Robin's Notes

[Unboxing Pictures]

This isn’t really intended to be a full review, and it’s certainly not intended to be unbiased or helpful to all possible kindle purchasers, although hopefully my perspective will be useful to some people.

The triggering factor leading me to buy the Kindle DX was the promise of native PDF reading.  Over the past few years I have accumulated a stack of PDF files - some of them purchased e-books, and some of them academic papers.

The e-books have proven useful as reference materials on occasion, in particular the computer related ones that I can keep on my laptop and refer to when I’m working, but the papers have generally been too difficult for me to read on the laptop screen, so my hope for the kindle is that it will enable me to read these comfortably.

I have also, to my great surprise, found it pleasurable to read novels using the Kindle iPhone app.

My Kindle DX arrived yesterday, and so far I would have to call my experience very positive.  Most importantly, I’ve been reading a lot - perhaps a few hundred pages already.

There was a letter from Jeff Bezos pre-loaded on the device.  He makes the point that the main design goal of the kindle is for it to ‘disappear’ from the reader’s experience and allow one to connect directly with the reading material, the way one does with a book.  For actual reading as opposed to skimming, there is no doubt in my mind that this goal has been reached.

The e-ink display is extremely sharp, and as Amazon repeatedly points out, it is much less tiring to read from than a backlit LCD.

I think there’s also a psychological factor at work.  The stability of the display and the impression of ink on paper, along with the limitation of the device, creates a feeling of calm about the text.  You know it’s not going anywhere - there’s no application to crash, no interrupting dialogs, or other windows to deal with, and you aren’t going to think of using the kindle to play a game or write a tweet.  What the kindle cannot do is as important as what it does in helping to create a suitable state of mind for reading.

The kindle does require a reasonable amount of ambient light.  Although the display is very crisp, the contrast is nowhere near as good as ink-on paper yet.  It’s fine under indoor daylight conditions, or the conditions under which one would read a novel, but after using laptops and the iPhone in dimly lit rooms for years, having to consider the ambient lighting conditions is taking me some time to get used to.

The PDF reader does a very good job of rendering.  The documents look pretty much as they would on paper, and having read many pages already, I would say that any difficulties I am facing have to do with the subject matter and not the Kindle.

The biggest limitation, which could be corrected in software, is that the PDF reader is not aware of hyperlinks within documents.  This is particularly problematic for large reference books, forcing you to enter page numbers to navigate rather than selecting links from the table of contents.  This limitation is not present for kindle formatted e-books purchased from Amazon, so there would seem to be no technical reason for it not to be corrected in future.

Page turning in general is not super-fast.  In practice this is completely unnoticable during normal reading, but is quite obvious when ‘flicking through’ a document.  Image heavy PDFs are particularly bad in this regard.  The very worst documents are those PDFs which are just a collection of scans.  These are still perfectly good for long-form reading, but border on unusable for skimming.

So far I’ve been able to use the bookmarking system, and the go-to-page function to alleviate the need to flick through books.  Time will tell whether this will be a problem in general practice.

Generally when I decide to buy an electronic device, I consider whether it will improve my life on a daily basis in some way that I care about, and how much use I expect to get out of it before feeling the need to replace it with something better.

It will be a week or so before I have a real answer to these questions, but my impression so far is that the Kindle DX really does enable me to do a lot of reading that was previously inaccessible.

As far as lifespan is concerned, it seems likely to me that the display technology will have improved sufficiently, through improvements in contrast and the addition of color, that I anticipate wanting to upgrade in as little as two years.

At $500, that means that this device works out to about $20 per month, which doesn’t seem seems bad to me given the immense value of two years of reading.

It’s all very well to predict features a few days in advance of a release, and with plenty of leaks to choose from.  I thought I’d write down what my wild guess about what will be coming a year or so from now.  I’ve given it some thought since I’m about to commit myself to a new 2 year contract and face potential pain when the next upgrade comes around.

1. More speed and memory naturally, because it’s now just like all of the other computers and will get bumped progressively, and their competitors will be doing this too.
2. Software Defined Radio.  This would allow Apple to support CDMA networks without increasing the silicon count, and thus make a single hardware design work across all networks.
3. Front facing video camera.  It won’t be an ‘eye’ by the microphone. It will be hidden behind the display. This could make for some interesting game control and UI innovations through gesture recognition in addition to the obvious videophone.
4. Higher screen resolution.  I’m just throwing this in - it’s not essential,  but there’s still aliasing visible on the iPhone screen - double the resolution and it would be like paper.  I suspect if it’s technically feasible, Mr Jobs and Mr Ive would like to get this done.

I think if these emerge along with another year’s worth of software innovations we’ll have yet another killer device.

Would it be enough to make me upgrade?  I skipped the 3G and have no regrets about that.  I’m about to buy a 3GS which has numerous features that will be a step change in usefulness for me.  My guess is that I’ll be on the fence, but that it will be a delight for the 3G users who don’t upgrade this year.

At Apple’s press conference this morning, they announced that iPhone, and iPod touch applications would have the ability to control external hardware via either bluetooth, or the dock.

Oddly, most commentators seem to be intent on making a big noise about the inclusion of copy and paste, but are completely overlooking the major implications of what seems me to be by far the most significant feature.

Steve Jobs stated that the company now regarded itself as a consumer electronics company and renamed it from Apple Computer to Apple Inc to highlight this at the same time he first announced the iPhone.

Apple has experimented with this kind of device control, both in the form of the iTunes remote application, and in the form of the nike + iPod application, so they have a good understanding of how this type of application works.

By enabling the iPhone / iPod to function as the user interface for devices, Apple has created the infrastructure for a whole new class of consumer electronic devices.  Compared to previous generations of device where the UI was self-contained they have the following qualities:

• They can be smaller, because they no longer need a display surface or physical controls.
• They can be cheaper, because they no longer need display or control components.
• They can have a far superior user interface, because the UI can be multi-touch, 3d, media enabled, and make use of any and all of the features of Cocoa Touch.
• They can be internet connected - via the iPhone as a hub, and they can have multiple remote controls - one for everyone in the family.
• Their firmware can always be up-to-date - because it can be distributed as a payload within the iPhone application, or because the iPhone application can act as the upgrade agent.

The example used by Apple today (the blood glucose monitor) demonstrated the power of making the device part of a connected application, but almost any consumer electronics category you can think of can be improved using this technology.  Here are a few examples:

Camera - Who need a self-timer when you can trigger the shutter from your phone? Especially if you can check that the shot is right with a glance at the phone screen.  DSLRs could offer a lot more, including immediately available contact sheets of ‘phone-res’ proofs for quick emailing etc.

Lighting, Fans, HVAC- Obviously a high-end air conditioning system could have a powerful control system and programming interface that would run on the iPhone.  What is less obvious is that low end separate units such as floor standing lamps, desk fans, floor heaters etc, could also expose a bluetooth control interface without adding a huge premium, and that a group of such devices could then be controlled from a single device though a beautiful interface. Even if all it provided was an off switch so you could turn everything off at night, this would be a big win, but obviously an app could provide a lot more, including energy monitoring, and all the usual home automation goodness without the need to geek out with X10 etc.

Cars - Aside from gimmicks like remote starting, a car application could show tire pressure, service history, fuel level / range, and any other sensor data that were available, even while the car was parked in front of your house or in the garage.  As well as being able to access this information from the phone, it could be uploaded to a web application.  The application could help you monitor fuel efficiency, wear and tear etc, and could even help in establishing a trustworthy value for used vehicles.

Need I go on? It’s hard to think of a type of electronic device for which a personal, secure, connected, and powerful interface as could be afforded by an iPhone/iPod touch would not be of benefit.

Solving these problems has been the holy grail of consumer electronics hardware and software makers for decades.  Nobody else has anything else even remotely competitive in place whilst Apple now has all of the pieces - the best UI (we can argue about the future, but for now there is no contest), the internet infrastructure, the customer base, the credibility, and the experience.

Furthermore, because of the cost and simplicity implications of the new devices in comparison with those who try to put the UI on the device, there is an opportunity for startups and smaller companies to produce these ‘headless’ devices and compete with industry giants like Sony, Phillips, etc.  Until there is a competitive infrastructure, devices that rest on Apple’s infrastructure will be better, cheaper, and more connected than those that don’t.

The only real alternative that I can see is for someone to out-open Apple and connect devices directly to the cloud.  This may be the ultimate way of things, but as of yet attempts have been fumbled, leaders have not emerged, and Apple is currently in control.

I bought a Taro root a couple of days ago at the grocery store.  After consulting the googles, I learned a few interesting things about it:

1. It’s poisonous.

2. If you attempt to peel it, it ejects tiny crystalline needles which cause skin irritation, and lodge in your eyes and take up to three months to heal.

3. It tastes bitter.

These issues are resolved by cooking it.

Apparently in Northern India it is cooked by frying, smashing it and then frying again.  I have been using this method with plantain recently, so rather than changing two variables at once, I decided to apply it to the Taro.

Method:

I started off by using gloves and peeling it using a sharp knife. Next, I sliced it into thin slices (3-5mm). I sautéed the slices in olive oil for about 3 minutes a side - basically until the edged were browning and the middle was starting to look translucent but still white and not browned.  I then took the pieces out, dried them on a piece of kitchen towel, and put them on a chopping board. I smashed them hard using a potato masher, until the ‘skin’ was obviously broken. I sautéed them again for about another 4 minutes a side, until they were golden to brown all over.  Sautéing them is tricky because they tend to curl, so I had to keep pressing them down with a wooden spoon.

Results:

The result was crispy slices, somewhat similar in texture to fried potato although a little more fibrous and fruity.  The flavor was a little nutty, slightly sweet and with a hint of bitterness, although not enough to be unpleasant.  I’m not sure whether the bitterness would be eliminated with more cooking.  Next time I’ll probably try using a lower heat and cooking for longer.

They seem nowhere near as balanced as plantains in terms of flavor or effect - it seems as though they would complement something else quite well, but for my mind didn’t stand too well on their own.  

Here is a recipe I made up today from various random sources:

Ingredients:

• 2 cups red lentils
• 1 teaspoon turmeric powder
• 7.5 cups water

1. Put in a pan, bring to the boil and simmer for 45 mins stir every now and then.

More ingredients:

• 2 habaneros 
• 8 cloves of garlic - basically a whole garlic
• 1 large onion
• 1 teaspoon cumin powder
• 1/2 teaspoon cumin seeds
• 1 inch section of fresh ginger
• olive oil

While the lentils are simmering, prep the vegetables:

2. Remove the seeds and the middle fibres that the seeds are stuck to finely and chop the habaneros - I mean really fine

DO NOT TOUCH HABANEROS WITH YOUR SKIN AT ANY TIME - USE GLOVES.  

I would suggest having a small container ready for when you’ve done the chopping. Put the chopped pieces in the container and set it aside, and then immediately wash the knife and chopping board.  This isn’t a joke, you don’t want any habanero juice getting in someone’s eye, not to mention various other unpleasant possibilities.

3. finely chop the garlic - or put it through a press

4. finely dice the onion

5. peel and finely chop the ginger - into tiny little pieces

6. put the ginger in the pot with the lentils and stir it in timing is not critical for this but you don’t want to wait until after the frying

After the 45 minutes, keep the lentils simmering and then do this:

7. put a couple of tablespoons of oil in a frying pan and heat fairly high

8. add the cumin powder and cumin seeds - let them sizzle for a minute or two

9. carefully add the habaneros sizzle them for a minute - be very careful about spitting oil

10. add the garlic - let that sizzle for a minute

During all this sizzling, mix well with a wooden spoon or similar implement.  You’re really wanting to mix the flavors into the oil.

11. Add the onion.  Fry the onion in the oil for about 10 minutes - the garlic will be browning, and the onion will be translucent.  Stir all the time so that the onion gets ‘dirty’ with the spices and chili.

12. After the 10 minutes, add this mixture to the lentils and stir well.

13. Now simmer the lentils slowly until the dal is at the consistency you desire.  For me this took another 30 minutes or so.  I like it fairly creamy with just a little liquid.

I let it stand for about 15 minutes with the pot lid on while I cooked some rice.

I served it over the rice with a big chunk of fresh chopped cilantro leaves, and some halved cherry tomatoes.  Squeezing a little fresh lime juice over the top added zest.  

The recipe made enough for 4 people.  I’m expecting the flavor to develop more as it sits overnight.

It’s spicy enough to stimulate the organs, but stays true to the spirit of dal and shouldn’t blow your head off.  I’d call it a medium spicy dish.  The general tone of it is wholesome and clean tasting.

Let me know if you have suggestions.

I love the idea of a kindle.  I think it would be good for my mental health.  Reading material of substance, as well as getting away from the instant crack hits of social media seems like a good thing.

Appealing as the new Kindle 2 is, it’s not ready for me because:

1. An Apple tablet would make it redundant.  By Apple tablet, I mean something about 3x the size of an iphone, running the touch OS.  It would have a screen large enough to be comfortable for reading, but more importantly, it would be just one device to carry around.  It would be weird if Apple releases such a thing and Amazon doesn’t release a Kindle App to go with it - so that’s what I’m waiting for.

2. As far as I know, PDF reading still sucks.  As I understand it, you can pay Amazon to convert a PDF by emailing it to them.  However the results are not good,  I have numerous technical books that I have purchased in PDF form, which I’d love to have on a tablet device, but they need to be converted really well.  There are also all sorts of academic papers that I’d read in this form.

3. Most of the books I really want to read are not available yet.  That’s a hard problem.  Many of the books I want to read are already on my bookshelf.  I don’t like the idea of having to buy them again, but for some of them, it would be worth it if they were available.  But they aren’t yet.

4. Books are expensive.  You can’t by eBooks secondhand.  I used to have a policy of buying any book I wanted, new, because I love knowledge and the books represented something personal to me.  I stopped doing this when I started having problems managing my library, and the ratio of unread/read books was heading towards a limit.  Now I buy books 2nd hand and don’t think of holding onto them.

I’d buy one today if I could pay 99¢ per chapter!

There are a lot of books that are not to my liking.  I’d like to be able to read a part (for pay) but then not pay for the whole book if I’m not going to read it.  I’d be reading all the time if I didn’t have to pay for material I wasn’t reading, and I suspect this is true for a lot of people.  When I buy a book, I’m not only deciding whether the topic and author are interesting to me - I’m deciding whether to take the risk that I won’t get value for money out of the purchase.  

If Amazon took this risk out of the equation, I suspect they’d do a lot for reading as a whole.  They’d help to improve the market by channelling more money towards books that actually get read as opposed to poor material that has a lot of marketing behind it.

It could be made transparent - as you turn to the second page in a chapter, you get billed for it.

I am sure existing publishers and authors would resist the idea because it would disrupt their models, but Amazon could offer it as an option for self-publishers and if it proved effective (as I suspect it would) they would be forced to adapt.

There are too many icons on my iPhone already, and I expect to end up installing more applications and web-app links.

To make it easier to use, I’d like Apple to make Springboard context aware. Specifically, I would like to be able to choose some of the rows of the grid to show a prioritized list of icons of applications that I last used in that context.

What would ‘context’ be?  It would be a combination of location, time of day, and day of week.  There might be other factors, such as previous application run that might be worth taking into account. How would the context be narrowed?  

Probably the system could actually work by constructing a bayesian network based on the previous uses of each application associated with those variables.  It wouldn’t actually identify contexts per-se - it would just list the applications you’re most likely to use at that time and place in reverse order of probability.

“I’d like to give a site access to my Google contacts, but I don’t want to give them my google username and password.”

OAuth solves this and similar problems.

OAuth allows one website to access a users content from another website without needing their username and password.  In OAuth terminology, the site that wants to get the content is called the consumer, the site that has the content is called the service provider and the content is called protected resources.

A concrete example (taken from the OAuth site) is of a user who wants to use a printing service to print photos they uploaded to a photo sharing site.  The printing service is the consumer, the photo sharing site is the service provider, and the photos are the protected resources.  If both sites support OAuth, the user can give the printing service access to the photos without giving it their username and password to the photo sharing site.

To do this, the user makes a choice on the printing site to access the photos from the photo sharing site.  The printing service generates a request and sends it to the photo sharing site.  The printing service then directs the user to the photo sharing site where the user authenticates themselves and approves the request.  The photo sharing site then provides a token to the printing service which it can can use to authenticate requests for the photos.  To reduce the chances of abuse, the token is typically only honored for a limited period of time.

A more detailed introduction can be found here.

I spent some time today having syntax problems with scala function by-name function declarations. It turns out that a space after the colon is critical. In the end someone on the scala irc channel at irc.freenode.net. by the name of ‘mapreduce’ helped me out (thanks!).

This session should clarify what works and what doesn’t work:

scala> def some1 [T] (interval :int) (f :() =>T) (noContent:T) = {}    
some1: [T](int)(() => T)(T)Unit

scala> def some2 [T] (interval :int) (f :=>T) (noContent:T) = {}   
<console>:1: error: ':' expected but identifier found.
 def some2 [T] (interval :int) (f :=>T) (noContent:T) = {}
 ^
<console>:1: error: identifier expected but eof found.
 def some2 [T] (interval :int) (f :=>T) (noContent:T) = {}
 ^

scala> def some2 [T] (interval :int) (f : =>T) (noContent:T) = {}
some2: [T](int)(=> T)(T)Unit

scala> def some3 [T] (interval :int) (f: =>T) (noContent:T) = {} 
some3: [T](int)(=> T)(T)Unit