|
| News | Profile | Code | Photography | Looking Glass | Projects | System Statistics | Uncategorized |
Blog |
It's the end of 2010, and 2011's just around the corner! In the past, I've written "a year in review" types of articles about what I did and accomplished in the previous year, or post new year's resolutions that I've never kept. This time, I won't do either. Instead, here's a few predictions and comments on topics that interest me.
The US (and world) economy won't collapse in 2011 and the world won't end in 2012.
Ok, now that we've gotten that out of the way, let's continue…
We've seen some of it already (healthcare, net neutrality, talks of cap & trade, etc.), but I think the United States' government will continue to grow much larger than it's already become in the last decade or so.
Net neutrality bills and other type of broadband initiatives will be approved and implemented, requiring broadband providers to upgrade their infrastructure to certain speed requirements, while losing the freedom to engineer or block traffic. Blocking ports like TCP/80 and BitTorrent will become illegal for anyone providing Internet access (form coffee shops to large MSOs).
MPG requirements for automobiles will continue to rise, eventually prohibiting (through high taxes or special licenses) the sale of sports cars to the general public, and causing auto makers to manufacture more efficient vehicles. Electric and hybrid cars will be promoted by the federal government and an international charging standard will be created to prevent plug incompatibilities of plug-in hybrids. LED headlights might make an appearance in the automotive industry!
In some way or another, the good-intentioned fight against [childhood] obesity in America will result in the government slowly starting to dictate the eating habits of Americans. I'm not implying that bacon will be taken off the market overnight, but I do think that the suggestion of a healthy lifestyle may go too far. Just like salt is on its way to be banned in some places, we'll start to see items offered by some fast food chains drop off the menu, more and more sugar substitutes appear, and portion size at resturants regulated. Eventually, I think hamburgers will be banned (remember this seaQuest episode?), too!
Anti-terrorism screenings and procedures will result in even more invasion of personal privacy, banning more arbitrary items from planes and trains (think iPhones and laptops).
Texting will become illegal, and more cars will start to support voice recognition via an extension to the Bluetooth protocol (ie, send and receive SMS/e-mail messages, post to Facebook, etc. by voice only).
Smart grid technology will continue to be deployed in most cities and modern neighborhoods. From the data obtained by such a system, households with more than some amount of "waste" (eg, computers running w/out power saving) may be taxed in some way or another, regardless of actual power consumption.
Home-schooling will probably be banned in the next several years, due to a divergence in the curriculum used and the inability for the US DOE to control it.
More whistleblowing sites like WikiLeaks will appear, and cause companies and governments to become more paranoid with the safeguarding of sensitive documents.
President Barack Obama will be re-elected in 2012.
Overall, software will become more complex, inefficient, slower, and buggier. I just don't see this changing, at all.
Apple, Google, Amazon, and Microsoft will all continue to embrace cloud computing, and coerce everyone into storing all of their personal data on their servers.
Facebook won't kill e-mail. However, e-mail spam volume will decrease.
Newer versions of desktop operating systems (Windows, Mac OS X, Chrome OS) will start to be "locked" to users, like iOS. So-called "jailbreaking" or "rooting" will be required to access the shell and filesystems. For most users, this won't be an issue because of cloud computing. The sale of software will migrate to online "stores" like the forthcoming Mac App Store.
Laptop, tablet, and netbooks will replace traditional desktop computers and workstations. Only in high-performance computing environments will desktops and workstations exist.
Desktop Linux will continue to fail, with the majority of Linux users migrating to Mac OS X or Chrome OS (I don't consider this to be desktop Linux).
IPv6 will continue to be deployed throughout networks in the United States over the next year or two, but it will be 3-4 years before the majority of content delivered to end users will move from IPv4 to IPv6. Router and switch manufacturers will find a way to produce cheap CAM and routing table bloat will be less of a concern, resulting in the maximum prefix length allowed into BGP by most providers to be at least 48 bits. Provider independent (PI) allocations from the RIRs will be given out to small companies and individuals, and the need for multihoming hacks (NAT66, SHIM6, etc.) will be reduced dramatically.
Whitelisting of DNS servers to provide access to AAAA records (like GoIPv6) will become a bit of a fad in 2011, but drop off dramatically in 2012. DNSSEC will start to see usage among the popular sites (Facebook and Google) in 2011, as well.
The majority of TDM circuits (DS-3, etc.) will be replaced with Ethernet over the next several years, and the PSTN as we know it will go away. Voice will start to move over IP everywhere, and phone numbers will start to become obsolete.
MSOs will become nothing more than ISPs, and video will start to be more subscription-based. The concept of "TV channels" will be gone, and all content will be viewed in a web browser or on specialized hardware. There will be more NetFlix-like services available.
More and more wireless ISPs (like Verizon, Clear, etc.) will appear with LTE Advanced services offering speeds that rival FiOS and DOCSIS 3.0 cable deployments. MSOs will eventually convert to FTTx and abandon coaxial cable altogether. Some (Comporium) have already started to do this.
There will be larger credit card security breaches in the next couple years, resulting in online merchants migrating to PayPal-like services. For retail payments, NFC will be used. IT departments around the world will rejoyce and not have to deal with PCI.
Merry Christmas!
If you can, unplug from The Grid and spend some time with friends and family.
Tomorrow there will be a vote on some potential network neutrality regulations. The agenda, according to the FCC, will include the following:
Open Internet Order: An Order adopting basic rules of the road to preserve the open Internet as a platform for innovation, investment, competition, and free expression. These rules would protect consumers' and innovators' right to know basic information about broadband service, right to send and receive lawful Internet traffic, and right to a level playing field, while providing broadband Internet access providers with the flexibility to reasonably manage their networks.
This is very vague, so I don't know exactly how detailed these discussions will be. If you choose to keep reading this article, be prepared that this is ALL my opinion (not the company that I work for, etc.). So, shield your eyes if you want!
Even in light of recent issues that might be viewed by some as a net neutrality problems and new services that will make it easier for ISPs to throttle, block, and charge for certain types of content, I haven't really seen a compelling reason for any government regulation.
Quite simply, if your ISP wants to block content, ports, protocols, or throttle certain types of traffic, I believe they have the right to do so. It's a free country. You, as the customer, also have every right to complain about it, and ultimately cancel your service. If that's the only service in the area (which is starting to be hard to believe with all the wireless services available, these days), then move, or deal with it.
Now, to step back a little bit, other than blocking new inbound connections to TCP/80 and throttling peer-to-peer traffic, are there any ISPs out there today in the United States that are charging extra for so-called Facebook (I say so-called because with CDNs, anycasting, and GSLB on complex sites like Facebook, it may be difficult to classify all Facebook traffic vs. advertisments or other third party media) traffic? Ok, how about blocking traffic to all search engines besides Yahoo! and Google? Ok, well, how about intentionally slowing down YouTube traffic? None?
Well, that last point can be a little more complex, since some might argue that (if you read the link above) Comcast is intentionally throttling content that transits their peering with Tata. If Twitter traffic is engineered (think BGP attributes like AS_PATH prepending, MEDs, LOCAL_PREF etc.) to traverse the Comcast/Tata link, is it considered throttling? Intentionally throttling? Well, maybe, but it doesn't really discriminate on exact content.
I do think that ISPs need to present the customer with exactly how they're blocking and throttling content. Maybe this requires some legislation, but honestly there may be cases where this isn't appropriate (for the customer or the ISP). It's difficult to predict how laws will impact things down the road.
Also, I think residential ISPs should probably change their model and charge for raw bandwidth, and I think they should have started it from day one. Your tier 1 and other large Internet backbones do it (Cogent, Level3, etc.), today, so why isn't that passed onto the residential subscribers? If you only use 10MB (ok, maybe proxy ARPs themselves amount to more than that, but anyway..) of traffic per month, why should you have to pay the same as the power user across the block who chews up 80GB per month? Why is consumption-based billing (CBB) so evil? And, if the ISP wants to charge an obscene amount for real unlimited traffic volume, then they should be able to (but guys, unlimited is unlimited, don't say unlimited and then unplug subscribers after they hit the imaginary 250GB cap).
Now, I digress, but start thinking about network neutrality and carrier grate NATs (CGNs). When the remaining /8s are handed out to the RIRs, and then the RIRs run out of blocks to give out to residential ISPs, customers are going to start receiving RFC 1918 addresses (NET-10, most likely) on their CPE interfaces. Yep, this means no more public IPs. IPv4 traffic from the customers' RFC 1918 addresses will be translated to public IPv4 addresses at hub sites or head ends (whatever you want to call it - POPs). Will network neutrality make this illegal, because ISPs are intentionally blocking inbound traffic to all TCP and UDP ports on the customers' CPE? Uh, well, there's no way around it with many-to-one NAT. Just think about it, it can get pretty ridiculous.
Also, if you want a quick summary of my views on blocking, limiting, etc. please check out my comment on Ian Gulliver's Net Neutrality? blog entry.
I saw Tron: Legacy in 3D, yesterday. It was, in a word, awesome.
I had re-watched the original Tron film from 1982 a few weeks ago, just to bring myself back up to speed. It isn't really required to understand the plot, since the audience is brought up-to-date in the first couple minutes of the film. However, it was neat to compare the graphics and special effects between the films, which are almost 30 years apart, both being state of the art at the time.
Anyway, the film was fantastic. The lighting, graphics, and special effects are second to none, with the musical score clocking in at around the same level (Daft Punk being a no-brainer choice for the musical score). The plot was mildly interesting, too, and the acting was decent. The illuminated costumes and bodysuits worn by the cast were done nicely (erm, Olivia Wilde looked great). It's too bad the film hadn't come out earlier in 2010, I would have liked to see some Tron Halloween costumes!
There was a little blurb at the beginning of the film that instructs the audience to keep their 3D glasses on, even though part of the film is intentionally displayed in 2D. It turns out that the only 2D parts of the film consist of flashbacks and scenes shot in the real world (compared to scenes shot in the Grid, which are 3D). This reminded me of, among other things, The Wizard of Oz's use of Technicolor, which was fairly effective in distinguishing Oz from Kansas.
Also, as seen in the theatrical trailer (and the one shown on TV, too, but not as prominent), the terminal that brings Sam into the Grid is apparently a Unix -variant named SolarOS:
Um, SolarOS running on sun4m with an i386 architecture? And uname output from iostat with a Linux-style top, to boot! Certainly doctored, but funny, regardless. At least some of the ps and other commands Sam typed afterwards more closely resembled reality. I always enjoy seeing screenshots and text from real operating systems in movies and TV shows. For example, routinely seeing MRTG graphs (via rateup, not RRDtool) and BitchX in 24 always cracked me up…
My only regret is that I didn't see it in IMAX! Maybe a second showing is in order.
As I mentioned earlier this week, I've moved over to the dark side.
Yep, I bought an 11" MacBook Air (I may refer to it as MBA or orion, its hostname).
This isn't my first Mac. I own two Mac minis, but their roles are specialized (HTPC, etc.). This is the first Mac I've purchased that will replace another computer that has a more generalized role, my Asus Eee PC 900 running Debian GNU/Linux. Yep, you read it right: I'm replacing a Linux box with an Apple computer running OS X.
I'll go over the reason for my switch, a short review of the MBA (plus some general OS X comments), and some conclusions.
A number of reasons prompted my switch. Speed and display resolution won out, I think.
The Eee PC 900 is a great little netbook. Over the course of its life (2+ years) I've upgraded the memory (2GB), SSD (32GB), and Wi-Fi adapter (802.11a/b/g/n). It continues to run well. However, software and the web continues to bloat, requiring more and more hardware on the client-side, not to mention more bandwidth. Websites (long articles on Wikipedia, for one) that normally require a second of rendering time started taking 8-10 seconds. Flash Video, maybe better now than it used to be on Linux, continues to demand more processing power. I don't think I can watch any videos from YouTube on the Eee anymore without having to use youtube-dl and MPlayer to play the videos offline.
The 1024x600 resolution is getting a little cramped, too. Many web sites nowadays are starting to require browser windows that exceed 1024 pixels in width, and things like Pidgin take up the whole screen, requiring the overuse of multiple desktops. Even 170 columns and 45 lines (the maximum I can get when fullscreening w/the fixed font) in a terminal emulator is starting to not cut it, anymore.
So, the solution is a new laptop.
I searched around, looking at Dell, Asus, and other brands, trying to find a 9" or 10" netbook with an SSD (I'm over hard drives in mobile devices, at this point) and Atom processor. I found one or two, but there was always something missing (no Bluetooth, etc.). I then started looking at Apple.
The MacBook Air models were enticing, to say the least. The 11" model seemed to fulfill most of my requirements (higher resolution, SSD, Bluetooth). However, I was initially hesitant to purchase because I wanted to run Linux. Well, I obviously gave up that requirement, and impulse-bought the 11" MacBook Air with the following specifications:
OS X has taken a few days to get used to. The hotkeys are easy to learn, and there are quite a few of them. I've hardly mastered the trackpad gestures, but I'm getting there.
I'm pleased with a good many features on the MBA. The main ones are battery life, design, and speed.
The Eee PC 900 started out with 3-5 hours of battery life when Wi-Fi was enabled. The MacBook Air seems to be pushing 4-5 hours with the same amount of usage, and that's with the screen on the whole time. I'm not going to complain about it. Most of the time I only need battery power for 2-3 hours, anyway, so this is more an enough.
I was tempted to fool around with CoolBook and try undervolting the Core 2 processor, but figured it might be more trouble than it's worth. An extra hour or so of battery life at the cost of possible instability didn't really interest me.
Although the MacBook Air has a fan (I hate fans), it's really quiet.
I'm not kidding when I say it's quiet. I have to put my ear up to the back of the display hinge to hear it. Speaking of, the exhaust vent is not on the sides of the unit, it's facing the hinge for the 11" display. I suppose this is part of the reason why there's no [obvious] option for the system to remain on when the display is closed - it'd probably overheat.
One thing I just noticed, too, the MacBook Air runs very cool. Using a temperature gun pointed at the bottom of each unit, here's how the Air compares to my Eee PC when idle:
I suspect the aluminum frame helps to disperse the heat much better than a plastic frame, in the case of the Eee PC.
Along those same lines, the MacBook Air has a ton of temperature sensors, which can be viewed with the Temperature Monitor application:
(orion:13:41)% /Applications/TemperatureMonitor.app/Contents/MacOS/tempmonitor -a -l CPU Core 1: 46 C CPU Core 2: 41 C SMART Disk APPLE SSD TS128C (Y09A40DNK0WK): 37 C SMB NORTHBRIDGE CHIP DIE: 44 C SMB NORTHBRIDGE CHIP DIE: 50 C SMC BATTERY: 29 C SMC BATTERY CHARGER PROXIMITY: 39 C SMC BATTERY POSITION 2: 29 C SMC BATTERY POSITION 3: 28 C SMC CPU A DIODE: 45 C SMC CPU A PROXIMITY: 40 C SMC LEFT PALM REST: 26 C SMC MAIN HEAT SINK 2: 30 C SMC NORTHBRIDGE POS 1: 42 C SMC POWER SUPPLY POSITION 1: 36 C
I've started graphing a few of them with the help of mrtg-rmt, too.
The power adapter is one of the smallest ones that I've seen. Proving a maximum of 45 watts, Apple gave it the flexibility of becoming a brick or a wall wart, which is nice. Also, the cable is nice and thin so it moves around fairly painlessly, unlike others.
To feed my Unix addiction, MacPorts gives me most of the familiar tools I'm used to on Linux. It works well, but some ports don't built, right now (GIMP, dsniff, etc.). I should probably file a bug!
I've got my VPNs working on OS X with no issues. For OpenVPN, I'm using Tunnelblick and either Miredo (from MacPorts) or manual 6in4 over OpenVPN using the following BSD-type ifconfig statements:
# start IPv6 tunnel ifconfig gif0 up ifconfig gif0 tunnel 10.3.254.233 10.3.4.20 ifconfig gif0 inet6 2001:48c8:1:116::2/126 route add -inet6 default 2001:48c8:1:116::1 # stop IPv6 tunnel route delete -inet6 default ifconfig gif0 -tunnel ifconfig gif0 inet6 2001:48c8:1:116::2/126 -alias ifconfig gif0 down
I don't think there's currently a way to setup 6in4 tunnels using the GUI. Disappointing, but not surprising.
I hate for this section to be longer than the previous one, but I'm often overly critical of computer and electronics equipment - plus I'm not one of those folks who worships Steve Jobs and all things Apple. Anyway, let's start.
There are MAJOR Wi-Fi issues in OS X (or maybe the Broadcom hardware in the MBA?). Specifically, 10.6 has some problem with reconnecting to saved networks that happen to not broadcast their SSID in the beacons. At home, I've got a Cisco 1232 WAP that provides 802.11a/b/g with a separate non-broadcase SSID for 802.11a using WPA2-PSK. I can type in the network and key manually, and join the network - this works fine. What doesn't work is reconnecting to the network after saving it. It just never finds the network.
I managed to find some two command-line utilities that sort-of helped diagnose this, airport and networksetup. I used airport to turn on some debugging. Here's what happens during the periodic probes when OS X can't find my SSID. Let's assume my SSID is MyHiddenSSID:
Mon Nov 29 22:12:49: <airportd[56086]> Scan for ssid = MyHiddenSSID, 19 channels
Mon Nov 29 22:12:49: <kernel> IO80211ScanManager::startScan: Directed scan request received from pid 56086 (MyHiddenSSID).
Mon Nov 29 22:12:49: <kernel> IO80211ScanManager::startScan: Initiating scan.
Mon Nov 29 22:12:50: <kernel> IO80211ScanManager::getScanResult: All scan results returned for pid 56086.
Mon Nov 29 22:12:50: <kernel> IO80211ScanManager::startScan: Directed scan request received from pid 56086 (MyHiddenSSID).
Mon Nov 29 22:12:50: <kernel> IO80211ScanManager::getScanResult: All scan results returned for pid 56086.
It just does that over and over for every network in the list. Nothing special seems to stick out for MyHiddenSSID that's different than the others. I know it can see my hidden SSID in this scanning state, because I can force a probe using the airport utility:
% airport --scan=MyHiddenSSID
SSID BSSID RSSI CHANNEL HT CC SECURITY (auth/unicast/group)
MyHiddenSSID 00:19:06:20:ef:43 -66 64 N -- WPA2(PSK/AES/AES)
Here's what the log says when I manually enter the SSID and PSK:
Mon Nov 29 23:21:52: <kernel> en0: Terminating supplicant.
Mon Nov 29 23:21:52: <kernel> RSNSupplicant: Releasing authenticator for 00:19:06:20:ef:43
Mon Nov 29 23:21:52: <airportd[56417]> airportd started
Mon Nov 29 23:21:52: <airportd[56417]> _APXMIGwakeupAirPortdServer: received 'Associate Event'
Mon Nov 29 23:21:52: <airportd[56417]> process_associate_event: ifName=en0 networkName=MyHiddenSSID securityType=WPA2 Personal, wep40=(null), openSystem=(null) password = ****, oneXDict=(null), remember=true
Mon Nov 29 23:21:52: <kernel> IO80211ScanManager::startScan: Directed scan request received from pid 56417 (MyHiddenSSID).
Mon Nov 29 23:21:52: <kernel> IO80211ScanManager::startScan: Initiating scan.
Mon Nov 29 23:21:55: <kernel> IO80211ScanManager::getScanResult: All scan results returned for pid 56417.
Mon Nov 29 23:21:55: <kernel> IO80211ScanManager::startScan: Directed scan request received from pid 56417 (MyHiddenSSID).
Mon Nov 29 23:21:55: <kernel> IO80211ScanManager::getScanResult: All scan results returned for pid 56417.
Mon Nov 29 23:21:55: <airportd[56417]> __ACNetworkAssociate(MyHiddenSSID, ****)
Mon Nov 29 23:21:55: <airportd[56417]> __ACNetworkAssociate: Making association request with apnPrivate->isWep40 = (null) and apnPrivate->isWepOpenSystem = (null)
Well, that certainly looks a bit different. Since it tries to associate. Unfortunately, I don't have a fix (well, a good fix - turning on SSID broadcasting fixes it, but I don't want to do that), and lots of other folks have the same problem with 10.6.
Terminal.app is nice, but it's slow and laggy. If you compare it to running rxvt in X11.app, you'll see the difference. It took some time to get used to the key combinations, especially having to use the Fn key to send PgUp & PgDn keys to the terminal itself. Also, just like Konsole, Mutt suffers from the backspace problem.
OS X doesn't read the Linux ext3 filesytem. This stinks, but I'm not surprised, at all. I ended up converting an external USB HDD of mine to HFS+ (instead of FAT32, due to silly small file size limits), since Linux has a mature driver for it, and it's a good compromise.
There aren't many soft cases for the 11" MacBook Air, yet. I did find this one on Amazon, but a corduroy case strikes me as a little strange. I'll probably wait for a better selection, though.
The escape key on the keyboard doesn't depress properly when struck on an angle, and seems a little fragile. I find this happens to me when I'm editing in vim, otherwise I don't really hit the escape key too often. I don't think the other MacBooks suffer from this, since the escape and function keys are a bit larger. They're pretty thin on the MBA. I'll just have to get used to consciously hit the escape key top down instad of on an angle, I guess.
Why, oh why, is there no way to set the system hostname in the GUI? The Bonjour and Bluetooth names can be set, but not the hostname. Out of the box, the default screensaver and terminal prompt will show the DHCP-assigned hostname, like h-10-227-43-200 or whatever. I had to dig around a bit before I found out that this can be set using the scutil utility:
% sudo scutil --set HostName orion
The manpage for scutil(8) discusses command-line arguments for displaying and overriding the other service names.
This is an odd one. Apparently, there's no easy way of determining if the MBA is powered on if the display is shut off. The Apple logo in the back of the screen is lit only when the LED backlight for the display is active. I can put my ear to the area of the hinge near the exhaust vent, to listen for the noise of the fan, but that's about it. Seems silly. Why not just add a tiny white LED somewhere near the keyboard?
OS X has some audio problems. Specifically, there's a huge delay (at least 300ms, from what I can tell) before the audio system starts playing a sound effect or stream. For example, if I trigger the bell in Terminal.app or press the volume buttons on the keyboard, there's a delay before I hear a sound. If I play sounds in succession, there's no delay, because apparently the sound system is active, but if I wait 15-20 seconds before triggering another sound, I'll get a delay. This is much worse when connected to Bluetooth via A2DP. Sometimes the delay is 1-2 seconds!
While the Wi-Fi problem is certainly annoying, I'm getting used to the MBA, and think I'm starting to like it. OS X is still taking a little time to get used to, and I might try booting Debian this weekend, just to see what happens.
Although sometimes it might seem like the world is going to hell in a handbasket, we've all got a lot to be thankful for.
Think about it for a few minutes. Write it down on note cards if you want (that's what we're doing later today).
Have a Happy Thanksgiving.
I've had lots of issues with GNU/Linux's Bluetooth stack, BlueZ, for awhile. Incomplete documentation, constant API changes, and a complete lack of CLI utility support. The whole development team seems to be very GUI-centric. This bothers me. Back in the 3.x version of the BlueZ stack there were a few nice CLI utilities (hcitool, sdptool, l2ping, hciconfig, hidd, rfcomm, etc.), but most of them are deprecated now.
Anyway, I've tried to embrace the GUI which is bluetooth-properties(1), but I've had some problems with my Bluetooth headphones not wanting to reconnect to the host upon powering on. Unfortunately, the Bluetooth GUI doesn't provide a way of initiating a connection to a paired device (without repairing, of course):
I'm not sure if the developers are expecting applications that use the BlueZ stack to initiate connections themselves, but PulseAudio certainly doesn't do this, because the device doesn't even show up in pavucontrol(1) if it's not connected.
So, I ended up writing a tiny little Python script that tells BlueZ to connect to the specified audio device, if it's in range.
It's pretty simple, just run it with no arguments to show the list of Bluetooth devices:
% ./btconn.py Codes: P - paired, C - connected, T - trusted Name: (none) Address: 00:02:76:0E:79:9E Flags: P Path: /org/bluez/26825/hci0/dev_00_02_76_0E_79_9E Name: 9xxBackBeat Address: 00:1C:EF:67:E7:48 Flags: PT Path: /org/bluez/26825/hci0/dev_00_1C_EF_67_E7_48 Name: Microsoft Bluetooth Notebook Mouse 5000 Address: 00:22:48:DD:F3:BC Flags: TC Path: /org/bluez/26825/hci0/dev_00_22_48_DD_F3_BC
Or, run it with a partial name of a paired device, and it'll try its best to connect:
% ./btconn.py 9xx Connection to "9xxBackBeat" was successful.
It only connects to audio devices, right now. Maybe I'll add support for other types of devices in the future, but right now I don't see the need since I haven't seen problems with input devices upon reconnecting. I don't have any experience with Bluetooth printers or other devices, but I'd suspect the applications actively call connect() as part of their normal operation.
Grab btconn.py from here. It's been tested against version 4.66 of the BlueZ stack on Debian GNU/Linux.
It's coming.. faster than you think. I've noticed that more and more mainstream sites are either adding ipv6.* AAAA-only resource records, or just taking the dive and adding AAAA RRs to the existing DNS records of their public-facing websites. Here's two lists of some you may or may not be aware of:
ipv6.* (scaredy cats!)
A/AAAAs (gutsy!)
Hopefully there will be an http://ipv6.{some other MSO}.com/ appearing in the next couple of months.
Displaying page 14 of 121 of 965 results
     |
This HTML for this page was generated in 0.013 seconds. |