Headworx

Headworx is a collection of brainstorming ideas and thoughts on technology. Most are inspired by a group of friends of mine and many interesting things I come across everyday.

Subscribe: [RSS Feed]


Twitter updates

    Wednesday, August 20, 2008

    Predictive Mobile Navigation


    The vacation summer season is over again. I really enjoy traveling and it happens we do travel quite a bit by car in June-July-August season. Looking around it is hard to find a car without a portable navigation system these days. Some would say we have just passed the GPS revolution. But really? Surely having a nice map guidance in a car is a good thing. And as the roads are more and more crowded, electronic maps help detour traffic jams. Actually I use two devices together to make it through.

    The first one is from the analog era - a CB radio. Extremely popular in Poland now, helps me learn what lurks before me. A traffic jam, a speed camera or a radar / laser equipped police officer hunting for a pray. Recently I was contemplating what would it take to recreate the CB Radio experience in the digital era we live in. An ad-hoc, peer-to-peer social network with alerts, chat rooms, where people you are connected with are selected on a location basis - the ones within a 10km range around you. We tried to approach this problem the "digital" way with one of our "digital" customers not long ago and we failed. Neither Facebook nor Google can do in digital domain what CB radio is doing in the old analog domain. Plus CB radios are not computers, so they cannot execute any code (read: virus) changing their behavior, so the huge security issue known from the Internet world does not simply exist. Plus the devices are so easy to operate. Turn on, select your chat room with a knob (a channel selector), set the volume and press a button when you want to send a message (speak to a microphone). Amazing...

    The second is the Garmin GPS unit built in the dashboard of my Subaru. Subaru sells cars equipped with a big navigation screen but they do not have maps covering Poland, so the system would be completely useless, if not one smart guy in my local dealership - he managed to connect an off the shelf Garmin unit to the Subaru touch screen. The benefit of this experiment is the USB cable hanging under my dashboard. I can take any Garmin map software - be it City Navigator NT or City Navigator North America or GPMapa for very detailed coverage of Poland. Good electronic map is 95% of a GPS unit. The rest (device, its design graphics, gadgets) is just 5%. With poor map software and latest and greatest hardware you will get nowhere.

    From the mobile navigation point of view, CB combined with GPS makes a system somehow similar to a GPS equipped with a Traffic Message Channel receiver, provided a TMC system is operational in the area you are traveling. But believe me, this all what people call "revolution" is just the beginning. We need one enabler to carry us to the new navigation revolution and it is the feedback channel.

    A GPS unit knows a lot. Where you travel, how fast, your preferred types of routes (be it highways or off-road trails). It knows where you are and where you are heading (you punched in the destination address, didn't you?) and your via points. Imagine now what would be the collective knowledge of all the GPS navigation systems along the route you have just planned? Imagine a business intelligence database collecting all the statistics from every car on the move. And possibly every car that will be on the move tonight or tomorrow... with their preferences, destinations and via points. Such a system would be able to accurately predict where traffic jams will happen and SUGGEST to you taking an alternate route or alternate time... Two weeks ago a friend of mine covered the route from Gdansk to Krakow in 19 hours. Following her exactly 24 hours later I covered the same route in 8 hours.

    Google already tries to do something in that direction, with the [Traffic] option on their maps.google.com service. Today what they have is only based on historical data. May be in the near future they will derive some information from the queries people enter on their web site. But this is just the infancy of what will come when the technology matures and gets widespread.

    I do not believe we are able to get many new roads and highways anytime soon. And there are places we will never get them, as the areas (think: Paris or New York) are already saturated with road infrastructure. But I do believe there are many people who foolishly follow each other like Lemmings to the Sunday - afternoon "back home" traffic jams. They would really love to see the REAL predicted time of their journey based on how many others are planning to be on the same road at the same time. And many of them would choose to stay an hour or a day more to avoid all the hassle. Empty roads are a scarcity. Computing power and machine - to - machine communications are in abundance. We should utilize this abundance to improve the quality of our lives. Predictive Mobile Navigation will be just one of the future developments improving this quality. And this will be an important market. We love intelligent gadgets and we do not mind paying for them.

    Labels:

    iPhone 3G, Cell Breathing, And Intersystem Handoffs



    So everybody is talking now about the dropped calls on the 3G iPhones. Well... I wrote a piece on this subject two years ago: http://headworx.slupik.com/2006/11/disconnected-umts.html. So let us try to understand once again what is happening.
    1. 3G iPhone is a UMTS device. This measns it works on 3G/UMTS networks known as WCDMA networks (Wideband CDMA or Wideband Code Division Multiple Access). WCDMA is an evolutionary generation step for GSM networks, but the two standards differ so much, this can hardly be named evolution. It would be like saying iPhone is an evolution of a rotary phone. Sure, both allow people to talk to each other, but are so far away technology - wise...
    2. WCDMA networks generally have much smaller cells. This means the distance between the mobile terminal (iPhone) and the base station is relatively short. And this means there have to be many base stations to cover a given area. Many more than 2G GSM base stations. Network planning has to be very careful, especially in urban areas, where buildings greatly affect the coverage. The cells created around base stations must overlap each other. This means when a phone is moving out of the coverage of one base station, it has to be moving in the range of another base station. This process is called a handoff and is relatively seamless between two 3G (or UMTS) base stations. Or at least it should be seamless.
    3. The trouble starts when a terminal (like an iPhone) has less sensitive 3G/WCDMA radio than assumed by the network planners. There is an unconfirmed rumor iPhone has some problems in that area. iPhone prototypes the networks were tested with, supposedly were more sensitive compared to production units. This means they passed network interoperability tests but now fail to comply. With less sensitive antenna / receiver, the iPhones are simply falling out of one base station coverage before they are able to attach to another. The reason may be somewhere else. There are opinions the problems originate in immature 3G chipsets inside iPhones. If it takes too long for the Infineon chipset that powers the iPhone to negotiate with the second base station, the connection may be dropped because the user moves out of the coverage of the first base station.
    4. There is a chance the problem has always been there but simply it was too difficult to prove any correlation. Before iPhone there was not a single model of a 3G phone so closely examined by the community and the analysts (ans lawyers!). But let us consider the facts:
      - There are many more handoffs in 3G (due to smaller cell sizes) than in 2G
      - Each handoff may fail (simply because the cell the user is moving into is already full and cannot accomodate another extra call)
      - The coverage of 3G base stations is less predictable due to the characteristics of the frequency bands of 3G networks (especially the 2100MHz that has so many problems penetrating concrete walls, so you may experience multiple back and forth handoffs simply walking around a room in a building, sometimes moving closer, sometimes further from a window)
    5. The above means iPhone users are very likely to experience so called inter - system handoffs, or handoffs from 3G/UMTS/WCDMA to 2G/GSM networks. Such handoffs are very likely to fail for several reasons:
      - First: the capacity of a 2G cell is usually much smaller than the capacity of a 3G cell, so simply there may be no free slot to accommodate new user.
      - Second: usually the 2G cells are part of a 2G sub-network and 3G cells are part of a 3G sub-network, and the entire conversation context has to be moved from one network to the other. The success rate of that handoff is never 100% and varies depending on the infrastructure equipment vendor. Nortel and Nokia are generally considered poor in that area while Ericsson and Huawei usually work fine.
    6. Intersystem handoffs are even more difficult when there is an active data connection (in parallel to the voice connection). Even more. Intersystem handoffs fail more often when there is data session in the background. Now imagine, due to its functional design, how likely it is an iPhone has an active data session during a call?
    7. And if that was not enough, WCDMA cells are prone to the "breathing" phenomena. A cell may contract (means: drop the users at the edge), when the overall signal-to-noise ratio gets worse. And this may happen just because another handset / terminal enters the cell.
    Considering all of the above, I would not count on fast remedy Apple may have to the problem. Dropped calls are the inherent nature of 3G networks. As long as the 3G cells are not overlapping each other by a wide margin (minimizing the number of handoffs and eliminating intersystem handoffs all together), we will be experiencing dropped calls. Apple has probably done nothing wrong here, they just brought everybody's attention to the problem.

    Sunday, August 17, 2008

    ActiveSync For Blackberry?


    The second week of my holidays is just about to end. Of course these days holidays are really on-line/connected holidays. No broadband yet, as spending a quality time in remote, very little populated areas leaves me with just a bar or two of GPRS/EDGE coverage. I somehow anticipated that, packing a 3 meters long USB extension cord in my luggage, so my Option USB 225 cellular modem could be fixed to a long wooden stick with a duct tape, and placed outside the window in order to catch the precious data packets. I was not able to make or receive phone calls (attaching my mobile phone to the same wooden stick did not work), but s-l-o-w d-a-t-a c-o-n-n-e-c-t-i-o-n was working filling my Inbox and emptying the Outbox.

    On the other havd there were days when coverage was working only on the beach and for obvious (sand!) reasons I was not willing to bring the laptop with me. So the Samsung took over the communication tasks. Samsung is a great piece of phone. It has its drawbacks (sound quality is average at best and display is hard to read in full sunlight), but application - wise it does the job. Business keeps on going and our company Exchange Server 2007 keeps on pushing messages via the ActiveSync protocol over the air. During the long strolls along the beautiful and almost virgin beaches of the Słowiński National Park, I contemplated the options of replacing the aging i600. Nokia E71 is one of the options, but still there is the open question whether it supports encrypted messages (during my short tests I failed to read the encrypted mail message despite successfully installing the personal certificate containing my private key). The new iPhone is more of an unknown, as the encrypted s/mime messages are not tested yet (imagine that I have not had my hands down on the new iPod yet!) and I am really not sure how far the virtual keyboard will take me into the irritation and frustration land.

    Then there are the BlackBerries. The perfect devices at all angles. But I have ignored them so far for two reasons. No 3G/UMTS and no direct ActiveSync support. The first problem is solved
    now with the arrival of the Bold / 9000. The second one could have always been solved by purchasing the BES (BlackBerry Enterprise Server) or signing up for BIS (BlackBerry Internet Service). Both expensive (very!) and unnecessary (from the technical standpoint) components. We (as many small and not so small companies) are running the Microsoft Exchange 2007 server to handle all business emails, contacts, calendars and more. The Exchange has very simple an reliable way of exchanging and seamlessly synchronizing ActiveSync - based mobile devices (all Windows Mobile - based, Nokia smartphones and iPhones with v2 version of the software). One day I imagined the Blackberry Bold running ActiveSync. yes, this would be the perfect match. The best business mobile device on the Planet working direct with my company infrastructure over plain, universal data access (be it GPRS or EDGE or UMTS/HSDPA or WiFi). The missing link? The software. BlackBerry makes a lot of money selling the servers, so they will not allow their hardwer to bypass their servers. What about the third parties then? Googling a little I found this: http://www.astrasync.com/. Looks promising:

    AstraSync establishes a direct https connection between the BlackBerry® Smartphone and the Exchange 2007 or MailSite Fusion server. Customers using AstraSync do not need to purchase a BlackBerry Enterprise Server® and are not dependent on the BlackBerry Internet Service.

    But will it deliver? Well... I will be happy to report to you... But this so important piece of software is not available yet. I signed up with my real email address to be notified. And I am waiting for the story to unveil... Hope they make the encrypted messages right. Otherwise I may stick with my Samsung...

    Labels: ,

    Sunday, August 10, 2008

    The FujiFilm, The Samsung and The Nikon


    It is holiday time again in August and holidays mean taking a lot of pictures. What a change in just a few years. Now we all travel with digital cameras, flash memory cards, chargers and laptops. Compare that to taking a few rolls of film four - five years ago. I was late to the digital photography, as I believed the final results were far off compared to analog films. Then in 2002 I bought the 4-megapixel Minolta F100, my first digital camera. And suddenly I realized there is no way back to analog.... Surely the quality was not there yet, but the flexibility and instant preview of the results meant better pictures and much faster learning process (the digital feedback loop is a few seconds - you shoot and you see the picture versus days / weeks long analog feedback loop). There was only the way forward - to improve the quality of digital pictures to match and surpass analog.

    Chasing the quality I soon realized the battle is not of megapixels. I no longer make prints. I store my photos on a media server and either have them cycling on several digital photo frames around the house or have the PlayStation 3 play them on the FullHD TV screen. FullHD is 2-megapixel image (1920x1080). So 6 megapixel file is more than enough, even for 50-inch. The difference between 6-, 8-, 10- or more is simply not detectable by human eye. But there is more than resolution that meets the eye. The most visible difference between the old "analog" and the new "digital" pictures is in the dynamic range, or transition from dark to light. Digital sensors measure light by capturing photons. Each "pixel" captures them. As cameras tend to be smaller and smaller and have more and more pixels, the pixels are shrinking. And so is their ability to capture photons. Imagine the extreme model: a pixel so small it can capture just one photon. No photon captured it is black. A photon captured it is white. No differentiation... no whiter shade of pale :). This is what happens with more megapixel - packed tiny, sleek and sexy digital cameras. The dynamic range of their output is shrinking.... There are some basic fundamental laws of physics that rule this world. And no electronics nor software tricks are able to get around them. So is the nature of digital photography. Pixels have to be large to get good, analog - beating quality. This means the trend reverse in nature to what we have now. Big sensors (meaning big lens) and fewer pixels. Is there a sweet spot? A balance of physical dimensions, resolution and dynamic range quality? Sure, but it is different to each of us. Myself I found that balance a year ago in form of the FujiFilm FinePix S5 Pro. The FujiFilm uses a special kind of sensor with two photo diodes capturing each single pixel. One captures the darker part of it, the other the brighter. The result is one of the best dynamic ranges on the market. I love the pictures it takes. But it is large and heavy.

    So recently tempted by the advances inb microelectronics, I purchased a second, point-and-shoot camera. The Samsung NV-24HD. The Samsung seemed to have everything. Wide angle zoom (24mm), OLED display shining in full sunshine, ability to record HD movies at 30fps. And very cleverly designed transfer / charging USB cable. But unfortunately after two months I realized I do not use it at all. The pictures just do not deliver. I mean they are not bad, at least on par with many other small cameras... but after the FujiFilm... they are pale.

    So last week I decided to swap the Samsung for the Nikon D40. The D40 seems to be great design by Nikon. As inexpensive as a D-SLR can be (around $400 including lens), it is - to quote DPreview - "perhaps one of Nikon's most important digital SLRs":
    It's certainly their smallest and lightest, their most affordable and ships with a fairly decent kit lens too. (...) In everyday use the D40 is just what it set out to be, a very capable, compact, lightweight and easy to use camera which makes a perfect first step for anyone wanting to get into digital SLR photography. It provides enough control and a large enough range of manual settings to enable you to experiment and learn but also helps you to take great pictures in the process. It's one of those cameras you can just pick up and start shooting without fuss, that you can hand to a friend who's never used an SLR and know they'll be able to do the same.
    6Mp mean the pixels are relatively large to deliver. Being D-SLR type of a camera, it teaches the proper approach to photography from the start. I find it to be a great entry - level camera to introduce the beginners, and at the same time has a lot of potential to experiment and deliver high quality shots. In my particular setup, having both high - end FujiFilm S5Pro and the low - end D40, means compatibility of accessories - particularly flash Speedlights and lenses. Yes, it won't fit in my pocket, but being much lighter and smaller compared to the FujiFilm, it is a great second camera I won't be afraid to carry with me in many weird places or give it to the youngsters to play and learn.

    Labels:

    Sunday, August 03, 2008

    Global Village... Or May Be Not?


    At the beginning of the change there were jet airplanes. Letting us jump from one place to another in a matter of hours. Then automated long distance telephony services came along. It was enough to add two or three country code digits and we could here somebody on the other side of the ocean. Initially this was expensive. And these actions reaching the other end of the Planet were taken knowingly. You never boarded an intercontinental flight by chance. Dialing from Europe to Australia was never by chance either. But then mobile phones and the Internet arrived. Dialing a mobile phone you always dial the same number, regardless where the B-party is. Actually you do not know... And you do not pay any extra fees for talking to somebody who is traveling far away - it is the B-party who covers the extra roaming charges. By the way the roaming charges are relics of the past. On the Internet you never pay more for TCP packets sent across the ocean. Telephony operators use more or less the same networks that carry the Internet traffic. Yet a call from London to Tokyo costs more than a call from London to Bristol. On the Internet there is no difference where you are and where your packets are going to.

    Or is there?

    Technically the Internet is a uniform global network and nobody charges the traffic by the destination of transmitted packets. But for some services it does make a difference where on the globe the packets originate and end their trip. Yes, I am talking about music - related services. Take the Pandora radio, mentioned here a number of times. It is available only to the US - based subscribers. It means when you live in the US and your computer connects to the Internet on the US territory, they will let you in. But if you happen to be somewhere else, they will block you. But it does not make sense. Imagine you are a US citizen living in North America. And you are a paid Pandora subscriber. It works. Now you pack your laptop and go for a trip to London. There you arrive at the hotel, get in your room, connect to the complimentary WiFi network the hotel provides, press Play on your Pandora application only to get a message this operation is not allowed outside US. Nice, ain't it? But easy... this has not been invented by Pandora, but rather by the labels. Yes, the same people who introduced the region codes on DVD movies.

    I live in Poland and I am a paid Pandora subscriber. I often travel to the US and enjoy listening to this service, as it introduces a lot of new music to me (yes, very often that leads to CD purchases - fortunately CDs do not have the region codes). So what do I do when I am back home? Well... I pretend I am in the US. How? By using a VPN connection. VPN is like a tunnel. My packets travel in that tunnel to the US and there is the point where they enter the Internet. To the Pandora servers these packets look like genuine US-based packets, so they are treated accordingly. In return the music packets sent by the Pandora are addressed to a US-based address that is the address of an entrance to the VPN tunnel to my home. I have this VPN setup on my home router (not every device can do that, but the almighty Dlink DFL-800 can), so every host on my local LAN looks to certain selected Internet services as US based. Some may say this is not legal (I am not sure to be honest, but I feel I have the right to do this, as I pay for the services I use that way). But this just helps me feel equal with my US friends... and I really do love the Pandora, so rest assured I will continue doing whatever is necessary to have this service working down here in Poland. We live in a global village... or may be not?

    The second service in the "not available in your country" category is the iTunes. iTunes takes different approach. They do not look into the origin and destination of TCP/IP packets. They look at your credit card instead. You are allowed to get in only when your billing address is US-based. And this is not that easy to work around. Or is it? The iTunes store accepts gift cards. The Gift Cards are just codes you have to enter when you make a purchase. You pre-pay them in advance. My friend in the US can buy a $100 iTunes Gift Card and send it to me via email. Then I can use the Gift Card to make my purchases. What if I do not have any friends in the USA? Well.. there is always eBay or Allegro, its Polish equivalent, where there are plenty of Gift Cards available for purchase.

    The moral? We do live in a global village. So dear content and service providers, please stop wasting your efforts to convince us otherwise. The world is NOT flat...

    Labels: , ,