Right now, the Motorola A1200 Ming currently has my vote on the best handheld available. However, this does not mean it couldn’t be better. In fact, I personally think that these devices are moving at a crawling pace, often in observance of market interests. Regardless of the economics, the tech of the One Handheld to Rule Them All ™ is a simple concept - converge all the small devices you could possibly think of. There must be some considerations made for cramming this into one device, and how it will work out. Ultimately, it would be One Handheld to Rule Them All ™.
..:: Interface
As I pointed out, the Ming uses a touchscreen interface that changes depending on which mode is being employed. This is currently my favorite interface, although some have expressed disinterest in the lack of tacit response. In other words, they want a keypad that they can feel so they can use it better while driving. While it takes a lot more skill to “know” where a button is on a flat touchscreen, as a matter of public policy, people also shouldn’t be texting while they’re driving (and this may be an indirect method of reducing this behaviour, but could also create more dangerous texting, as people mght completely ignore their driving to send a text).
..:: Cell Phone
Use GSM or some other world compatible set of bands. This is a given, and a pretty simple concept. However, cell phones haven’t seemed to add anything to the actual phone aspect since I had my first erricsson back in 1999. It was a monochrome, 1-line display, but it had voicemail, caller ID, call waiting, and conference call. Here we are, nearly 8 years later, and nothing has changed except adding more transmission frequencies, customized ringtones, and identification images. To remedy this, I have a few ideas.
Voicemail listings: One thing that current cellphones do not do well is manage voicemail. You get a recent/missed call list, a notification that you have voicemails (some phones will tell you how many there are) and that’s about it. The voicemail system is rather embarrassing. It should be more like a voice recorded message sent directly to the handheld as a file. In the user’s list of missed calls, ones accompanied by voicemails would have a voicemail icon, and the user could easily listen to and manage any voicemail independent of the others. This concept is roughly taken from Vonage’s voicemail system, where users see what caller left messages.
Conversation Recording and Contact Sorting: Another failure of innovation in cell phones is recording. Why do no phones (or very few) support conversation recording? With decent compression technologies, spoken voice quality recordings could be made with infinitesimally small space requirements. This is where my favorite idea comes in. Gaim supports chat logging. Also in gaim, the user can set it so the buddy list automatically sorts itself by log size. In other words, the people you talk to the most float to the top. Applied to cell phones, the people with the longest total recorded time would float to the top. Since the recordings would be saved in a nice open format (lossy are alright, but ogg/flac anyone?), they could be exported from the phone to a computer for editing or archiving. When exporting, some sort of micro-database would store the record length so user’s contacts would maintain their total talk time even if the recordings were no longer on the phone. Even if recordings weren’t enabled, call logs would monitor durations, thus enabling the smart sorted contact list. Implementing this should be possible on any open phone. The logging would also enable advanced call time monitors. Currently, monitors only check minutes per month and minutes since the last reset. The monitor does not discriminate between unlimited nights and weekends, unlimited mobile (or specified friends, depending on your carrier), and peak hour calls. However, the monitor should monitor all three, all separately.
Recording Protections: What good is all this recording, if someone can steal a phone so easily? This is where some sort of encryption system is needed. A user could create some password of a hefty length (”The quick brown fox jumped over the lazy dogs”, or a voice signature maybe?). This phrase would be used to create some encryption key to allow exportation and decryption of the recordings. The key would not have to be entered as long as the phone is used once every n hours. Regular users could set it to 12 hours (16 might be better - some usage statistics would best measure this), and for the most part never notice the problem. High security users could drop it lower, or disable recording entirely, while low security users could disable the recording protection system entirely. This is effectively an open DRM format, but with security in mind rather than anti-competition and profit insurance. Implementing this should be possible on any open phone.
Signal Notifier: This is a simple software based module, somewhat like an alarm. The idea is the user sets an alarm not based on time, but when a signal reaches a desired level. This would be incredibly useful in times when out of major metropolitan areas.
Update: some of the newer motorolas have this feature. My dad turned it on, and after being out in the mountains of Virginia for a few hours, he couldn’t stand it anymore.
..:: Syncing
Sadly, SyncML can hardly be called a standard, or at least implementations of it have been notably ineffective. My comparison of Motorola PhoneTools vs ActiveSync is like a manatee trying to take out a leviathan. SyncML should be easily usable on any system, regardless of whether it’s Windows, Linux, or Mac.
Hashed Playlist Synchronization: Playlists and media should also be easily synchronized. While support of DRM (like Apple’s FairPlay) is obviously a nicer feature, DRM should never be mandatory (like it is on the Zune and iPod). Under no circumstances should the media format be altered or further limited by the sync process. Why file paths cannot be transparently managed is inexplicable. Files can be identified with hashes and other algorithms. A regular playlist could be able to be converted into a hashed playlist, and then when transported, the files could be transferred and a playlist on the handheld automatically generated from the hashed playlist. This would innately and transparently handle playlists where some files are already on the handheld, in addition to basic pathing issues.
PIM Synchronization: Four methods of synchronizing personal information are easily able to be designed: PC overwrites handheld, handheld overwrites PC, ask on conflict, prioritize latest. The first two are simple; one system simply overwrites the other. The third detects that two entries are conflicting, and asks the user to choose one to remain. Taking a feature from Wikis, the last n changes should be able to be undone (changeable at a user’s preferences). The fourth is a simple concept: with a Wiki style transaction manager, timestamps of changes are stored. Whichever timestamp is the latest prevails in the change. Synchronization should also be possible over any supported method of wireless data transmission.
..:: Personal Information Manager
Also known as a PIM, this is an entirely software based system that manages contacts, events, tasks, and calendars. Windows Smartphone got the home screen right. It should show the current day’s events, the number of emails/messages unread, and the current number of unfinished (and overdue) tasks for the day. Contacts should be able to support multiple numbers (business, home, mobile, fax, etc) for contacts, as well as multiple addresses (business, home, mailing). Nextel’s had the idea right with the dialing interface. When the contact list pops up, each name has a small icon next to it. The icons would be for business phone, home phone, mobile phone, email1, email2, etc. The user presses up and down to scroll through contacts, and right/left to change the number with which to reach a contact.
Contact list systems are neither new nor complicated. Look at gaim which supports sending instant messages to cell phones as SMS/text and also supports AIM, ICQ, MSN, Yahoo, Gtalk/Jabber, IRC, and then lesser publicly known systems like groupwise, SIMPLE, and sametime. It’s easy to set a single contact up that can have multiple screen names on different services and even cell phones. If my brother is away, my message, by default, will go to his cell, but I can right click to bring up the other contact methods. And, because of the contact list sorting and logging, it’s on a contact basis, not a screen name basis. Additionally, the names that I see actually mean something - not useless childish information like flowergirl123 or hondaspeeds2000. The handheld should be able to use these methods just as transparently as gaim does.
..:: Media Player
The handheld should have a proper media player. Windows Media Mobile and RealPlayer mobile both lack many basic features found in top tier PC media players. The media player should automatically generate smart playlists (Most Played, Never Played, Newest, etc) like the top tier media players (Windows Media Player, Itunes, Amarok). Further, the library should automatically scan (at some interval) for media, hiding songs in playlists no longer on the handheld. The Media Player should be able to utilize an available internet connection to play (and record) streamed internet radio.
When playing video, the handheld should be able to rotate to play on the screen in a landscape mode. The video should support subtitles from separate text files, were the file lists the time and text to be displayed. This is really all software based. The only required hardware for a media player is a 2.5mm headphone jack and an internal speaker for using without a headset. 3.5mm headphone jacks are alright, but they’re also larger, and every millimeter counts, especially in the One Handheld to Rule The All ™. When receiving a call, the handheld should switch to call mode, pause the music, and give the user the option to take the call. When finishing the call, the music should pick right back up.
Network Streamer: The handheld should also run a webserver. When at a laptop or PC on the same network, one could visit the handheld’s IP address and stream audio/video to the computer, like gnump3d or icecast. The software could be further adapted to use the playlists on the phone, and actually use a flash interface for those playlists. This could also be transmitted over bluetooth, but it would not be effective over GPRS or EVDO/EDGE as those services are typically too expensive. This is entirely a software feature, as long as the phone has a wireless data connection.
..:: FM Radio
Simply put, make it digital. It’s important that the frequency can be altered to the 5 hundredths place. In other words, the handheld should be able to reach 123.40, 123.45, and 123.50. There are many cities (particularly NYC), where there is band overcrowding, and having the hundredth’s place on a digital tuner really matters. Other than that, multiple presets, and the ability to record radio to a nicely editable format would be nice features. With all the DJ’s talkover on radio, it would be unlikely that users stop purchasing music.
..:: Applications
Most applications for handhelds are pretty worthless. There are loud annoying games with horrible controls, and plenty of document viewers, but very few compatible editors. This requires multiple things:
Open System with SDK: without an SDK, applications must be written as java midlets. Unfortunately for Sun, the concept of “compile once, run anywhere” just doesn’t hold water. Screen sizes, input methods, and connectivity options are all different. Since Palm was introduced almost 10 years ago, closed development environments have been a clear issue in the handheld industry. There was tons of software to be tried, and very few, if any buying it. The same still holds with java midlets for cell phones. With an open SDK, we would see the spark of open source genius that has resulted in the Linux distributions, and various tinker projects like Xbox Media Center, and Linux on PS3/Wii, various firmware customizations for multiple mp3 players, and even emulators. Connectix was a group that made a free playstation emulator. Eventually, Connectix became a company, and Microsoft bought them out, using their knowledge in developing Virtual PC. Open software is a necessity to spark growth.
Package Manager: Having something as deep as apt or portage would be nice, but neither repository system appears to have a method of maintaining commercial software. Because it is laughable how outdated deployed software gets, the need for extensible update daemon (via APIs) is incredibly necessary. This is a completely software driven module, but requires some sort of data connection, preferably wireless. Software installation should be as easy as Ubuntu’s Synaptic makes it. When a user wants to install commercial software, the commercial software would invoke the update daemon’s API to add itself to the commercial list (which sandboxes commercial apps from editing other competitors update entries, and free software entries).
..:: GPS
Both the Neo 1973 and the Mio A701 have GPS. The hardware required, is obviously, a GPS module. The software should include a course plotter (where different courses can be chosen - that always bugged me about mapquest) and voice directions.
Handheld Locator: when the phone has an accessible Wifi connection (or if the user enables the GRPS/EVDO/EDGE connection) the handheld could wget on a timed interval some page on a remote webserver with it’s phone ID and current coordinates. A parent could track their child, or the user could use the Locator to help find the handheld when lost. When camping, hiking, or traveling in general, one could use this as a safety feature to let friends and family keep tabs via the GPRS/EVDO/EDGE signal. While you might think that a user would just make a phone call, the Locator would be automatically pushing the coordinates, whereas making a phone call physically requires the user to check for a signal.
..:: Memory and Expansion
Personally, I would prefer a 100 GB microdrive like iPods have, even at the cost of no longer having some sort of SD slot. Transflash is nice and all, but the 2 GB limit, nowadays is just unacceptable. With a large microdrive, the possibilities would be endless. Conversation recording would be incredibly easy, and users could carry around entire digital collections. Add this to the network streamer system in the media player, and the handheld would be a total portable media center solution.
..:: Camera
The camera is a toss-up. Any successful handheld should be offered in two models - with and without a camera. Many business users stick with Blackberry because of the lack of camera. This is because many government buildings (especially courthouses), and many high security corporate offices do not allow cameras.
For the camera embedded model, the system should be able to take stills up to 4 megapixels with a flash and at least 4x optical zoom. Digital video recording should also be available, with at least 720×480 resolution. There should be no hardcoded time limit on the length of the video. Most video recording devices that are not specifically cameras have an automatic hardcoded cutoff time - anywhere from 30 seconds to 3 minutes. For some devices, this is a technical issue - the early Canon ELPH series supposedly didn’t have enough ram to continue 640×480 video for more than 30 seconds. However, we are now multiple generations descended from those early technologies, and these basic issues have been solved. The truth of the matter is that manufacturers do not want devices cutting into other single-task devices’ sales. Video editing on a handheld usually is a laughable task, but mostly because the device does not have much graphic space. The camera itself is mostly hardware based, and the software is rather simple ports of digital camera firmware.
..:: Wireless Data Transmission
Delivered data is the Next Big Thing ™ and it comes on multiple fronts: EDGE/EVDO/GPRS, Bluetooth, Wifi (802.11a/b/g/n), and IR.
GRPS is the wireless cellular internet that has been around for quite a few years now. It’s embarassingly slow, and incredibly overpriced. EDGE and EVDO are the telecoms’ next generation of wireless cellular internet, and are effectively wireless broadband. However, the wireless broadband connections (regardless of which network it is) are grossly overpriced. Many run about .2-.5 cents per kB. Downloading a single 3MB mp3 will run you $6-$15. That’s just outrageous. Even worse, Verizon’s EVDO terms of service are even more atrocious. Verizon will actually drop heavy broadband users who have unlimited plans. Additionally, the EVDO service agreement specifically says that the service cannot be used for “uploading, downloading or streaming of movies, music or games, … server devices or with host computer applications, including, but not limited to, Web camera posts or broadcasts, automatic data feeds, Voice over IP (VoIP), automated machine-to-machine connections, or peer-to-peer (P2P) file sharing.” That covers just about every bandwidth intensive application there is. Apparently, Verizon only wants you to check your email, and I can’t imagine Cingular’s limitations being any different.
IR is not exactly a necessity, but it would be nice to be able to get rid of the 47 different television remotes and have one device to handle it all. Preferences for television (like channel listings) could be kept on the handheld, as televisions typically do not have the stored memory for such an application. Profiles for different remotes could be kept, so that the interface would remain nearly the same but the different TVs would all use the same handheld. This is easy as there are already learning remotes which can memorize IR frequencies using a basic capture and imitate sequence.
Bluetooth currently suffers from a serious case of carrier-fearus lockdown-itis. Most carriers lock out all Bluetooth functionality except use with headsets. This is fairly well intentioned, but horrible for innovation. The idea is that if a users’ handheld had unfettered Bluetooth usage coupled with user accessible storage space, simply going to school, the mall, or any public meeting place would be like walking into a private music piracy party. However, this lockdown has also come greatly at the cost of innovation. Many phones are unable to even sync with computers via Bluetooth solely because the carriers disabled this function. Stifling innovation to safeguard the recording industry’s profits is a lost cause.
Wifi (802.11a/b/g/n) is the nicest and cleanest feature to have. Use the internet anywhere you get a wifi signal. Most universities have blanket wifi networks, and many apartment complexes have them too. Even many home users have wifi. Here’s the problem: carriers don’t want wifi. Wifi means users can get on the internet without being subject to astronomical data charges. Additionally, users can easily navigate to free websites that give out games, ringtones, and even full mp3s. Further, Skype charges about 2 cents per minute, and often runs promotions to purchase credits below their normal cost. Users with wifi could cut the carriers out of a substantial amount of revenue if more voice over IP (VoIP) was used. My plan currently would not benefit from VoIP usage, as I get 450 minutes, unlimited nights and weekends, and unlimited calls to Cingular customers. Although this is the cheapest plan Cingular has, the majority of my calls are made using the unlimited call methods, and I rarely break even 100 minutes in a month in the peak usage minutes. However, many users who have $70 or $80 plans could cut their plan and use VoIP if they had wireless access points. At 2 cents per minute rather than 30, many would save ridiculous amounts of money. Even nastier for the carriers is voice chat, offered by Google Talk and Skype. This is internet-internet calling and both companies offer this for the unbeatable low price of no-strings free. This also means lost revenue in text messages, as users would no longer pay 10 cents per text message. Instead, users would just import their AIM, MSN, or Yahoo buddy lists. Checking one’s email remotely would cost the user nothing, and this means lost revenues for the carriers. All in all, the wifi is the carrier’s bane.
Wifi Profiles: WinXP does a very good job with this, and unfortunately, Ubuntu’s wireless manager is not as good. WinXP automatically picks up a wifi network, and connects the user to the top available network that a user has prioritized in the profile list. This process occurs transparently, and requires no user interaction. This is completely software driven, and only requires the wifi module.
Network Auto-switching and Transparency: The user should be able to specify network priorities. For instance, I would specify Wifi: Preferred 1, Bluetooth: Preferred 2, IR: Preferred 3, GPRS/EVDO/EDGE: Request Confirmation. In other words, the data transmission should occur transparently regarding which network interface is used. Data is data is data - it doesn’t care how it is transmitted. The only ones who care about how data is transmitted are the carriers, because they can charge to be the gatekeepers. Without that expense, users have no need to care about how data is transmitted. The PIM should list Google Talk and Skype screen names, and when viewing a contact, if the caller is online, the system would check whether the contact is available on one of the two services. This could even handle chat over AIM, MSN, or Yahoo if available. When contacting a user, the network connection would be chosen transparently.
VoIP Manager: This is a completely software based system, and is used to log minutes. Since calls could be made using the GSM/CDMA or VoIP, the user should be able to explicitly choose which to use, and but also prioritize systems for different rules. For instance, if the time is between 8pm and 8am, or it’s Saturday or Sunday, use GSM/CDMA. If the target contact is a free mobile-mobile user (like unlimited Cingular-Cingular, or a Verizon user in the IN network, or on the Tmobile 5 list), then use GSM/CDMA. If a voice chat contact is available, use VoIP. If the current month’s limited times call duration is over n minutes, and a wifi network is available use VoIP. This would all occur completely transparent to the user.
..:: Closest Contenders
There are currently no devices that meet all of these specs, and not even one that comes close. Although the state of handhelds is pretty sad, at least it is moving forward. Here’s a list of the current closest contenders:
The Motorola A1200 Ming is a Linux based handheld with a 2mpx camera, quadband phone, media player, Bluetooth, FM radio, touchscreen interface, USDC expansion, and PIM, but due to Motorola’s intent on keeping the system closed, the extensibility is definitely below par, and software support is dismal. The syncing options also are quite restricted, and the handheld lacks wifi and GPS.
The OpenMoko Neo1973 is yet another Linux based handheld slated for release in March 2007. The initial version has multiple confirmed features: GPS, touchscreen interface, media player, USDC expansion, PIM, quadband phone, Bluetooth, package manager. The second generation is rumored to add Wifi. The big thing about this handheld is that it’s an open Linux format and it comes with an SDK. It’s rumored to be priced around $350. This handheld is incredibly anticipated in the Linux community.
The Nokia 770 is an internet tablet which has wifi, Bluetooth, and a media player, but it lacks a basic GSM/CDMA phone connection, GPS, a camera, and it is rather unwieldy.
The Nokia N90 has a 2mpx camera with flash, a PIM, a media player, and Bluetooth, the phone lacks GPS, FM radio, and is only tri-band. Additionally, the N90 still uses a keypad, and the video recorder only captures at 352×288, a small and nonstandard resolution.
There is also the massive class of Blackberry clones like the Nokia E62, Motorola Q, and the Treo series, but these all suffer from the same things: mini-qwerty keyboards are annoying because they take up a ton of space, and are only rarely used. The buttons are too small and there are too many to use without looking, so the tactile response argument goes out the window. The qwerty keyboard would be better replaced by a few square inches of touchscreen. Additionally, none have wifi, GPS, and many have substandard media players, gimped by slow processors.
PDAs often have wifi modules, but they’re generally larger than phones, and lack GSM/CDMA cellular phone calling.
..:: Conclusion
I currently have the A1200. However, if I can get my hands on the Neo1973, I definitely will get one. I would love to write all (or most) of this software for these phones, but development on closed systems is just incredibly difficult. Not only is the documentation simply unavailable, the handhelds have “security” features that are constantly and repeatedly working against the developer. Hopefully, the Neo will change that.
