An AP story I wouldn't have expected the Globe to link from their Web front page:
Praying for other people to recover from an illness is ineffective, according to the largest, best-designed study to try to examine the power of prayer to heal strangers at a distance. The study of more than 1,800 heart bypass surgery patients found that those who had other people praying for them had as many complications as those who did not. In fact, one group of patients who knew they were the subject of prayers fared worse.
—Rob Stein, "No benefit of prayer found after surgery," AP via Boston Globe, 2006-03-31
DNS resolution for bothton.org, domain of Bothton (Boston Lisp announcements email list), is unreliable due to
DDoS attacks on Joker.com:
DDOS Attack on Joker.com Nameservers
Joker.com currently experiences massive distributed denial of service attacks against nameservers. This affects DNS resolution of Joker.com itself, and also domains which make use of Joker.com nameservers. We are very sorry for this issue, but we are working hard for a permanent solution.
Thank you for your understanding,
Your Team of Joker.com
If you're having difficulty subscribing or posting to the Bothton list, you can email me directly.
I'd been waiting all winter for spring, so that I could buy a new walkaround lens for my Canon Digital Rebel XT. After flirting with a 24-70/2.8, 17-40/4, 24-105/4 IS, and even a triplet of fast primes, I opted to try cheap IS while the Boston light is nice. So I bought a Canon 17-85 IS with the intention of trading it in next winter for a faster, possibly non-IS lens.
I haven't had a chance to test the 17-85 yet, but I can say that IS is spooky, the lens is the most front-heavy I've had on my XT, and the zoom ring is a little stiffer than I'd prefer.
In a month or so, it'll need a hood and a 67mm circular polarizer. I'm not sure I want to get an always-on UV filter, since the lens is already slow, and a high-quality 67mm filter is pricey.
In the last few days, some emails to me were delayed up to six
hours, and others were lost completely. This seems to have been due to
multi-day breakage of the Joker.com servers I was using. Today I moved my DNS
to Pair.com, and pretty much everything should be back to rock-solid.
(There's one more thing I need to do, after I hear back from Pair.com about a
possible issue with mod_rewrite, but nobody should notice.)
Messages from Verizon Wireless phones to email addresses trip many
spam-detector triggers. If you want to whitelist both "TXT" and "PIX"
messages, the two From header addresses seem to be:
TenDigitPhoneNumber@vtext.com TenDigitPhoneNumber@vzwpix.com
For example, to whitelist emails from phone number 617-555-1212, the two addresses would be:
6175551212@vtext.com 6175551212@vzwpix.com
When debugging why such Verizon Wireless messages are not delivered, be aware that Verizon's message status interface can claim 1-4 minutes after the message was sent that the message had now been delivered, and at a specific time -- when in fact a connection to an MX host for the addressee had not even been attempted. I verified this with a test mail server after wasting time with a hosting-service server that I could not monitor intimately. Header timestamps on one message that did ultimately get through showed a layover within Verizon's internal servers of many hours after the time the interface claimed the message had been delivered.
If you've experiencing problems with email delays to/from me, or with accessing my Web site today, it's probably due to all the nameservers at Joker.com being borked.
I'll probably move DNS to Pair.com, once Joker.com is back up enough that I can do so. I hadn't wanted to do that, due to a convention Pair.com has or used to have, but all Joker.com DNS dying at once is unacceptable.
Found a snapshot of dessert I'd taken at Pastiche when the girlfriend and I visited Providence earlier this winter. I'm not sure anywhere in Boston, even the North End, quite compares.
(The DOF of the photo is so shallow because I was trying to discreetly snap a single shot at maximum aperture.)
Incidentally, we stayed at The Old Court Bed & Breakfast on Benefit St, which was also pleasant.
This afternoon, I finally received my Motorola E815 that I ordered from Verizon Wireless 10 days ago, which prompted me to quickly make my Web site more friendly to "handheld" Web browsers.
(Review of the phone, firmware, and Verizon Wireless is forthcoming, once I have more experience with it. I'm already considering canceling Verizon, and moving to a different carrier and different phone.)
All that was required to make the site handheld-friendly was to
make a trimmed-down version of my CSS file, and use the appropriate
media attributes on XHTML link elements:
<link rel="stylesheet"
media="print, projection, screen, tv"
title="Default"
href="/neilvandyke-blueboxes.css" />
<link rel="stylesheet"
media="handheld"
title="Handheld"
href="/neilvandyke-bluehandheld.css" />
I'll fine-tune it when I have more time, but that few-minute solution got me 99% of the way there.
Dale Vaillancourt, et al., have released an DrScheme-based frontend for the ACL2 theorem-prover. (Initially, it sounded like a reimplementation of ACL2 in Scheme.)
Dale is working on getting it into PLaneT.
I just heard more info from Verizon Wireless on the backlog I mentioned yesterday (2006-03-08). The backlog is actually on phones and service ordered from Verizon's Web site, which, for reasons Verizon could not explain, are processed separately from telesales orders. Counterintuitively, telesales orders are processed much faster than Web orders, and do not have the huge backlog.
The CSR also said that today they are processing orders placed March 1st, and that my order would be processed next week at the earliest.
The CSR offered to cancel my Web order and connect me to telesales for placing a new order, which could be shipped in 2-3 days. She said the pricing would be the same, but the rebate on the phone, which on Web orders is an instant rebate, is a mail-in for telesales orders. I was in the middle of work things, so told her I don't have time for this right now, and would make a decision later. I added that waiving the $35 activation fee due to the delay would make me happier waiting for the Web order to be fulfilled, but she demurred.
If I didn't want the Verizon in-network calling with my girlfriend, I would've already ditched Verizon in favor of Cingular, Sprint, or T-Mobile. I'll decide after business hours today.
Three digital camera-oriented Web sites got together to define what a "review" is.
Quite a lot of other product areas could use review quality standards.
After I ordered a Verizon Wireless package on Sunday (see 2006-03-06) yet still hadn't received shipping confirmation by Wednesday evening, I called customer service. The CSR reported an order-processing backlog of 5-6 working days. She said the people who process Web orders had left for the day (they leave at 7pm, and I'd called a few minutes after), but she estimated shipment would be on next Monday, and shipped overnight.
Had I known about this backlog, I would've paid the extra $50 at the mall kiosk to have the phone immediately.
I finally put up a page, GNU/Linux on the IBM ThinkPad X20.
This joins my pages on putting GNU/Linux on the IBM ThinkPad 560E and Toshiba Satellite 4005CDS.
After almost 5 years without a mobile phone, yesterday I ordered Verizon Wireless service with a Motorola E815.
I haven't needed a mobile phone since most everyone I know prefers email, and I've been online almost constantly. However, the girlfriend is not an always-on email person, especially when we're trying to find each other. Secondarily, I've just started job-hunting, and it's good to be always accessible to prospective employers by voice.
Once I'd decided to get a cellphone, there was a nice-to-have requirement of being able to check on my servers and email without carrying a laptop everywhere. I was leaning towards a RIM BlackBerry with a full QWERTY keyboard and Idokorro Mobile SSH, but the required data service would've been about $50 a month atop the voice service, which I wasn't ready to cost-justify. I decided instead to get a phone and service that would give me simple Web-browsing, and to rig up some Web CGI scripts that will let me do limited email checking and do some of the other things I might want to do remotely. (Incidentally, I'm displeased with RIM for the industry-wide impact of settling that patent suit, although I can understand their business justification.)
Phone-wise (and before I'd decided on Verizon), I really liked the looks of the black Motorola RAZR V3, but an engineer friend who works with them could not say enough bad things about their RF performance. I also wasn't thrilled with their battery specs and on-screen user interface. Hence the Motorola E815, which is not near as cool-looking, but does not appear to have the RAZR V3's problems.
I tried to buy from the Verizon kiosk at the CambridgeSide Galleria yesterday, but he couldn't negotiate low enough on the phone cost, so I went home and ordered online. The main advantage of buying from the kiosk was that he was ready to get me their $5/mo. "Mobile Web 2.0" service. When you buy an EvDO phone through the Web interface, they offer you Mobile Web 2.0 only as part of the $15/mo. "VCAST" package. I decided to give VCAST a try. If I don't like it after a week or so, I can ask a customer service rep switch me to the $5/mo. or cancel my service.
I'll blog a review of the phone once it arrives and I've gotten some experience with it.
My Mozilla Firefox 1.5 had been slow for weeks. Clicking on an unexposed tab would result in a delay of 3 or 4 seconds before display, for example. The only extension I had installed was BugMeNot, but I'd previously installed and uninstalled several extensions.
On a hunch, I blasted all traces of extensions from my Firefox profile. Sure enough, Firefox immediately resumed being peppy.
I can't be certain that something with extensions was the cause, but I think that's not an unreasonable hypothesis. I'll leave the real work of investigating to someone else.
Now that the somber news is public (in a particularly poor MIT Tech article that I won't link), I'll mention in my blog...
Push "Pushpinder" Singh of MIT was found dead in his apartment Tuesday.
Push was one of my favorite people at MIT, and I can characterize this event only as tragic.
Site © 1994-2008 Neil Van Dyke neil@neilvandyke.org XHTML 1.0 Strict CSS2 Legal