2012-Nov-18 18:45:03 +0000 @822
It’s been over two years since I posted anything new here. Actually, I’m mostly tweeting, using Google+ and posting to some other group blogs.
I do get traffic from people who find the information on this blog useful. As of now, I don’t think anything is so outdated that it is wrong… but use caution. I might have more to post here in the future, but I’m mothballing it because I don’t have time right now.
Be safe, everyone!
2010-Aug-27 21:29:34 +0000 @937
How far away does the “Close Call” feature on Uniden scanners work?
I just tried a test in a suburban area in a large parking lot. An FRS radio worked reliably at 100 feet and 80% of the time at 250 feet.
One thing I did notice was that if the signal gradually increases, such as walking toward the scanner with the talk button down, the scanner did not activate.
There didn’t seem to be much actual numerical information about this feature on the web, so I hope this test is useful to others considering a scanner with this feature.
2009-Dec-22 18:10:28 +0000 @798
I just tried a couple of tests by posting a 155k file on a few types of web servers to measure the load time. DNS was factored out and all client side caches were cleared.
Rackspace CloudServer 1.35 to 1.65 seconds
Rackspace VPS running Apache 2, consistently 1.05 seconds
LimeLight CDN 0.404 seconds
The differences are pretty profound, I think. I’ll be looking at CDNs in the future…
2009-Jul-07 20:03:16 +0000 @877
I tried AirTran’s Internet service on a recent flight. The provider was GoGo. The trip from Atlanta to Seattle had about 3 hours of usable time for $12.99. It was decent enough for email and web. I was able to establish a Skype connection, but not hold a conversation.
The transit speed was about 1.6Mbps both ways, but it seemed to have some sort of use limitation. When transferring files, they would go at 1.6Mbps for the first 2 megabytes, then drop instantly to around 24kbps.
The latency varied greatly between 120 and 2000 milliseconds, making SSH connections hard to use. Additionally, packets were sometimes returned out of sequence. Service quality seemed to come and go, perhaps a sign of our proximity to ground stations or the ebb and flow of user traffic.
Connectivity was nice, but there still isn’t enough room on the seats to type, so don’t expect to get work done – just kill boredom with YouTube.
64 bytes from 220.127.116.11: icmp_seq=688 ttl=53 time=778.819 ms
64 bytes from 18.104.22.168: icmp_seq=689 ttl=53 time=128.547 ms
64 bytes from 22.214.171.124: icmp_seq=690 ttl=53 time=162.566 ms
64 bytes from 126.96.36.199: icmp_seq=692 ttl=53 time=156.998 ms
64 bytes from 188.8.131.52: icmp_seq=693 ttl=53 time=165.286 ms
64 bytes from 184.108.40.206: icmp_seq=694 ttl=53 time=181.222 ms
64 bytes from 220.127.116.11: icmp_seq=695 ttl=53 time=124.276 ms
64 bytes from 18.104.22.168: icmp_seq=696 ttl=53 time=138.904 ms
64 bytes from 22.214.171.124: icmp_seq=697 ttl=53 time=146.648 ms
64 bytes from 126.96.36.199: icmp_seq=698 ttl=53 time=174.594 ms
64 bytes from 188.8.131.52: icmp_seq=699 ttl=53 time=427.510 ms
64 bytes from 184.108.40.206: icmp_seq=700 ttl=53 time=1039.029 ms
64 bytes from 220.127.116.11: icmp_seq=701 ttl=53 time=159.956 ms
64 bytes from 18.104.22.168: icmp_seq=702 ttl=53 time=179.433 ms
777 packets transmitted, 711 packets received, 8% packet loss
round-trip min/avg/max/stddev = 119.437/303.578/2274.701/289.599 ms
2009-May-14 17:15:23 +0000 @760
Today I asked some of my friends to Audio Book recommendations because I’m going to use up my credits on Audible.com and cancel for a while. I have well over 10 days of podcasts and audiobooks that are backlogged, with more arriving each day, so I figure I’ll save the fee until next year.
I completely forgot that I should make my recommendations since I have so much material. Here are a few that top the list:
2009-May-11 12:52:59 +0000 @578
I’m sure I’m not the first to do this, but I just figured out you can create a “smart playlist” on your iPod to show the most recently downloaded podcasts. This makes it much easier to listen to new media instead of searching the 50 podcast feeds.
2009-Mar-16 15:03:50 +0000 @669
I guess I’m a machine. I can’t get these CAPTCHAs from MegaUpload.com.
I’ve been working on my own ray-traced CAPTCHA for a few weeks now, and I’m finding it very difficult to create a CAPTCHA that I can’t write software to break – unless it’s darn near impossible for a human to solve. I think these are dead. Time to move on to something else.
2009-Feb-03 15:00:46 +0000 @667
For those of you keeping track, the Unix Epoch will be 1234567890 on Feb 13th. (Exact time depending on your Time Zone… unless you want to be UTC about it.)
2009-Jan-11 15:13:21 +0000 @675
Yes, part of the idea of going to Burning Man is to escape all the stuff back at home, but the reality is that if something goes horribly wrong back at work, people could loose their jobs – so I carried some satcom gear just to check in once a day. Since there is little posted on the Internet about telecommunications from Burning Man, here’s a contribution of my experience.
I rented a BGAN Immersat ethernet transceiver and a Iridium satellite phone. The BGAN is really expensive, not only the weekly rental fee by the $10 per megabyte charge. Both work as promised, but no better.
Realistically, light only travels so fast and the trip to orbit, Houston, and back adds substantial lag. The throughput isn’t bad, but the time between packets precludes gaming completely.
The Iridium phone worked every time we tried it. While it is advertised as “you must be able to see the sky” it was able to receive text messages and “hear” incoming calls while inside a metal vehicle and inside my two story wood house. If the phone rang, I would have to go outside so the satellite could “hear” the less-powerful unidirectional handset. The phone does have a lot of compression artifact and the bit rate is low. It sounds like talking on early VoIP applications. While it is “full duplex,” you really need to concentrate on taking turns talking because your conversations are half a second or so out of phase with each other. I did have trouble calling some people with fancy pants handsets like Treos, which seem to have their own sample rate that never meshed up. Then again, I have trouble calling those people from analog lines too. The phone worked in sandstorms. I only saw one cellphone that worked at Burning Man. Someone brought a tower and an old analog phone. That tower was scheduled to go off-line on 30 September 2008.
The BGAN Immersat satellite phone worked consistently… consistently slowly, that is. If you needed Internet, it was the best thing out there. If you call up a web page, it might take 5-10 seconds to it to start appearing, but once it started loading it was much faster. Streaming operations worked with the same startup delay.
The system worked great on my Apple MacBook. The software was a little more complex than needed, but was helpful for a first time user. You might not actually need it. The system has a GPS to figure out where you are, then it helps you aim it. There is a magnetic compass on the top (which won’t work if you put it near a big metal thing like your car). It starts beeping and beeps faster as you get closer to the satellite. Our satellite was parked over Brazil and we were nearly at the end of the service range. Packet latency to our servers in New York ranged from 800-2300 milliseconds. Response times were scattered randomly on each packet, but it was consistent. It never dropped a packet. Email was fine. Web pages were like dialup on a 14.4k modem. Using web based programs like Gmail was similarly slow. SSH was painful and I found myself counting keystrokes to move the cursor around and change a little bit of code.
Several other types of Internet connections were supposedly available. I frequently saw people at center camp with laptops and I heard rumor of a 45 megabit microwave connection and another rumor of 3-6 megabit. While I was able to see a lot of access points, only half would give me a DHCP address and the other half could never route for me. Someone said distant sand storms affected the connection. I also saw a few people with WildBlue and Hughes backfeed dishes. Some fancy RVs have gyroscope based systems as well.
If you get a BGAN, make sure you watch your bandwidth. At $10/MB, automatically downloading that video someone emailed you, your operating system’s automatic update, your podcasts, etc. will run up a multi-thousand dollar bill in a hurry. Turn that stuff off and make sure you have a visual bandwidth meter at all times!
Next year, I plan to take a satellite phone (which ultimately cost less than roaming in Europe) and leave the BGAN behind, though if I go with a large enough group, we might get a Wild Blue and share with other people.
My final word of warning is that sand gets everywhere. I only brought out my laptop when there was no wind and it still got lots of sand in it just from what falls off your clothes and hair. Airtight ziplocks or Pelican cases are a good idea, but ultimately don’t bring any equipment where you would cry if it was utterly destroyed. This is a use for your 5 year old laptop.
Burn on, everyone.
2008-Apr-28 23:53:36 +0000 @037
Here’s some fun technical stuff. If you don’t program SQL, this will make no sense to you. Just move along. I wanted to put this out for the Google to find so it benefits other programmers.
I have some medium sized tables that I’m doing duplicate removal on and I tried a variety of techniques to see what works the fastest. The differences are amazing.
The server is a dual Pentium 1.8, 256 Meg SCRAM single Seagate SCSI 7200rpm drive running Linux and MySQL 4.x using MyISAM. I have a table with 2.8 million phone numbers and tried SELECT DISTINCT and GROUP BY with and without indexes. The phone number field is a BIGINT with NO NULL.
- SELECT DISTINCT with no index took 9 minutes 3 seconds.
- Using GROUP BY took 13 minutes flat with no index.
- The index took 18 seconds to create.
- SELECT DISTINCT with the index took 8 minutes 25 seconds.
- Using GROUP BY took 25 seconds flat (zero minutes).
WOW! What a difference the phrasing can make. I would almost consider it a bug that there is such a difference.
Keep in mind that creating this table without an index and doing the inserts unindexed was faster. Making the index once is much less work for the database engine than 2.8 million index updates.