I have two load-sharing servers. One of my two webhosters moved my site to a new datacenter, breaking that copy of my site in the process. (Their motivation was to improve reliablity. ahem. ) So half the time, CSC users were looking at broken webpages.
I've switched the traffic over to the one working server, so CSCs should be displaying again.'
However, it might take me a couple of days to fix the root cause of the problem. During that time, if you can't see a CSC image, try using the url http://server2.cleardarksky.com (instead of http://cleardarksky.com). But don't book mark it. I reserve the right to stop supporting server2.
Time to shop for a new webhoster.
In a moment of complete-lack-of-imagination, i've decide to call them "clear sky charts" instead.
Annoying. But this is the reality:
Note: As you read the rest of this file, please replace any occurence of the word "clock" with "chart".
(I'd do it myself, but I hate re-writing history.)
I'm re-writing a noticeable amount of infrastructure. So if you see some of my web pages garbled, or unreachable; or if some of the click or mouse-over features dont work: please wait 15 minutes and try again. If it's still hosed (that's the technical term), then please email me.
I'm still working though a large email backlog. But I will reply --eventually-- to everyone who has emailed me.
However many of you have been asking for details of my legal problems. Sorry about being vague earlier. I didnt want to say anything until I had found legal advice and figured out what to say to the complainant.
So this is what happened so far:
On June 28, I recevied an email from Skyclock company. I understood the email to mean legal action against me over the use of the word "skyclock" for which Skyclock company has a trademark registered in the USA. I have zero experience with trademarks or legal conflicts so it took me over a week to calm down enough to find legal advice. In the midst of all this angst, I received a second email from another individual at Skyclock company.
This email was polite, considerate and reasonable. Since then I've spoken with a Mr Rob Baxtresser of Skyclock company and find him to be a consderate and reasonable person. Since this was a sharp contrast to my first impressions of Skyclock company and since many of you were incensed at the thought of harm being done the CSCs, I thought CSC users should see Rob's words.
Reprinted with the permission of Rob Baxtresser.
|
|
So it would appear that Skyclock company an the CSCs will not be in conflict. (Now I can start regrowing my stomach-lining.) Even better, it looks like conflict will be replaced with cooperation.
I'll still have to add "legal bills" to the list of things that sponsorship help pay for. But my legal bills look like they will be finite and will mostly go towards perparing for the future.
Again I wish to thank everyone for their considerable and kind support. And I'm very glad to say that i'm still generating the CSCs.
After a huge amount of stress, I can now say that it looks like I will not be sued by an USA trademark holder for trademark infringment. I will post details soon.
But for now I'd like to express my thanks and awe at the enormous support from CSC users. You guys are great. I have a very large number of people to thank. I will also reply to every email I've received, however I have a big backlog. Please forgive me for not following up in all newsgroups.
Special thanks to Norman Van Treeck.
However, I've been threatened with legal action by a company in the USA over trademark infringement. Personally, I can't even figure out whether USA or Canadian law applies, or whether I need zero, one or two (one for each country) lawyers.
Legal fees are something I've never faced before in operating this site. It seems I must soon add "legal fees" to the list of things that sponsorships help pay for. So now is a time when sponsorships are particularly welcome.
Further reports, and possibly louder cries for help, as events develop. As always, I will continue to do my utmost to keep the CSCs updated.
The CSC website is no danger. It's just a server problem, though it's an annoying one.
I buy hosting from two different webserver companaies. Someone is sustaning a denial-of-service attack on one of my webhoster servers, though not specifically on the CSCs website. That webhoster is responding by changing their ip addressess -- without telling clients like me. That explains why only some CSC users are having problems. It also explains why some CSC users are being routed to the wrong website.
I have manually switched all CSC traffic to the other sane server.
However, the problems may recurr for a few days as I experiment to fix the bad server. So if you continue have problems with the CSCs, try the url http://server2.cleardarksky.com. (But dont edit web page links to point to that url, since it is temporary). If you find server2.cleardarksky.com doesnt work either, dont hestiate to email me.
The analysis of new observations is on hold untill some database problems get fixed. You are still welcome to enter observations, but they wont be listed on maps untill the database problems are solved.
CSCs in Saskatechewan and Arizona, and other places that never use DST are not affected.
CSCs in Mexico do use DST, but they start DST in April.
To check that a CSC is properly set up, look at upper left of a CSC image under the words "Local Time". You should see something like "GMT -4.0" for the eastern timezone with DST, "GMT -5.0" for Central, etc. Email me if you see an CSCs with the wrong GMT offset.
Now you can make a color-legend appear just below the CSC image by clicking under "image control" (just to the left of the CSC image) on the check box labeled "Display color legend".
The problem is now fixed.
I'm now plotting CSCs on zoomable an panable Google maps. There is one map per state or province. The maps for California and Ontario take about minute to display the map. Also, there is no single Google map that plots at 2800+ CSCs. That's because Google's api doesnt scale well. Such a page would take 10 minutes to draw the map.
To see CSCs plotted on a google map, goto any list of CSCs in a state or province, and click on the "Map (may be slow)" link.
example
Unfortuneately this does not mean that anybody's sky has gotten darker.
Thanks to Jack Liebenthal for pointing out the error.
I've converted all Indiana CSCs to DST. However, some counties have also changed from the eastern timezone to the central timezone. I may not have found all the affected CSCs. So please let me know if you see a CSC with the wrong GMT offset. (It's written in small print in the CSC image under "Local Time".)
No word yet if I can get coverage restored.
A few CSCs, mostly in Alaska and South Texas, now get the cloud forecast. Previously, they only had transparency and seeing.
CSCs now update with new cloud and transparency data about 20 minutes earlier than before. However, only the CSCs with the best update order (i.e. sponsored) will see the benefit since I'm still struggling with limited upload bandwith. It will be about a week before I can install a replacement router.
I've had to re-write a fair amount of my code to accomodate the conversion. The basic CSC functionality works, but there may be a few bugs in the less-used stuff. The javascript errors people were having yesterday should now be fixed. Let me know if you see any more. In particular, I've temporarily stopped updating the CSC climate/history pages as I still need to do lots of work for them to use the new forecast data.
These new astronomy-forecasts at CMC are now officially one of their products. Check out their website.
In difference of opinion like this, the webhoster always wins. My only recourse is to buy webhosting space from another webhoster and move. I have done so.
The offical Environment Canada astronomy weather page is http://www.weatheroffice.ec.gc.ca/astro/index_e.html.
It will take me a some time to convert the CSCs to use data from the new images. In the meantime, I will continue to generate the CSCs from the old stream of data.
Looks like I have a lot of work to do upgrading my computers.
The server outages and resulting email backlog means that there are a number or requests for new CSCs that I have not yet read. I'll start creating new CSCs as soon as I can.
Sorry about the mess.
I make it a point to reply to all email. If you've sent me an email recently, and if I haven't replied in a few days, it's probably because I lost your email (spam filter problems). Please resend.
I'm now displaying Allan Rahill's temperature and humidity forecasts on all CSCs. These forecasts are generated by the same high-accuracy model that generates the cloud, transparency and seeing forecasts.
To keep the new forecasts from cluttering the CSC images, I've made some format changes. I've also changed the legend pages a bit. So email me if you find broken links, bad splleing, or non-functional features.
I've also made changes to the machine-readable ascii CSC files. If you use them, email me for details.
I'd like to thank sponsors for helping me provide the extra computing power and hosting space needed to handle the new forecasts.
to access CSCs, please note that none of those urls work properly anymore.
Please use the offical url: http://cleardarksky.com/csk only, especially as I seems that I have to keep changing web servers. But I'll always keep http://cleardarksky.com pointing at the right server.
In keeping with the notion that blue blocks are good, Allan has labeled calmer conditions with blue.
However, it's just barely possible that the change to my name servers has not propagated througout the internet yet. So, if you get error 404 when you try to load any page from cleardarksky.com, try quiting and re-entering your browser. Also, if you have any direct links to the old server, like http://66.78.6.237/...., replace them with http://cleardarksky.com/
The history pages just summarize past forecasts, but they sure make me wonder if there are better places to observe than where I am. To aid people, like me, who are looking for climate data for other CSCs, I've changed the coverage map to subtly indicate which CSCs have a forecast history page.
In the interests of fun, I've added a CSC history page to selected CSCs. For example.
The url http://cleardarksky.com switches automatically to point to both servers, or just the working one, if the other fails. It's clever, but the switchover takes about an hour.
To give people the choice, I used to publish direct urls to each server directly. One was http://leo.lunarpages.com/~cleard2.
Please dont use http://leo.lunarpages.com/~cleard2 and http://server1.cleardarksky.com any more. Because of obstruse software interactions, none of my PHPs, including the ones for displaying CSC images, will work via those urls.
This page will always list the currently valid main and alternate urls for getting to the CSCs.
This is probably the first time so many peope will use CSCs to forecast transparency for an event right on the horizon. So I'd like to remind people that the CSC colors represent a calculation of the transparency of a *vertical* column of air above the observer.
To make sure you are reading conditions correctly for the transit, look at the full forecast map for the hour of your observation. Look at the few pixels immediately east of the cross marking your location. If that is clear, you may still be able to observe, even if the CSC is showing white.
To see the full forecast map, goto the legend page for your CSC (browse or search from here) and click on the colored block for your hour of observation.
And remember, never look at the sun without proper eye protection. If you don't have have a properly filtered solar telescope, watch the transit on the web.
Readers unfamillar with the transit of venus might want to start here.
However the http://server1.cleardarksky.com/csk alias for the CSCs is gone. Use http:/cleardarksky.com/csk instead for the CSCs.
The seeing observation database browser functions are down because of server config problems. I have no idea when I can get them fixed. In the meantime, it is still possible to enter new seeing observations. (I just cant re-display the them).
The CSCs, however. Are still working.
To make it easier to see timezone errors, I've added the timezone to each CSC image. It's in small print just below the words "Local Time". Please email me if you see any CSC images with the wrong timezone.
I hope no one noticed.
The new server is automatically load sharing with the old one. So no one should have noticed any changes.
But this means that should one of my webhosting servers fail, (or get messed up by the administrators of a certain webhosting company I'm too polite to describe in more detail), the other hoster will take up the load.
If that happens, you may see 404 or other weird error messages about half the time. That's because half of all CSC users would being sent to the dead server. But if you keep exiting and re-entering your browser, you will eventually be connected to the good server. After a few hours, I will automatically re-configure my name servers to point only to the good server. (Takes 15 minutes for me to switch, but a few hours to propagate through the internet.) That should cause all the errors to disappear. Some time later, the dead server will come back up and resume load sharing. It should be *very* unlikely that both servers will be down at the same time -- even on a new moon friday!
People who dont want to reload pages in order to get to the good server when one is dead, may want to access them directly through alternate urls. Server status and alternate urls are posted here.
Don't panic. My plan is to move to a new webhoster with no interruption the the delivery of CSCs.
I plan to do the switch over during the early hours of April 6th. Users should continue to use http://cleardarksky.com to
get to the CSCs.
Ok, for some that's not enough. Here's some more detail. I've rewritten my CSC generating infrastructure
to upload to two mirror sites. The new server is already up and is being updated with CSC forecasts in
parallel to the current server. I'm hopeful that the new server will be much more reliable than my
current server which is difficult to describe in polite company.
On April 6, I plan to switch DNS configurations to point cleardarksky.com to the new server. That means
requests for CSCs will automatically be routed to the new server. In the event the new server fails,
quit your browser and retry to load a CSC page. Eventually you'll be routed to the old server which
will still be running as a backup.
In the event that the DNS reconfiguration has problems, here are aliases to both servers. But I will be working to make cleardarksky.com to always point to the best server. So dont make permanent links to these;
Aliases for New server:
Aliases for Old server:
I will fire^H^H^H^Hretire the old server sometime before Septempber, so dont keep permanent links to
http://68.78.6.237
In the event of a total disaster, remember that you can still get the astronomy forecasts, in map form,
from Allan Rahill's site:
http://www.weatheroffice.ec.gc.ca/astro/index_e.html.
Those alternate urls are not working properly. Please use instead:
In the event of a total disaster, remember that you can still get the astronomy forecasts, in weather map form,
from Allan Rahill's astronomy-weather site at CMC:
.
Now that sounds like "famous last words".
However, since I'm paranoid, I will post a switch over date here (when I'm ready). I'll also post
work-around urls for both old and new servers.
If you're even more paranoid than I, feel free to ask me to email you to remind you of the switchover a day before it happens.
All CSCs were down for 24 hours. People who saw the "account cancelled" message should not
take it personally. My webhoster was talking about my account with them.
As I thought, the problem was simply the admin division of my webhoster being confused. They restored my
service without explanation. This hasn't been the first time. [....unchartiable remrarks deleted in the interests
of maintaining some semblance of decorum ... ]
Soooooo. I've bought space on a new hoster and am working on transferring the site. I will try to do the transfer
without an interuption in CSC service and without changing any urls -- I hope. I'll pick a time which
isn't a weekend or a new moon! I'll post plans here when I have them.
I've also changed registrars and am looking into getting a mirror site.
I'd like to thank all of the many people who wrote to offer supprt.
In other news the WAP or cell phone CSCs I released a few days ago are now obsolete. New model cellphones
with WAP 2.0 browsers can display the graphical CSCs. So CSC users with the latest uber-geek equipment
can dispense with the clunky WAP CSC and just use the preview (also called "icon CSC") images.
However, since most of us still have WAP 1.0 cellphones, I'll be generating the text CSCs for cellphones for a while.
A small step towards getting CSC forecasts to observing sites: the WAP or cell phone version. If you have a web-enabled cell phone (aka WAP capable cell phone) you can now
read a simplified text version of CSCs whenever you are in a digital coverage area.
Because of costs of generating and hosting an extra copy of the CSCs, only sponsored CSCs have a cell phone version at the moment.
Look in the nifty links on your CSC's legend page.
A new feep to help solve this problem:
As my S.O. would say "feep...feep...feep".
So instead of automatically closing the popup after your mouse has left the colored block, I leave
it open. That allows you to move or resize the popup window so you can see the numerical fields properly.
You may need to uncheck and recheck the "Show block details when you mouse-over" to regenerate the popup
window in some cases.
I've rewritten the descriptions on the lightpollution pages. Feel free write me if you think your site's lightpollution is significantly different that what the maps show.
2004 April 2
Moving to a new server on April 6
http://leo.lunarpages.com/~cleard2/csk/
http://66.78.6.237/~cleardar/csk/
2004 March 29
Some people using two alternate urls that I posted in September:
http://www.pobox.com/~clearskyclock
http://66.78.6.237/~cleardar/
2004 March 25
I've purchased space from a new, more expensive, and more reliable (I hope) web hoster. I'm working on
transferring the site. My current plan is to do the switch with no visible change to any CSC users.
2004 March 21
Webhosting Follies
2004 March 9
I've found a way of getting USNO's sun & moon data without revealing he latitude and longitude of a CSC site.
So I've re-enabled the "sun & moon data" link for CSCs with location privacy.
2004 March 6
Whenever I goto star parties, I'm frustrated by the lack of internet access. I need to see my CSCs.
(No I'm not addicted. I can stop any time I want.... I just dont want.)
2004 February 3
Desparate for photons? Clouded out? Willing to drive a 100 miles? You can see a tantalizing clear spot on
the forecast map but cant tell exactly where it is?
Only works from non location-privacy CSCs.
2004 January 7
The red "No Data From CMC" banners on all the CSCs is because of a problem with the astronomy-weather
computer model running at the Canadian Meterological Center. CMC knows about the problem. As soon as they
resume modeling, my software will detect the new data and will automatically update all the CSCs.
Unfortuneately, I cant say exactly when that will happen.
2004 January 4
In legend pages, when you check "Show block details when you mouse-over" and then run your mouse over a colored block in the CSC image, a little text window pops up to show numerical details of the forecast for the hour
of the block. However, browser incompatibilities have made it too frustrating to position that little window
properly.2003 December 21
A small feep: I'm now showing the locations of other CSCs on each CSC's lightpollution map. You can
go to those CSCs by selecting "on click: find CSCs" in the lightpollution map page. Then click on the little
crosses that mark the other CSCs.
2003 November 30
A couple of people tell me that the light pollution maps are wrong, being too pessimistic. In reviewing my own observing notes and the bortle scale, I conclude the maps are about right. But some of the descriptions on my
light pollution pages were out by one whole catagory.
2003 November 10
I was originally inspired to create the first CSC because I was too lazy to download all of Allan Rahill's
forecast maps. But I've been surpassed (in lazyness).
A user has asked for a faster way of seeing the light pollution rating of a CSC. Apparently he got excited about the forecast for a friend's CSC only to realize later that it wasnt as dark as he thought. Clicking the "light pollution map" link would have been too much work.
I've added a light pollution column to CSC listing pages and search outputs. But I'm not convinced it's it's worth increasing the clutter in the CSC images with another color code. So, I put the color code for the CSC's light pollution rating into the text portion fo the CSC's legend page.
For the tech weenies: Mark is using the machine readable versions of the CSCs. I wrote them just so 3rd parties, like Mark, could come up with new applications. If you have other ideas for machine-readable CSCs, let me know.
This feep isnt available for CSCs that have been requested as "location private". It would have been easy to find the exact location of private CSCs otherwise.
Thanks for everyone's patience during this mess. Special thanks for the very many people who offered support, help and added new adjectives to my vocabulary.
(I knew the machine-readable version of the CSCs would come in handy some day.)
I'm quite amused -- really. I started the CSCs because I was too lazy to download all of Allan Rahill's astronomy forecast maps. Mark has gone one step further.
Some people have reported that, on CSC legend pages, the text appears, but the CSC image does not. In all cases so far, the problem has been that some ad-killer software, added by users to their browsers, have decided that the CSC images are advertisements.
To see if you are having this problem, compare what you see with these two links: Link 1 and Link 2.
If you can see the CSC with Link 1 but not Link 2, then you probably have ad-killer software that needs to be told not to nuke gifs from cleardarksky.com.
If you cant see the CSC with either link, then my webhoster is probably down.
I'm converting the CSC generation system to use the new hourly data. The CSCs will be updated this evening. There are a few bugs in the "next" "previous" links of the full-map display that should be fixed tormorrow evening.
The problems with the numerical model at CMC have been fixed.
Originators of existing CSCs are welcome to ask me to convert them to French.
I could also support Spanish (same limitations) if anyone is interested.
Now serveral users of one CSC have reported it not updating for as long 5 days.
Here is my third and hopefully final attempt to prevent CSC .gifs from being cached. It appears effective. But people who want the fix for CSC images displayed on their own web pages need to change their html. Use the sample html at the bottom of the legend page for your CSC.
If your CSC appears older than 12 hours, and you are using the current html to display CSC, please write me.
This change has also made the text of the legend page load about 1/3 faster.
I have this feeling that the number of CSC requests that I am getting is about to increase. I figure I will eventually get requests for CSCs that will not really be used. So I reserve the right to delete CSCs that I believe have fallen into disuse. For the moment, by definition of "fallen into disuse" is "the legend page of the CSC hasnt been used in 30 days".
If a CSC you care about has vanished, the remedy is simple: request it again and it will live for at least another 30 days.
Sorry, I can't make all news items as exciting as the Oct 9th entry.
Last week, news of the imminent cancellation of the CSCs made it to usenet and a number of astronomy mailing lists. That news was followed by calls for emails to be written to to certain civil servants, followed by news to not write civil servants, followed by the news that everything is fine. Now that we can breath again, here is the story:
The struggle had been quitely going on for a year.
I generate the CSCs from publicly available forecast maps from the Canadian Meterologial Service. They are the creation of Allan Rahill, CMC meterologist, and are incredibly accurate. But they are not an official product of the CMC. Allan created them on his lunch hour. Well, during a couple of years of lunch-hours. He always knew he was taking a chance releasing a unofficial forecast.
In August, a department in a different division than Allan, told him that his forecast was in violation of Environment Canada's policy against specialty forecasts and unofficial use of government computers. Allan was told to take down his forecast. That would have starved the CSCs of data within 36 hours.
The CSCs themselves were never in danger since cleardarksky.com and I have no connection to the CMC. However, Allan's forecasts are simply the best for astronomy. I don't believe I would have been able to find a replacement accurate enough for the hourly display of the CSCs to be useful.
News of the status of Allan's forecast hit the net last week. The situation was exaggerated to appear to be a crisis and was followed by calls to write Allan's directors. In an attempt to keep Allan from getting fired, I posted request to re-direct the traffic to me. Allan posted that emails were in fact welcome, but that they should be sent to him. In the confusion each of Allan, I and two CMC directors each received between 100 and 200 emails over the course of three days.
It turns out that it was not a crisis. Allan had been planning to present his directors user-written emails of support. Allan did not get in trouble. His directors actually expected to receive some emails directly, but not in the hundreds. Allan presented them with the full set of emails from people supporting his forecast. There were over 2000. Almost all of the people who wrote those emails said they used the astronomy forecasts though my CSC interface. (I think thats pretty cool.)
The volume and geographical extent of the emailed support for the astronomy forecasts was a factor in the CMC directors deciding to work to make the Astronomy forecasts into an offical CMC product. That might take a year or so. In the meantime, they will leave Allan to continue to develop and tune his forecasts.
So the short story is: Yes the CSCs were in danger of disappearing. They are safe now. Huge thanks to all those who wrote in support.
As long as Allan can generate his astronomy forecasts maps (and as long as they are the best :) ), I will continue to create CSCs from them.
We now return you to your regular, more boring, web site news.
The "CSCs within 100km" page lists distances from the orgin CSC.
Some may note that some distances are labeled as "<24km"or "<15mi". Those are CSCs for which the owners of the site requested that I not reveal their their precise location.
I've added a "light pollution map" link to each CSC's legend page. It's a small exerpt from the 130meg tiff image of North America from The World Atlas of Artificial Night Sky Brightness centered on the CSC site. Used by the gracious premission of the authors: P. Cinzano, F. Falchi, and C. D. Elvidge.
I've created a public mailing list about weather forecasting for astronomy. I've been inviting those few meterologist who I know are also astronomers to join in the hopes that I may bathe in their combined wisdom. Or maybe just get an answer or two. We can also use the list to discuss usage and accuracy of weather forecasting websites, like the CSC, if we feel like it.
Help tune the seeing forecast used by the CSCs by submitting your observations of astronomical seeing. The database in currently in beta test. Everyone is welcome to play with it and report bugs or suggestions to me.
I will reset the database and make it ready for real observations around August 20th.
If you'd like a CSC for your observatory, but are worried that the CSC would make your site's latitude and longitude too public for comfort, there is a solution. Mention your desire for privacy in your email to me and I do my best to not reveal your latitude and longitude on cleardarksky.com.
I've added reminder in May I get my own CSC? and a link to the privacy policy that explains the choice.
The people at Observatoires du Mont Megantic have asked be for a less "pupil destroying" background color for the background of legend page for their CSC. Apparantly they've been projecting it onto a huge screen at night. They seem to prefer these colors which are a little easier on night vision.
The orginal requestors of any CSC are welcome to ask me to set their legend page to the same color set. However, since having separate colors for each page is a pain, I will take each request as a vote to change them all. Even if you've never requested a CSC, email me if you'd like all CSC pages in night vision colors -- or not.
