Author: Matt Croydon

Via NewsForge, Integration Developer News reports on a new WS- spec: WS-Discovery.  The spec is designed to make web services work among both wired and wireless devices on a local network.

When reading about WS-Discovery, I keep thinking of a similar technology that is already out there: Rendezvous/Zeroconf.  WS-Discovery seems like a very specialized solution that Rendezvous/Zeroconf may have already solved.  I’m sure that the way everything works is quite different, but on the surface it just “feels” the same.

I’m quite excited about what WS-Discovery might be able to do for mobile and context-specific applications.  For example:

mobile device: Hey!  I’m here!  What services are available to me?
WS-Discovery server: there is local access to the following services: foo, bar, baz.
mobile device: Sweet!  I know about baz.  I’m going to use it.  Thanks, WS-Discovery server! <cheezy grin>

Update: after some poking around, I managed to find the WS-Discovery spec at MSDN.  I already have a headache from a cold/flu, so I’m not going to read it in detail right now.  It looks just about as complex and painful as any other WS- spec that has come down the road.

Via Phonescoop, Cingular nabs AT&T Wireless for $41 billion:

Cingular Wireless LLC, a joint venture between SBC Communications Inc. (NYSE: SBC) and BellSouth Corp. (NYSE: BLS), announced today an agreement to acquire AT&T Wireless (NYSE: AWE), creating the premier wireless carrier in the United States. Today, the combined company would have 46 million customers and one of the most advanced digital networks in the U.S., with spectrum in 49 states and coverage in 97 of the top 100 markets. The combined 2003 annual revenues of the two companies would have exceeded $32 billion.

This really should be good news to everyone involved.  Cingular definitely seems committed to GSM technology, AT&T is the only one with EDGE nationwide.  With the combined capital and customer base, ATTular should be poised to take a jump at full-bore 3G as soon as it is an option.

I am a little worried about how this is going to affect roaming agreements with my carrier, T-Mobile.  I know that in several markets (San Francisco in particular) Cingular and T-Mobile share a lot of towers and coverage.  I’m worried about some of these roaming deals may dissapear.now that ATTular has the clout to do what it wishes.

It will also be interesting to see how the merger/takeover/purchase will affect phone and plan prices.  When my contract is up with T-Mobile, will it be worth switching over to them?  It would really take a lot to get me to switch over.  I’ve been nothing but impressed with T-Mobile since day 1.  Their customer support is quick and awesome, they don’t treat you like dirt at their retail stores (I’ve been treated very badly at Sprint stores and when asking questions at AT&T Wireless stores), and their coverage is good (but not perfect) and their billing is simple and consistent.

Once everything is figured out, this new company is going to be a force to reckon with.

Congrats on the press release, Russ and Wavemarket.

My body hasn’t quite figured out that it is now back in GMT-5.  Waking up this morning was hard.  Expect a slow couple of days while I decompress, digest, and recover from ETech.

Ewan and I managed to get a little touristy this afternoon.  We had lunch and beers at the Kansas City BBQ, also known as the bar from Top Gun.  They had great beer, excellent ribs, and everyone there was very nice.

We had to get the obligatory pictures, of course.  We even managed to leave behind a card.  Can you spot it?

I am sitting near the gate, next to a power outlet, and as close as I can get to the main terminal.  I am currently getting a weak signal from inside the United club, but still getting excellent throughput.  I snagged the non-dead tree version of this months JDJ journal, and hopefully I’ll have a chance to read it at some point between now and tomorrow morning when I make it back to DC.

Chris Heathcote is bringing up the rear of the conference with a talk on geolocation.

• no magic bullet
• GPS is not the solution
• appreciate the toolbox
• measures
• accuracy
• availability
• reliability, trust, etc
• What’s good enough?
• 20-50m?
• too much costs too much
• to little isn’t useful
• Here we go
• 0. assume: The Earth
• 1. the time: is it light? dark? what time is it here?
• 2-7. cultural clues: which cel phone operators? which wi-fi? phone number syntax? newspapers available? accuracy: country
• 8. Ask someone.  Accuracy: 10 meters (if you’re lucky)
• 9. Use a map.  Accuracy: 10 meters to 1 mile
• mobile tech
• 10. cell ID.  You have to go through the operator.  accuracy: 50m to 2 miles.
• 11. cell ID (local lookup): you’ve got to keep track of geo data youself
• 12. angle of arrival (AOA)
• 13. time distance of arrival (TDOA) 30-50m
• 14. observed time difference (OTD) 25-250m
• 15. assisted GPS: mainly in japan.  cel operator assists GPS chip with your location for more accuracy.
• Geo Tech
• 16. GPS: pretty good, but doesn’t work everywhere.  Are there satellites above me?  kills batteries.
• 17. WAAS: improves accuracy for GPS.  Even more sattelites required.  2m-25m
• 18.differential GPS: needs two receivers pretty close.  1-3m.
• street furniature
• 19.postcodes and zipcodes.  usefulness varies
• 20. street names.  Not all countries have street names.  hard to enter when mobile. 20m-hundreds of miles
• 20a. street corners, intersections provide more accuracy. 10m-miles
• 21. street numbers: great, if available.
• 22. biz names. go out of date really quickly.  lots of some types of businesses in some locations
• 23. landmarks and littlemarks.  what can you see?
• 24-26. public transport.  bus stops, street lamps, traffic lights.  bus stop UUIDs.  Data is proprietary
• 27. location street signs.  dedicated geolocation street signs.  10m accuracy.
• 28. geowarchalking.  rock!  pirate geo graffiti.  geolocate the world.
• emerging tech
• 29. dead reckoning. accelorometers, compasses, really accurate measurement of relative positions.  needs accurate location and time source to start with.
• 30. wi-fi triangulation. active campus
• 31. broadcast TV/radio triangulaton.  needs broadcast reception from three different locations.  not likely in many areas.
• IP lookup: great for a country or continent but not much better than that.  Varies.
• location advertising
• 33. encoding of location in access point name/location points.  part of SSID or whatever is advertised.
• 34. local servers/rendezvous: where am I. fixed machines that say where they are.
• 35. bluetooth.  dink.  here you are.
• bonus
• 36. RFID.  anything that transmits can give you location.  Is this scanner geolocated?  or in reverse, card senses if scanned and potential lookup
• a social future
• 37. who are you near?  where are they?
• 38. objects you are near. are they broadcasting their location?  what’s more accurate?  do something with it.
• 39. the road most traveled. recording and aggregate accurate flows.  time, speed, quantity of movement.  maps autogenerate themselves on the fly. better directions, see who has been where recently.  WAAG.
• winding down: location is important.  what if you want to be lost?

the presentation is here.  #geo on irc.oftc.net is another place to check out.  locative.net.  Become a geowanker.  Great presentation, Chris!

Edd is going over Dashboard.  It’s pre-alpha, runs under Linux in C# using Mono, and just all around rocks.  I’m really impressed with the information that it snags when you go to one of Edd’s page.

It’s sort of like jibot for the rest of your computing life.  It is absolutely brilliant.  I’ve been looking at screenshots of it for quite some time, but Edd’s presentation makes me want to go through the trouble of installing it.

As usual, dajobe is logging everything with realtime IRC notes in #foaf.  The notes are excellent as usual.

It’s off to the Plaza room for a hardware hacking session that sounds really interesting.

Andrew Huang, or “Bunny” has some suprisingly easy hardware hacks to share.  He’s the Hacking the XBox guy.  Right now he’s going over the basics of hardware hacking, reverse engineering, and the like.  Current slide: Is RE Legal? (Answer: Yes- but IANAL).  Hardware hacking and reverse engineering is sort of a a checks and balance system.  For example: does that computer indeed have the chipset in there that it says it does.

Watch out for the DMCA, it bites.

Why reverse engineer?  Why not?  It can be a curiosity.  Tweaking and and innovation is another reason to hack or RE.  Accountability is another reason.  Hacking is balance.

Printer cartridge ink chips suck.  Lock in sucks.  Hardware hackers can help by reverse engineering the chips and lock in system.  Get it out into the public.  Tweak your cars.  Chip your cars.  Why ship an engine but not enable all of its power?

Hardware hacking appears to not be as easy as it used to be, but it’s still pretty darn easy.

Emerging trends: circuit boards and dev boards are really cheap now.  They used to be really expensive, but now they’re cheap.  The barriers for entry still apear to be high, but are quite lower.  No single trend solves the problems of hackers, but several trends are here.

• cheap circuit boards.  You can send design files to a company, they make them, send them out without having someone touch the board in the process.  Breadboards were great, but these things are better.
• You can get really cheap PCBs which you can program and extend.  He has a list of sites up, but my eyesight is pretty bad so I can’t transcribe the URLs.  The talk should be online later.
• FPGAs are helping.  ASIC: Application Specific Integrated Circuits are great, but really specific and not good for general hacking.  FPGA (Field Programmable Gate Array) chips are a hardware hackers friend.  They can do all kinds of stuff.  They can be really fast.
• Design tools are getting better.  For example, WebPack is free.
• Open Source hardware is here.  The tools and designs can be open.
• Soldering and desoldering is a lot easier than it appears.  ChipQuick has an alloy for easily removing chips from boards.
• Probing boards has gotten better.  The leads are really small, but the tools are definitely there.  Micrograbbers are less expensive and very efficient for hearing what an individual pin is doing.
• A lot of the stuff that costs $25,00 can be done with much lower tech and a few weeks time.
• IC Analysis is up and coming, getting easier.  The tech is a bit above my head, but cool no less.
• Back Doors.  It’s all about back doors.  Use the back doors, they are your friend.

I was hoping for some more practical hardware hacking, though the book Hardware Hacking Projects for Geeks by Scott Fullam should cover a lot of that.  I’m excited about lower cost FPGAs and PCBs.  I must check in to that some more.

After a fine Denny’s breakfast with Ewan, I snuck in to the GeoURL session a few minutes late.  Right now Joshua Schachter is showing off a ton of cool sites that use the technology.  I really hope that the links from the talks are collected somewhere, because all of these sites are really cool and worth digging in to.

As with anything with google juice, spammers have been trying to take advantage of it, but Joshua is on the job, so don’t worry.  If your blog is not yet GeoURL aware, get with the program and add it now.

Joshua will be putting his presentation up on the web, I’ll link to it as soon as I find it.  Lots of amazing stuff has been built on top of the simple but effective GeoURL tag.

Danny O’Brien presented an excellent session on life hacks.  He sent out questionaires to alpha geeks who do work publicly and interpreted the results that he got back.  When he asked for screenshots of said alpha geek desktop, the most common theme was shells.

And shells, and shells, and shells, and shells…

The command line is the lowest common denominator when you work publicly.  The command line is also scriptable.

The other common trend was the use of a todo.txt file or something similar.  Alpha geeks tend to work with tools that they know, and todo.txt or otherwise organizing life in text files.  The common view among alpha geeks is to not trust your software more than you have thrown your computer in the past.

The email client is also an excellent organizational tool.  Alpha geeks use it a lot.  Private blogs and interntal/secret RSS feeds are also becoming more important, and often replacing email as an organizational tool.  Alpha Geeks are rapidly spending a fair amount of time in their RSS aggregators, and are using the software that they trust.

I was suprised that wiki didn’t come up as an organizational tool.

Scripting the daily life is also a common theme among Alphas.  Personal scripts tend to be short, for specific use, may have a limited use life, and are often embarrasingly encoded.  Often a script is written, used shortly, and promptly lost or forgotten.

Danny is an excellent speaker, and it was great to peer into the minds of alpha geeks in the public space.

I’m sitting near the back of the FOAF session.  Complete notes can be found in the #foaf log.  I’ll add my comments/notes here.

The big (but not really new) news is that Livejournal is rolling out a FOAF explorer, which could add another 2 million FOAFs.  Since Livejournal is such a community and friend driven site, I can only view this as a good thing.

Dan Brickley gave an overview of FOAF and there was some discussion about privacy concerns.  Edd Dumbill is currently demonstrating FOAFbot.

Christian Lindholm from Nokia is taking the stage a few minutes late for a talk that should include a demo of some cool blogging stuff.

What do people need from a mobile device?  It’s hard to find out, because surveys find out what people say they want, not what they want.  Focus groups are another way to find out, but you run in to many of the same problems.  Christian calls these Overt needs.  You can extract data from focus groups by looking at what is actually driving people, not just what they talk about.  Lifestyle studies, analysis, observations, etc help figure out exactly what’s going on.

Nokia’s product principles.

Virtuvius: three pillars of architecture: Firmitas (firm, solid, etc), Utilitas (usefulness, utility, usability), and Venustas (beauty etc).  If you don’t address all three, your product will suffer.

Sullivan: form follows function.  But what drives the function?  What comes before function?  A users needs drive function.

Convergence is a hot topic.  What are we supposed to do?  Shrink a pc down?  bulk a phone up?  Chopsticks vs. knife and fork.  Forkchops?  Sporkchops?  How silly is that?  Convergence has to make sense.  The smartphone HAS to be operated one handed.  It’s a big thing.  That way you’re not spending all of your resources computing, you can be living and computing at the same time.  He was talking about this a bit on the way back from dinner the other night.  (There’s a plce for two handed devices (ala Communicator, Series 90, etc), but for smartphones, gotta be one handed.

The intersection of functionality, form, and context is user delight.

For example, usability for the older phones was a big issue.  Do you go for something that is efficient or understandable?  Or do you go for readability on a small device?  Or do you go for beautiful with graphics and icons and stuff?  Nokia did their best to incorporate and compromise.  That is the Nokia Way.

Christian has a case study.  There’s a really good book called Mobile Usability.  I flipped through it earlier, and it looks quite good.  Here’s a case study: the digital Kazu.  Kazu is a person and also Japansese for balance.  He carried a notebook around with him and just wrote all kinds of stuff in it.  Timeframe: 1994.  He misplaced his book, and freaked out.  One day Christian wanted to build an electronic version of it.  He was using a Newton at the time and was quite happy with it, but alas.

5 years later, the book is an annotated scrapbook, with notes, pictures, clippings, etc.  The book as evolved.  The book was a legacy.  A sort of self blog.  Around 98 a trend was spotted: digital cameras, internet, multimedia, content explosion, lots to keep track of.

Who was driving this trend?  Younger people, socializing, nomatic, living in the now.  Another group was the family centric recorders.  Recording kids, events, everything.  Nokia started attacking this problem back in ’98.  They did tons of research on GPS tracking, trails, facial recognition, moving the web to mobiles.  But the devices didn’t have enough MIPS.  They test and analyze things thoroughly.

Design Comcept A was PC based, allowed you to record and review your data.  Four views: patchwork, single view, matchmaker, and slide show.  Design Concept B was a revolving sushi bar.  Have things fly by you and pick out bits when you need them.  Organize by a more organic timeline rather than so linear.  Design Concept C takes little moments in to consideration.  Storing different moments in your life.  Pictures, text, audio, video.

They took the brain from A, heart from B and clothes from C.  It’s another compromise, another integration.  Together it’s a sexy concept app.  It could be groundbreaking.  It’s personal blogging, note taking, annotation, and more.  It’s presented in a unique way with a UI for your phone and a UI for your PC.  A PC has almost infinite storage capability from a mobile guy’s perspective.  It ends up being a two way multiplatform solution for managing your personal content.

Buzzword compliance: It’s ubiqutious.  Your phone is with you all the time.  Wallet, keys, and phone.  Your PC at home is your archive or your log.  It’s trusted.  (Make sure you back it up!)  The web is your universal place for sharing.  You of course need to be able to sync across all three platforms.

I really hope that this app (or one much like it) makes it out of the lab and in to our lives.  Here’s a closer shot of the current concept user interface:

I’m back in real-time notetaking at the moment.  I’ll catch up on the notes taken with a text editor earlier.

Rob Gilmore from Via Telecom is talking about upcoming mobile data tech.  I’ve had spotty access to my VPN, and don’t feel comefortable sending passwords in the clear over wireless, especially with all of the sniffing going on.

Rob is going over the current cellular technology, such as CDMA, GSM/GPRS/EDGE, and the current/near-term stuff CDMA20001x, 1xEV-DO, and WCDMA  Of the latter group, WCDMA is nowhere to be found in the US, but is all over Europe, Japan, and Oceania.  1xEV-DO can pull down up to 1.5Mbit down, though only 64kbit up.  That’s going to be a huge crutch, just as current GPRS/EDGE is optimized for download, and if you want to send data, you’ve got a small pipe.

Anywhere/anytime is the Next Big Thing.  There are some technical hurdles that still need to be worked out.  Korea is expected to get 1xEV-DV by the end of 2004.  1xEV-DO will be upgraded to address the lower reverse link in the near future.

Cellular and WLAN (think: Multiradio) working together is going to be big.  You might be able to use WLAN to supplement your data services when it is available, possibly by connecting hot spots into hot zones.  A business model needs to be figured out for this to happen though.  Operators need to make money in order for it to be worth their while.  I would love it if my cel phone would use local WLAN instead of GPRS for data if it is available (both present and billable).

Right now range and interference are two big things holding back 802.11 from being more useful.  We’re suffering through that right now, with 802.11b, 802.11g, and Bluetooth all fighting for our computers’ attention.

802.11n could be the next big thing.  It’s still quite immature tech, but could offer greater than 100Mbit throughput to the user (rather than the usual speeds which include overhead)  MIMO (multiple input, multiple output) antennae might make this possible.  It’s going to require a lot of signal processing though.  It would occupy one WLAN channel, and be in compliance with the 802.11 standards.  One access point could speak 802.11a/b/g/n.

WLAN and 3G roaming.  It’s possible using Mobile IP.  Technical hurdles include handoff, security, billing, etc.  Mobile IP allows for this, but the infrastructure isn’t there yet.  Using Mobile IP, you would have a home agent located at home, which would maintain the location for each of your mobile, and hand off to 3G if neccesary.  A Mobile IP foreign agent wold facilitate roaming.

The slides are flying by too quickly to write it all, but the devices would talk to a foreign agent, which would communicate with the home agent, and would work out connectivity.

WMAN is another possibility for future high bandwidth access.  802.16 is Line-of-Site, which could offer between 2-155Mbit.  802.16a supports mesh, and can go 75Mbit and is non-line-of-site.  Obligatory mention of the last mile.  DSL-like services, interconnecting WLANs, etc.

802.16e is another standard that might help out mobility.  It is based on 802.16a and aimed at carrying signals for PDAs and laptops.

802.20 is packet based, and greated that 1Mbit with a radius of 15km.

Next up, CDMA the next generation.  1xEV-DV (data and voice), and 1xEV-DO (data optimized) are two ways of going about it.  EV-DV does both but has to optimize voice and data differently.  I love this technology.  CDMA is like time-sharing revisited.

EV-DV (which is fun to say) has a single 1.25MHz bandwidth share between voice and data.  It’s 3.1Mbps peak download.  The scheduling is very complex.  Voice users are usually scheduled first.  Dynamic allocation of the unused basestation power is then distributed to data users.  This makes doing multiple things on one channel quite easy.  The second rev of EV-DV (Rel D) will significantly increase the upload (reverse link) bandwidth.  Subsiquently, the next rev of EV-DO (Rel A) will increase the reverse link to up to 1.5Mbit peak.

The EV-DV scheduler is really really complex.  Really complex.

CDMA is really complex too.  I miss GSM.

It will be interesting to see how things pan out EV-DV vs. EV-DO.  EV-DV seems to have its advantages, but EV-DO is here now.  EV-DV has a smoother upgrade path from IS-2000, whereas EV-DO requires new antennae, hardware, etc.

EV-DO is due for upgrades soon (in a month or two), and EV-DV has upgrades coming out, even though it’s not even here yet.  Via is working hard on EV-DV.  1xEV-DV is due out in Q4 2004, and in the US test trials may start by the end of the year.  EV-DV might be in place en masse by 2005-2006.

EV-DV headsets will be made by Nokia, samsung, LG with chipsets by Qualcomm, TI, and VIA.

This presentation has been quite informative, but acronyms abound.

The first keynote of the morning was by Marc Smith, a sociologist at Microsoft Research,  Marc had some excellent visualizations for the data collected from Netscan, which scans and archives tons of data and metadata about usenet newsgroups.  As a sociologist, he is quite interested in graphing, visualizing, and interpreting conversations and relationships as they happen across Usenet.

Marc discussed Schelling Points, or obvious places that people go in order to meet certain people.  Schelling Points can be complex and quite interesting when you throw the net into the mix.  Another important concept that he looked at was Yhprum’s law.  Yhprum’s law (Murphy spelled backwards) is that systems that should not work but often do.  This happens quite a bit online.

Usenet is not well, but it’s not dead either.  Last year 240 million messeges were sent by 8.5 unique identities.  Netscan has collected about a billion headers, and Marc showed several ways of visualizing the data.  First up was a tree view of all of usenet, showing growth and decay in certain areas over time.  It was quite interesting, and quite trippy at the same time.

He also visualized the difference between between a tech support style group with say alt.politics.bush.  The number of posters, their frequencies, and the number of threads that they used were strikingly different.

Another interesting visualization was tracking one user over time.  You can generate a sort of histogram with the data and tell quite a bit about a users’ habits at a glance.  One type of user was one that never initiates a thread, but adds to threads all the time.  This is the answer person.  They

Marc also discussed mobile machine readble tags.  It’s quite interesting tech, tho identical to a presentation given at Foo Camp.  More information can be found at the Aura site.

Via Jim, Edd was sniffing the Bluetooth air last night.  14 devices isn’t bad at all, though during Russ’ tutorial, I counted 19.

Long live Bluetooth!

While I’m at ETech, PMA, a huge photo industry show, is ramping up.  PMA is the biggest industry show in the US and second only to Photokina in prestige.  Before the show, I had somewhat predicted a G8 replacement of Canon’s G5, but they one upped me: the Canon PowerShot Pro1.  It’s 8 megapixels (which I called), but also has a 7x USM L-Series lens.  Pricepoint: just under a grand.  Also interesting is the PowerShot S1 IS, a 3 megapixel 10x camera with image stabalization.  More predictable is the announcement of the S430, S500, and SD110.  Those are all pretty much rez bumps to their ultra compact line.  Also new is the A75, raising the bar for mid entry level a bit.  The last thing that DPReview links to today is the Canon A310, which is a budget entry level camera, and looks like the Fuji Finepix A310 and is in the same segment.

All of this similar naming is really confusing.  We’re dealing with the same thing on the mobile phone front: how many series 60 devices are there that contain either the letter S or X?

Yikes, it’s too early.

Russ is currently trying to distribute an app via bluetooth in a room that is just awash with Bluetooth and Wi-Fi.  It has been extremely slow and problematic.  From the back of the room I can discover 19 Bluetooth devices, many of them are the T616s that were issued for the tutorial.  The Wi-Fi network has also been slow this afternoon.

Wireless access is a good thing.  I can roam around the lobby or second floor of the hotel and have access.  But we’re reaching a saturation point here.  We’ve maxed out the 2.4GHz spectrum and there’s just not much we can do about it.

Russ, Ewan and I were playing FIFA 2004 on the N-Gage last night and the gameplay was much slower than when we were playing in the hotel room away from the wireless madness.  We are assuming (but could be wrong) that all of the other Bluetooth and 802.11 traffic was bogging things down as we faced off against each other.

Russ is going over the basics of J2ME for his talk at ETech.  He’s got a monster G5 and a big Apple flatscreen in front of him, but is presenting from his laptop.  Poor guy.

More pictures from ETech can be found on my moblog.

Update: The source code from Russ’ talk is online.