Re: End of Chumby as we know it...

For a Flash movie from "domain1.com" to load XML data or another SWF  from "domain2.com", "domain2.com" needs to grant permission by hosting a "crossdomain.xml" file.

Re: End of Chumby as we know it...

Would that be needed even if the mirrors were assigned chumby.com subdomains?  If you wanted to make it more restrictive, you could have all the hosts be subdomains of .mirror.chumby.com, like doktorjones.mirror.chumby.com if I were to host a mirror; then as long as the crossdomain.xml grants permission to .chumby.com or .mirror.chumby.com it should work, right?

Re: End of Chumby as we know it...

OK, so for a new "widget" server to have those SWFs be hosted on other sites just means that the hosting site has the crossdomain.xml file -- nothing on the chumby.com server has to whitelist the domain.  That's good -- ultimately, a channel of widgets could be served as a bunch of HTTP redirects to other hosting instead of eating bandwidth from the "new" chumby.com.

Re: End of Chumby as we know it...

The sites from which the widgets collect XML data must have a crossdomain.xml file that grants permissions to the domain from which the widget is served.

For instance, if the chumby channel XML provides a link to a widget SWF hosted on the "example.com" domain, that in turn fetches data from the "sample.com" domain, then the "sample.com" server must provide a cross domain.xml granting permission to the "example.com" domain.

Many third party servers granted permission specifically to SWFs originating from "*.chumby.com" which means that the widgets themselves must be served from that domain, or they will fail to fetch the data.

Chumby's own content server does this - see "http://content.chumby.com/crossdomain.xml" - you'll see that it grants permissions only to certain domains.

Of course, it would be easy enough to add domains to Chumby's servers, or simply wild-card them, but that only helps widgets hitting that domain.  For the rest of the widgets, you'd need to ask everyone to update their crossdomain.xml files, and that's impractical.

The options then are reduced to making sure the widgets come from *some* chumby.com domain, or find a way to circumvent the crossdomain check.  There is a crossdomain disable mechanism in the Chumby Flash Player, but it requires lowering cross domain security (duh), and I'm not totally sure it would work across all widgets.

Re: End of Chumby as we know it...

Am I right in assuming that the widgets themselves don't actually take a vast amount of disk space?

When I ran the widget downloader program some months ago it produced a directory around 130MB in size - is that pretty much the extent of the disk space required by the mirrors?

Re: End of Chumby as we know it...

Yes, that's about right for raw public widgets.  With the private widgets, it's actually around 1GB, including thumbnails and configuration templates.  However, we're going to nuke the private widgets - if/when the full service comes back, people can re-upload any private widgets they want.

Re: End of Chumby as we know it...

I don't see the problem in setting up the mirrors as sub-domains either - as you said simply mirror01.chumby.com etc... I can't imagine mirrors are going to come and go too often. Besides, DNS changes are needed when new mirrors do come along to add them to the round-robin list, so why not add an extra A record at the same time?

Actually it makes much more sense to have mirrors on a dedicated vhost anyway so their paths can be consistent.

Re: End of Chumby as we know it...

I agree that it's not a significant barrier - just one to keep in mind.  In some ways it ends up being a way of controlling rogue mirrors.

Re: End of Chumby as we know it...

Right -- if a mirror stops updating, or begins hosting altered versions of the files or whatnot, it can easily be removed from DNS and will vanish in a few days.

With a little work, we could probably make a nice little administrative console that manages the DNS and rsync permissions all in one, and even does automated checks of the mirrors once a week or so to verify content... that way when someone wants to host a mirror, it's just a matter of punching in an IP address and some login info, and everything else happens automatically smile

Re: End of Chumby as we know it...

So far I can only count two mirror volunteers though - that's 2.5TB each yikes

I guess more would come forward once they realize the service is stopping....

Re: End of Chumby as we know it...

Do you have access to the extra Chumby "swag"?
If so, those would make a nice perk for Kickstarter donators if you choose to go that route.

Chumby Charms, stickers, bags, etc..
I would certainly donate $10 for a year of continued service without hesitation and assume many others would
as well.  For less than a $1 a month it should be a no brainer.

Re: End of Chumby as we know it...

I have *lots* and *lots* of  chumby charms and I think I have stickers.

I guess I could also put up the CafePress swag (cups, T-shirts, etc) for people to buy that I can get some income off that.

I don't have much else - 20 or so new Chumby 8's, a bunch of random parts.

I'm pretty much done with what I need to keep up the minimal service - now I'm working on a way keeping the whole service up and running.

Re: End of Chumby as we know it...

Duane,

I think Spoonman has a great idea. I would modify it a bit though. Maybe you could ask for donations and over a certain amount they would get a piece of Chumby memorabilia (I hate the term "swag"). Like NPR's donation drive without all the annoying pleading. Just a thought.

Do you currently have a way of accepting donations?

Tar, feathers, congress. Some assembly required.

Re: End of Chumby as we know it...

BoloMKXXVIII wrote:

Do you currently have a way of accepting donations?

Not yet - I'm waiting on the final paperwork for the new company.

Re: End of Chumby as we know it...

my opinion on the subject:

if it is possible (easily) it should give an offline version that will display at the currently active devices in the "control panel" the conversion. I suspect most users use only a limited number of widgets. (I use mainly as youstreams, clock, weather, RSS, image url ......) which will then be selected and saved as installable file.
I do not know if the authors of the widgets have to agree or will agfree, but this is possible

for the future, however, must be an installable offline version stored independently.

Re: End of Chumby as we know it...

Sorry to be chiming in here so late. I got wiped out by hurricane Sandy. Believe it or not, I did "evacuate" two of my Chumbys before the flood, left almost everything else behind. My PC and USB sticks with my customized doomsday Zurkware didn't fare too well after soaking in salt water.

I've just gotten back some web access, but I've been glancing at the forums via a dumb-phone regularly. Unfortunately, I didn't see this critical thread until today. Duane, if possible, could you list this "End of..." thread as a new forum rather than just part of the chumby.com forum? Or put a big banner on top of the forum home page? Others like me might be missing how important this thread is.

I agree with a couple of posters above wondering if it would help to get the word out to the 40,000 silent users. Can you insert an SOS info message like the short-lived ads were done? Those ads sure got everybody's attention, even if the response was negative. (We were so naive then.) I guess most registration email addresses are old, so that wouldn't work. Any chance of support from something like the Linux Foundation?

Unfortunately, I have no creative ideas to help other than those already stated. But I wanted to chime in that I would be happy to contribute money or time to help out.

I've been keeping my C1 unplugged for 2 months until I can move in to a safe new place. But I'm plugging it in tonight, in case these are the good old days. Duane and volunteers, thanks yet again for your heroic efforts in the real spirit of an open-source web. Please let the rest of us know what we can do, no matter how tedious or menial it may seem.
Bob

Re: End of Chumby as we know it...

I have considered trying to use the device ad infrastructure to deploy a message to the devices.

Unfortunately, when we were in the mode of reducing the traffic, I disabled at least part of the code in the Control Panel that hits the ad server periodically.  I could reenable that code and deploy a new Control Panel, however, nobody would get the new Control Panel until the next reboot, which doesn't happen very often with chumbys.  There is a way to remotely reboot the 3.5" devices from the server, but requires some manual management.

I'm also not sure if the ad code ever deployed to the C8 and I8 Control Panels.

As far as a mass mailing to users, that comes at a cost - and there would be an interesting question whether I should send email to those folks the have opted out of mailings.

Finding a way to contact the users is next on my list once I'm confident that I have the "worst case scenario" covered - the stub service I outlined in my first post.

Re: End of Chumby as we know it...

I could assist with the mass mailings if necessary; I'm sure others have the infrastructure to help too if necessary.  You could always title it "Important Chumby service announcement", with a short paragraph at the beginning explaining that this is a one-time notification, and that those who have opted out of mailings will not be receiving any further communications unless they log in and change their account preferences.

What about changing the control panel startup screen or home screen to include the announcement?  Some people still may not see it, but hopefully it'd get the word out to at least a few more.  Just have a short bit of text saying "IMPORTANT CHUMBY SERVICE ANNOUNCEMENT" and a shortened link that would be easy to transcribe to a PC.

Re: End of Chumby as we know it...

I agree that a hack to the CP is a good approach - there are three that would need updating (CC/C1/I3.5, C8, I8).  I should probably put out an update to Android, though I might be able to have it run the stub service and come back up to the full service if/when that happens.  That would leave the TVs, though there's a chance they'd work, I have to take look at the source code (not high on my list).

Re: End of Chumby as we know it...

I didn't know and I fairly regularly lurk on the forums. Just didn't see the thread on one forum

I'd have thought an email (even to excluded  individuals) is very acceptable in a service close down situation. It's massively the lesser of two evils.

Re: End of Chumby as we know it...

I will probably require the users to agree to a new terms of service to use their old accounts with the new service - the new company is not party to any agreement between the user and Chumby.  I guess I will also need to provide an option to destroy their accounts altogether.

As to users being "extremely annoyed", I'm not sure how they can be too upset since the alternative service is: "nothing".  There is no Chumby Industries any more, and hasn't been for some time. If they don't want to use the new service, they can simply turn off their devices or find some other alternative such as zurk's firmware.  I'm certainly not forcing anyone to use this one.

As to Pandora, my goal is to support it on the new service.  If you plan on supporting it on some other service, you'd need to somehow obtain a Pandora API key.

Just like the old service, the new service will require the device to be authenticated to access Pandora, so offline systems will probably not be able to use it - otherwise, that would violate Pandora's TOS.

Re: End of Chumby as we know it...

By the way, I have no problem with anyone creating an alternative to my service.  I'd be glad to document the minimal set of network calls that a booting Chumby does to get going.

90% of what I'm doing in the new service is supporting the calls that the current set of Control Panels make after they're started.

Anyone making an alternative does not have to implement any of that.

Re: End of Chumby as we know it...

Maybe I missed something but will existing accounts need to reregister with the new company and reactiviate their chumbys? Also is there going to be a way to activate/deactivate chumbys on the new system? Will you have a new website similar to the old chumby site to add apps if they come about?
BTW what is there to get angry about? Duane is making a last ditch effort to continue the life of our chums and people want to get angry for chumby inc not living up to its promises? Chumby Inc is dead. I am thankful someone is actually trying to do something constructive instead a complaining. I tell my kids "whiners get nothing".  Something is better than nothing. I for one congratulate Duane and the other fine techys who are assisting in the transition.

I am surprised CNET or someone hasn't got the word out from you that Chumby is making a comeback. Maybe that is a good way to contact chumbyers. Send an announcement to the various techy outlets with a clear story of what is about to happen. It is rather amazing that the ole chum may have a longer life than expected.

Owner of 3 Sony Dash, 2 Info 8.

Re: End of Chumby as we know it...

No, if/when that comes back, I'll try to preserve the existing accounts and channels, and you would not have to reregister your devices.  At most, you'd simply have to accept a new terms of service to reactivate your account.

If you all *want* to re-register accounts and devices, and start it all again from scratch, I can certainly arrange that.

Re: End of Chumby as we know it...

No, I would like to wake up one morning and just have to reboot and see that there is still daylight for my household chums. I do think you may consider unplugging any whiners though.  HA! It is going to be interesting to see how all this plays out. You did say that MY Streams should work right out of the gate? Is there a possibility YouStreams will be added early on? That is a very nice app.

Owner of 3 Sony Dash, 2 Info 8.