quite traumatic and a lot of his reactions could apply equally to any quest based game. What does it mean if we’ll go commit (virtual) genocide just because an NPC in a position of authority told us to do it?
7. And although this may possibly be the least subtle link between links ever, I thought it was absolutely fascinating that The Anne Frank House were able to post up an actual video of Anne Frank on youtube this week. This is (obviously) from before the family went into hiding.
8. Keen writes a sharp, well observed post asking whether MMOs are being designed for too many players these days. And what do you lose when you decide to go for the mainstream?
9. And two thematically related posts: Tobold wonders how people like their games to be paced – if it’s all excitement all of the time then there’s never any downtime in which to socialise after all. And Andrew@Of Tooth and Claw asks how people feel about difficulty in games, and particularly about ‘cheating’ to sidestep the difficulty if it is getting in the way.
10. Hudson splits the CO community into two parts, conceptualists who try to stay true to a character concept and minmaxers who design their character concept around whichever powers work best at the time. Which are you?
11. Larisa wonders if it’s OK to apply to another guild while you are still guilded. After all, it’s OK to apply for new jobs while you’re still employed (at least until your employer finds out).
12. And another tor.com link, this time to a fantastic report on a reading and Q&A session with Michael Chabon (another of my favourite ever living authors) who is a dyed in the wool geek and proud of it. I’ll end with a quote from him:

… he goes on to describe the way fandom binds people together:

“For in playing, or writing, or drawing, or simply talking oneself deep into the world of a popular artwork that invites the regard of the amateur, the fan, one is seeking above all to connect, not only with the world of the show, comic book, or film, but with the encircling, embracing metaworld of all those who love it as much as you do.”

Growth, innovation, scaling, and the pace of life in cities

Solid surface counters are but the most preferred ones due to ease of maintenance and longevity. Another reason why many home owners solid surfaces is color diversity. Unlike natural stone, solid surfaces are made by man and therefore they can be pigmented in any way. There are lots of uniform colors for solid surface countertops, so that you can select the one you feel will be the best for your kitchen interior.
Solid surface countertops are stain-resistant. They have a non-porous structure, which rules out the penetration of bacteria and germs into the surface. This is also a reason why solid surface counters are less likely to discolor with age than granite counters.
These countertops are made from a piece of wood or a wood-byproduct, which is covered with several layers of plastic. Counters made from engineered stones are also termed solid surface counters. Solid surface countertops are easier fitted to various home interiors than granite or marble counters. They are less demanding in terms of maintenance and tend to be stronger.

I always hesitate when I type words like that – customisation – I was taught in Australia and New Zealand to spell it with an “s”, yet I find in my job, typesetting, that many here now prefer “z”. I’m guessing most readers here are US based, so I usually use a “z”. What an absurdly ridiculous thing to have to give so much though to.

Neuroprogrammer 2 is still a favourite of mine. I like the style of the supplied sessions. They’re cleaner and more simple than the Mind Workstation equivalents. I also like the ease with which the sessions can have quite dramatic changes made without editing. The “Customize” button makes accessible many of the capabilities otherwise accessible only via the session editor.

The “General” tab allows the session’s tracks to be muted – any combination of Background Sounds, Noise and Tone can be selected, although the suggestion is made that Noise and Background together may be unnecessary. Fade In and Fade out can be adjusted – worth remembering when worrying about start and end volume when designing sessions.

You’ll all know the Session Countdown Timer - here under the General tab you can change the countdown time. I find a longer settle time useful when I’m using biofeedback.

Digging deeper into General, we find “Customize Volume/Intensities” and the Advanced Customization dialogue. Here you’ll find most of the controls that would otherwise be found using the Node Edit in the session editor.

The controls are divided up according to the type of track they are relevant to – Tones, Noise, Background Sound.

Tones and Sound/Music volume are obvious – changes the volume of the sessions binaural, monaural or isochronic beats or the sound tracks respectively.

Noise also has a volume control. In addition, the smoothness of the noise can be altered. In its most basic form, noise is a random value in each bit time of the audio bitrate. Smoothing limits the random range, making the sound softer and less “hissy”.

The background track can have the intensities of the Pitch Modulation or the Crossfeed adjusted. Pitch Modulation varies the pitch of the sound track at the entrainment rate and to a depth selected by the Pitch Mod Intensity slider. Crossfeed swaps left and right channels at the entrainment rate, again to a depth set by the intensity slider. It’s is possible to trigger entrainment with any repetitive stimulus, but these methods are much more aesthetically pleasing than effective.

Noise and sound files can be pulse modulated at the entrainment rate. The noise track or sound track is modulated in the same way a tones track is with an isochronic beat. The Pulse Intensity slider determines how much the volume of the track is varied – maximum chops the sound into discrete pulses.

A Low Pass filter blocks high frequencies, allowing only low frequencies to pass. Low Pass modulation is similar to pulse modulation, except that it only acts on part of the frequency range of the sound track.

Autopan shuttles the audio perceived centre left and right, just like quickly flicking the balance control on a stereo back and forth.

Next along is the Scaling tab. This deals with all matters relating to session length. In the Session Options, when editing a session, a pair of Ramp Markers can be set, beginning and end. When scaling a session the entire session, including ramps, can be stretched or condensed, or just the part between ramp markers. Play to Infinity plays the last node forever – take care using this if the session has an end ramp. Start and end times for any selected recordings can be set, however this is best left alone unless you wrote the session – the main purpose is to begin suggestions when the session reaches the point where maximum suggestibility should have been achieved. “More Info” takes you to the Customization page of the NP2 Help file.

Want to experiment with lower pitch sessions? Under the Tones/Sound tab is the opportunity to over-ride the session’s designed pitch(es). Even if the session usually changes pitch throughout, specifying the Tone Pitch will make the entire session play at the selected pitch. This tab is also the place to replace the default background track with your own – either music or hypnosis/NLP content. If the session is longer than the background track, the sound file will be looped. To prevent noisy transitions at the loop crossover, a selected number of seconds can be overlapped and faded. “Advanced Customization” takes you to the same dialogue as the “Customize Volumes/Intensities” in the General tab.

The AudioStrobe Flashing tab contains all the controls for the visual side of things. AudioStrobe can be turned on or off. If you have sensitive hearing and can hear the high pitched AS tone, this can be a useful option if you’re not using light glasses. Adjusting AS brightness here can overcome some problems with getting a satisfactory lightshow and an acceptable audio volume on some mind machines. The Screen Flashing colors are used in conjunction with the “Use Screen Flashing” checkbox on the main NP2 session screen. Screen Flashing can be effective, but don’t let it put you off getting a proper AudioStrobe decoder and glasses.

From the Recordings tab the way suggestions and other recordings are played can be controlled. Usually recordings fade in and out over 20 seconds to prevent startle, but some recordings may need to be started more quickly, if the important content begins immediately. Panning the “speaker” left and right can be quite effective with suggestions – Pan Modulation will occur at the entrainment rate, with the depth adjustment varying how dramatic the left/right shift will sound.

The final tab determines what happens when the session ends. I find the “Shut Computer Down” option great for sleep induction sessions. The “Open and Play Session” option allows sessions to be run in sequence. By setting the session end options of the called session, any number of sessions can be chained together. If you have a habit of falling asleep during sessions, the “Play Sound File” option will let you sound an alarm. I have yet to need to have NP2 launch another application when it finishes, but it’s nice to know I can if I do.

At the bottom left of the Session Customization dialogue a button marked “Save Config As” has been lurking throughout. When you customize a session, this customization is saved under the same name as the session, and becomes the default customization for the session thereafter. If you want to keep particular session customization, you use the Save Config As button. When you hit the “Customize” button, the first thing on the Session Customization dialogue is “Saved Configurations” – from here you can select any pre-saved configuration, including ones devised for use with other sessions. Supposing you customize the “Alpha Alternating Hypnotic Relax” just the way you want it and you want to apply the same tweaks to “Alpha Hypnotic Relax” – open the Alternating Hypnotic session, click “Customise” and choose the config you saved for the plain Hypnotic session.

It is this customization facility that is a distinguishing mark of NP2. To change a session in Mind Workstation means getting in there and editing tracks. MWS is a session creator’s tool, with more controls than most people will ever find, let alone use. NP2 provides maximum flexibility for the user who is mostly going to want to play existing tracks. With NP2 significant variation in the character of a session can be achieved with negligible effort. The easiest way to become familiar with all the settings is to set up a session containing one plain tones track, a simple background and an unmodulated noise track. Change customization settings one at a time and listen for the difference.

There’s a lot I like about NP2. Of the Transparent packages, NP2 offers the most comprehensive selection of sessions, all designed in accordance with the best research and trial information available. It’s the first piece of real brainwave stimulation software I bought, it’s what got me totally hooked, and it still delivers some of my favourite sessions. I’m loooking forward to seeing what NP3 will bring – improved user interface and fully editable sessions, I believe, among other things. I hope it doesn’t stray too far from the clean, simple listener-oriented interface of NP2.

Cheers,
Craig

A while back I came across this very interesting presentation by Jeffrey Dean from Google titled “Challenges in Building Large-Scale Information Retrieval Systems” (video, slides) where he talks about the various challenges that Google ran into over time as they scaled up and how they solved them.

This abstract sets the stage pretty well

Building and operating large-scale information retrieval systems used by hundreds of millions of people around the world provides a number of interesting challenges. Designing such systems requires making complex design tradeoffs in a number of dimensions, including (a) the number of user queries that must be handled per second and the response latency to these requests, (b) the number and size of various corpora that are searched, (c) the latency and frequency with which documents are updated or added to the corpora, and (d) the quality and cost of the ranking algorithms that are used for retrieval.

In this talk I’ll discuss the evolution of Google’s hardware infrastructure and information retrieval systems and some of the design challenges that arise from ever-increasing demands in all of these dimensions. I’ll also describe how we use various pieces of distributed systems infrastructure when building these retrieval systems.

Finally, I’ll describe some future challenges and open research problems in this area.

What particularly appealed to my (gnarly) mind were the various encoding algorithms they tried to encode numbers in their index (pages 54 to 63 of the slides).

Having written a search engine I was particularly interested in how they approached that. I actually settled for what they call “Byte-Aligned Variable-length Encodings” (slide 56) ten years ago having done a bunch of benchmarking (which I repeated every few years to see how well it fared on newer generations of processors.) The encoding is very compact and does very well when you have powerful CPUs and weak IO. I did some revisions over time in the C code which drives it to steer the optimizer, but it has held up well. I used this encoding for the Feedster search engine. As an aside one of the pluses of this encoding is that it is endian independent.

So I was very interested by the “Group Varint Encoding” (slide 63) specifically the claim that the decode time was faster.

Google was able to achieve a decode speed of ~180M numbers/second for the Byte-Aligned Variable-length Encodings and ~400M numbers/second for the Group Varint Encoding.

So I set out to replicate the results and ran into some interesting things.

The machines used for this are an Intel DX58SO motherboard with a Core i7 920 CPU, 2.66GHz, 6GB of RAM, Fedora 11, gcc 4.1.1, and a Mac Pro with 2 Dual Core Xeons, 2.66GHz, 13GB of RAM, Mac OS X 10.5.8, gcc 4.2.1.

The test code is pretty simple, one test writes and reads number from the same area of memory and the second test writes and reads a sequence of numbers over 1GB of memory. I added in verification code to check that the number read was the same as the number written to simulate some in between reading the data.

My results for Byte-Aligned Variable-length Encodings were better, I was able to get ~226M numbers/second on the DX58SO and ~190M numbers/second on the Mac Pro. Removing the verification code I was able to get ~426M and ~405M numbers/second respectively.

And the results for the Group Varint Encoding were much better, I was able to get ~680M numbers/second on the DX58SO and ~500M numbers/second on the Mac Pro. Removing the verification code made reading pretty much instantaneous (less than 1 microsecond to read 4.8B numbers.)

UPDATED October 9th, 2009 – Obviously my own BS meter was off when I wrote that last paragraphs, as I point out in the comments, the optimizer was doing away with most of the code because the results were not used. Preventing it from doing that gave me the same results as the ones I got with the verification code in place and was certainly not instantaneous.

Thanks to Blake for passing this on.

“If you want to build a ship, don’t drum up the people to gather wood, divide the work, and give orders.  Instead, teach them to yearn for the vast and endless sea.”

-Antoine De Saint-Exupery,
Author of The Little Prince

I am teaching myself cross-fit principles and techniques. I want to live the 100-words of world-class fitness. I found www.crossfitbrandx. com through the main www.CrossFit.com site as a place offering a warm welcome to beginners like me.

I am interested in scaled WODs and a lifetime of fitness.  I will run my first half-marathon next month. I’ve completed three sprint-distance tri’s. I’ve found a great deal of success eating in the Paleo Zone.  Next year’s goals include an Oly distance tri, a Half Ironman, and if I can raise the funds, a full Ironman.  I want to train for these endurance events without overtraining or compromising my health.

So, Friday I finally started.   My mantra: start where I am.  

Gigi Sehat

Karena untuk sementara waktu aku tidak ada kegiatan luar, maka untuk agar dapur tetap ngebul aku ngelamar kerja jadi supri (Supir Pribadi). Setelah beberapa kali mendapat  tugas menjadi “supri” bossku dalam rangka mengantar anak-anak boss untuk perawatan giginya. Kegiatan ini disamping ada waktu senggang juga merupakan kesempatan yang langka bagi orang seperti aku untuk datang ke dokter.

Seperti biasa meski sudah janji melalui telepon untuk pesan nomor antri, namun sesampainya di tempat praktek itu sudah banyak pasien yang antri menunggu. Karena nomor urut masih cukup jauh aku sempatkan untuk melihat-lihat suasana poli gigi ini sekalian cuci mata. (Tuuh.. dasar mata keranjang, pasti deh nyari-nyari..).

Setelah menjelajahi siapa saja pasien yang antri dan pengantarnya, matakupun beralih menyusur ruangan tunggu yang bagus dan nyaman. Nyaman bagiku yang tidak dalam masalah gigi tentu berbeda dengan yang sedang bermasalah giginya. Serius kuperhatikan gambar susunan gigi, bentuk gigi, macam-macam penyakit gigi dan terakhir pada cara yang benar perawatan gigi.

Skema Gigi

Terdengar suara anak bossku dipanggil untuk masuk keruang praktek dokter. Segera kuantar masuk ruang periksa. Setelah ber “say hello” sejenak dengan sang dokter langsung anak bossku naik ke kursi pemeriksaan (dental chair) yang  sepintas kubaca merk china model MDTE – 638. Kuamati bu dokter yang masih muda dan ramah senyum meskipun tertutup masker. Dengan sabar dan lembut juga interaktif memeriksa gigi anak bossku itu. (jadi teringat bundo NakjadiMande, neh.. hik.. hik)

Rupanya hanya akibat kurang rajin merawat gigi, maka timbul peradangan pada gusi geraham akibat sisa makanan yang tertinggal. Setelah selesai pemeriksaan dan mencatat pada buku catatannya, bu dokter memberikan resep obat yang harus diminum. Dan pesannya “jangan lupa setiap mau tidur, gosok gigi dulu ya..” spontan dijawab “ya bu dokter”.

Sambil berjalan pulang, aku teringat akan pesan dokter gigi tadi dan cara perawatan gigi pada dinding ruang tunggu. Timbul suatu pengertian yang lebih dalam tentang perlunya perawatan gigi, agar tidak terkena resiko sakit gigi yang kata orang rasanya “sangat menyiksa”.

Dalam pengetahuan kebersihan dan perawatan gigi, sudahkah anda melakukan perawatan gigi dengan benar?

Gambar:

Dokumentasi drg. Evy Novianti.

Quando intendo caricare qualche immagine su questo log ho sempre la cura di ritagliarla e/o scalarla nella dimensione in cui voglio che venga vista da chi capita tra queste pagine: in pratica le operazioni di scaling e cropping le effettuo io in Gimp, non lascio che sia WordPress ad occuparsi della scalatura. La larghezza (in pixel) dell’immagine chiaramente la scelgo in base al tema che uso e quindi mi sono sempre permesso di usare l’opzione full size in sede di inserimento dopo il caricamento, sicuro che l’immagine non trabocchi fuori dal tema.

É questo il motivo per cui mi sono reso conto solo ora della nuova nozione di full size presso WordPress, a partire dalla versione 2.7. Con le versioni precedenti per full size di un’immagine si intendevano le dimensioni originali di caricamento. Adesso invece l’opzione full size effettua una scalatura dell’immagine nella larghezza massima supportata dal tema in uso (mantenendo comunque su sever l’immagine originale). In questo modo cambiando il tema del blog l’immagine viene ridimensionata in automatico.

Chi volesse, per varie ragioni, mantenere un full size effettivo deve seguire le facili indicazioni prese da qui:

1. caricare l’immagine com al solito
2. copiare l’URL dell’immaginea termine del caricamento (l’URL è del tipo http://kyklopslog.wordpress.com/files/2009/08/immagine.png)
3. passare alla linguetta URL di provenienza
4. inserire nel campo URL immagine l’URL copiata al passaggio 2
5. cliccare su Inserisci nell’articolo.