I remember when Telus shut down thousands of websites to take down their union’s website. Serverloft might have just done the same. The Seattle Post-Intelligencer has an article on the takedown of the Yes Men site. I did find an article from straight.com which claims to have the original complaint. The IP address in the complaint there is the same one that ec-gc.ca is currently hosted on. Ole Tange is the contact for PiWeb listed in the IP address for that whois.

At first I was convinced that Serverloft did in fact pull the plug on PiWeb’s netblock. A slashdot comment gave me pause. After all, the message from Mr. Tange is on a website controlled by the Yes Men, and his contact details are publicly available. As of this writing, the PiWeb status page makes no mention of related downtown. Maybe the Yes Men took down their own site as a PR stunt. Here’s a paste from the disputed website.

Website suspended

Serverloft blocked the IP-range for this server because of the content of the client’s website and would only unblock the IP-range if we suspended the website. The website was used in a spoof by The Yes Men.

Serverloft blocked the IP-range without a warrant and without calling us and thus affecting servers hosting 4500 of our customers’ websites until we ourselves discovered the problem, and convinced Serverloft to unblock. Serverloft did send us an email explaining that they would not unblock the IP-range until the websites were taken offline. The email was sent 5 minutes after they cut of the access to the mail server, so we only received the email after the 4500 websites were back online.

Convincing Serverloft that their systems had blocked access on purpose was hard because Serverloft frontline support claimed that all their systems were working fine and they therefore assumed that the problem was a configuration problem on our server. They refused to help troubleshooting the issue.

Serverloft could simply have called us and asked us to deal with the situation. We would then have asked the Canadians for a warrant. If the Canadians had shown us a warrant we would have taken down the site immediately. As others have pointed out the Canadians could probably just have gone through CIRA and have the domain suspended, which would not have affected any of the other 4500 websites.

As we cannot go through every single page that our customers put on their websites we anticipate a similar situation may arise again. We have therefore asked Serverloft to revise their procedures so we at least would get a phone call before they cut our connection. They have so far refused to do so. They have answered:

your net was blocked because of hosting phishing sites. I’ve attached the information, we have, below our signature. I’m sorry, but we cant call every costumer for abuse. In some cases we’ve to respond very fast and have to block the net or server.

While I appreciate Serverloft respond fast, it is no good if the collateral damage is more than 1000 times as big. Had they called I am sure we would have found an arrangement that would satisfy both of us.

For more information: contact Ole Tange <ole@tange.dk>

The only other website I can find running on the listed IP is good-cop15.org, which is also the default vhost for that IP. It’s a site about the Copenhagen conference and links to the Yes Men hoax. Judging from the reverse lookup on the address it’s probably a virtual machine, quite possibly run by the Yes Men. It turns out that there is actually a Bruce Moore at CCIRC, but he wasn’t on duty when I called. Ec-gc.ca isn’t on any phishing list that I can find so that charge was overblown.

I don’t particularly like the Yes Men. They go from satire to impersonation. I’d have loved to see CIRA close down their ec-gc.ca domain but the domain dispute takes a long while. I can understand the letter from the CCIRC. The Yes Men website will do the most damage now, not in the time after CIRA is through with its dispute process. I find it very plausible that the Canadian Government would send such a request.

I also find it very plausible that Serverloft would issue a knee-jerk takedown. It’s relatively common for service providers to take down material immediately after a complain, especially among low margin resellers. From the WHT opinions, it sounds like that fits Serverloft. If that IP address was limited to a virtual machine run by the Yes Men, and Serverloft was okay with risking their reputation by acting on the complaint rather than following up with PiWeb, they could have at least minimized damage by only nullrouting the VM’s IP address.

There’s enough negative material on WebHostingTalk that I wouldn’t want to use Serverloft regardless of the outcome of this case. I’d like to see some official statement from PiWeb on an official channel before I believe the details though.

Update Jan 2nd 2010: It looks like Ole Tange has updated his personal website with details of the takedown.

Related articles by Zemanta

• Yes Men take credit for Canada climate hoax (nationalpost.com)
• Canadian Censorship Takes Down 4500 Sites (yro.slashdot.org)

After a bit of play with Wave my first instinct is that it’ll be great for collaborating on taking class notes. The problem is that most of my class notes involve pretty big formulas. The solution? I’m writing a gadget that takes in LaTeX code and returns a PNG suitable as a blip. I’ll be using LatexRender for that part. The downside that I can immediately see is that most of my lecture halls don’t have wifi signal, but maybe releasing the gadget will be useful to someone out there.

I don’t yet have my developer sandbox account but I’m working with the Wave gadgets tutorial. It looks like an AJAX-enabled textbox will do the trick. The state handler will check if there’s a pre-rendered PNG; if so it’ll display that, and if not it’ll query the PHP script. I’m hoping that I’ll be able to instead replace the gadget with the static PNG render once the editing of the Wave is done but I haven’t discovered the hook for that yet. I also hope that the state handler will be called immediately after init, and I think that’s the case, but judging from the documentation I’m not entirely sure.

I’m hoping to have a simple version of the gadget that just takes LaTeX within two weeks of getting my sandbox. The pipe dream feature would be a WYSIWYG editor that outputs LaTeX, so for the foreseeable future I’ll just be copying and pasting from LyX.

Related articles by Zemanta

• Google Wave Is Easier To Understand Than… (techcrunch.com)
• Google Wave’s Little Secret: It Already Works On The iPhone (techcrunch.com)
• Google Wave: There Will Be Backlash (techcrunch.com)
• Google Wave Wednesday (jeffisageek.net)
• Why You Haven’t Got You Google Wave Invite Yet (jessenewhart.com)

I decided that I would install my own copy of Weave because I’d rather store my information on machines I own. The Mozilla folks are very unlikely to let my passwords get stolen or misused but that bit of paranoia hasn’t let me wrong so far. The installation of the Weave server scripts isn’t for the uninitated. Since it requires modification of apache configuration files you won’t be able to do it on a shared hosting plan, though you might be able to get around the aliases with some mod_rewrite voodoo.

It turns out that my paranoia isn’t warranted in this case. You only upload encrypted data. The data is encrypted with a passphrase separate from the Weave login, and the server never knows it. I don’t know how they implemented their encryption (other than the cipher is aes-256-cbc according to about:config) but I imagine the data is safe. The Mozilla folks say they can’t get at your data.

Weave is still too new to have a plan set in stone but Mozilla does have an idea of what they want it to do. The preferences for the client side extension list these as syncable items, though some are greyed out and don’t work: Bookmarks, cookies, extensions, forms, history, input history (location bar), microformats, passwords, plugins, preferences, tabs, and themes.

It looks like they’re trying to support OpenID too, but unfortunately it didn’t work on either my Basecamp login (it actually prevented me from login) or my WordPress install. The OpenID text field actually disappeared and was replaced with a “Sign in using Weave” button. I took a look and as it turns out, Weave assumes that you’re using Mozilla’s OpenID provider with the same username that you’ve signed in to Weave with. Not very helpful when I run an openid provider on imaddicted.ca and use tail.recursing.org for Weave. Fortunately you can click “(revert)” next to the button to get the OpenID text fields back. If you don’t use Mozilla’s OpenID provider you may as well change extensions.weave.openId.enabled to false in about:config. I’m sure that this behaviour is just a temporary thing while Weave is in alpha.

Fennec in need of Weave. Image by Getty Images via Daylife

The client side extension will also install in Thunderbird and Sunbird but won’t do anything as of yet. Fennec support is underway. Thunderbird would be easy to sync, keep your passwords in the cloud and let IMAP worry about which messages have been read. Sunbird is a little trickier. Will they sync calendars, or just the login details for shared calendars? Weave was intended to sync just metadata but calendars are more data than meta.

Weave has me intrigued. I see a lot of potential, especially in having a single sign-on for OpenID. What I’d like is for all Mozilla products to ship with Weave support. I want to be able to sit down at a friend’s desk, sign in to my Weave account (hosted on my own server, naturally) and have all my extensions plus settings and bookmarks at my disposal. I want Sunbird to snag all my calendars. I want Thunderbird to work. I want my Mozilla experience to be easily accessible whatever machine I’m on. I want to get rid of my USB key that contains Portable Firefox. I want it to work on Linux, OS X and Windows.  This is somewhat similar to Dion Almaer’s take on Weave in Ajaxian, although in its current form Weave is still useful to me.

Take a look at the screencast demo. Want to give Weave a go? Check out its Mozilla Labs page.

Related articles by Zemanta

• Weave 0.3 Released (labs.mozilla.com)
• Weave: Firefox Bookmark & Passwords Sync Add-on (shankrila.com)
• Mozilla’s Weave: (Too far) Ahead of its time (news.cnet.com)
• Who do I trust with my Identity? Erm, how about me? OpenID Weaves into the browser! (ajaxian.com)

I’m not sure I agree. The value of social bookmarking sites is heavily rooted in the network effect. Otherwise, they’re not really social bookmarking sites, just online bookmarking. In delicious‘s case they still have a big enough crowd for a popular link to really mean something, and enough of a bookmark base to power a really slick suggestion engine.

Right now Pinboard doesn’t have either of those. That’s not to say it’s worth skipping over.

There’s no spammers (yet). It’s not a juicy enough target (yet), presently pagerank 0. The small entry fee is a deterrent to the people who would usually register dozens of accounts to make hundreds of bookmarks to their network of sites.

Greasemonkey script suggesting tags in Pinboard

Delicious’s advantage of its huge user base powering its tag suggestion engine can be harnessed with a little help of greasemonkey. On the same day I registered I cobbled together a quick script that puts delicious tag suggestions into the pinboard bookmark popup. I’m sure eventually delicious will put a stop to it, but in a couple months I expect it to be moot, Pinboard will have the community necessary to power their tag suggestion engine.

Last but not least, the to-read feature of Pinboard is pretty slick. The Read It Later Firefox extension is very cool but I found it too heavy for what I wanted. I didn’t want to create yet another account. I didn’t want offline viewing mode, which ate up a lot of space on my university computer^[0]. I didn’t want

Pinboard’s toread is elegant. It’s a tag separate from other tags, so it won’t pollute the tag suggestion system.^[1] When I want to browse my reading list, I just click the “read later” bookmark. Simple. All I need to do is write up another greasemonkey script (or maybe a bookmarklet) that takes me to a random member of my reading list and I’m set.

Pinboard is simple and elegant but occasionally I’ll still go back to delicious for their popular page. If a resource makes it onto there it’s generally authorative. For my bookmarks I think I’ll stick to Pinboard for now.

Related articles by Zemanta

• Krumlr.com – A Fresh Twist On Social Bookmarking (killerstartups.com)
• Diigo – Web Highlighter and Sticky Notes, a delicious killer? (academicproductivity.com)
• Be Careful What You Save (davefleet.com)

I discovered Ohpan by A Type Studios through Twitter and decided to give it a go. You can too if you check out their latest blog post to find one of the beta keys. The ones that work as of this writing are ATYPE1 and GENYTO.

It’ll ask you to log in through an OAUTH provider like Facebook or Gmail. Once you do that it’ll create a user name for you, as well as check your contacts to see if you have any friends on the service so you can forward news to them.

Once you’re in you’ve got a two column news feed filled with rectangles that scroll leftward. You can change direction and speed using the left and right arrow keys, and you can generate new columns by using the search function. When you hover over the news boxes you get two options: star (vote up) or slash (vote down and remove box.) When you click on the box it creates a pop-up that shows you the entire RSS item which could be the entire story. It also has a link to the page.

The user interface looks very slick but the UI could use a few tweaks. Understandable since it’s a beta. Here’s what I’d change: Add a pause button (apparently a frequent request.) Turn off the arrow navigation when the search bar has focus, and allow the enter key to work on the search field. Lastly, allow a vertically scrolling ticker for those of us who still have a 4:3 monitor.

There are still a few kinks to work out. I downrated thirty sports articles hoping to get rid of the sports category only to find that it gave me nothing but sports entries for an hour after that. It doesn’t allow you to add feeds yet since it’s in early start-up phase, but they’re willing to take e-mailed OPML files in the short term.

It seems that they have an iPhone version coming up. I hope that they’d bring out a Windows Mobile version too, I’m sick of using reddit when killing time on the go.

ReadWriteWeb just posted a blog entry that also includes more beta codes.

Most people I talk to hear “microblogging” and put too much emphasis on one of the two benefits of blogging. Blogging is a printing press in the home of everyone with an internet connection and it’s also a conversation.

It’s hard to say much of worth in such short confines when you look at it alone. A stream of posts from a single author would have been much more conveniently read in a small blog posting. To think of microblogging only as a publishing tool is to overlook other areas of value. The problem I see is while there’s a blogsphere, there’s no microblogsphere.

Conversations on Twitter happen in through tracking keywords and hashtags.  Back when Twitter had an official IM client I used it to track dozens of keywords and hashtags. Tracking meant that Twitter would notify me the instant one was used, and I’d be able to respond instantly through my phone or IM client. Unfortunately that was killed off.

There’s still a few easy ways to track keywords on Twitter. You can subscribe to an RSS feed of a search query and mash it up any way you want or you can use Tweetdeck.

The problem so far is that this all relates to Twitter. Right now I can think of at least two open source microblogging applications, laconi.ca and OpenMicroBlogger. Jaiku will eventually be open sourced. There’s already dozens of proprietary ones and the number of installations of open ones is slowly growing. The community is too fragmented.

The two open source applications support the open micro blogging specification which lets you subscribe to microbloggings of users of another site, similar to how users of a Jabber IM can usually add someone to their list who’s on another server. It’s not implemented on the most popular servers, but it fixes half of the problem with conversation.

What it doesn’t fix is being able to reply. Since addressing a reply is done in-band, that is inside the message itself, including the server the recipient is on robs you of scarce message space. If you don’t include the server then you run into name collissions. Am I addressing a message to lance_ on Twitter or identi.ca or Jaiku or what?

The proposed OpenMicroBlogger 0.2 spec might solve that by including a hidden out-of-band variable  to the message, similar to how Twitter did so with the specific message ID that you’re repling to.  Each time you do something like this you’re fragmenting the users into “users of clients who support that feature” versus “users of clients who don’t support that feature.”

What’s it going to take to have a microblogsphere? Communities will need top open up and support OpenMicroBlogging or similar, and some hard choices are need to be made to compromise between preserving message space and vital features.

The interview with Nora Young of Spark went over general issues of net neutrality, which is why my questions weren’t asked. My questions were a bit more technical and Nora felt they were already answered either in The Current or an article in The Gazette. I’m still not content despite these three pieces.

I’ll lay out my biggest problems with the interviews so far: They all include Bell claiming they need to do this to maintain network integrity, they give no explanation of where the congestion occurs or why they moved their throttling from the transit level to the DSLAM level, and they fudge the differences between wholesalers and resellers.

For example, from the Gazette article:

What they do instead is they buy a wholesale, end-to-end Internet product and put their brand around it. Then they don’t have the ability to manage their own network. It’s the same network shared between retail and wholesale. Those ISPs that bothered to invest in their own infrastructure, this problem doesn’t affect them. The use of the term “leased lines” isn’t quite accurate and I see that in a number of newspaper articles. It’s a very, very important point.

If it’s wholesale, it’s not an end-to-end Internet product. If it’s wholesale, the ISP has to provide their own transit. Bell will just provide the point to point connection that connects the user to the ISP. The ISP provides the connection to the internet. TekSavvy is not an end-to-end Bell technology. It just takes one trip to their DSL sales page to realize that.

This service is intended as a two-tiered option where you can go DSL Unlimited over Cogent (5ms to 15ms more latency) or if you prefer a premium option, DSL over Peer1 (premium routing). The difference between Unlimited and Premium Capped service is in its use of internet onramps. Call for further details!

TekSavvy, along other wholesalers like Eagle.ca, is being throttled even though they’re using an entirely different transit provider. The big question is whether or not this is covered under the tariff as reasonable network maintainence. Is the DSLAM the point of congestion? Is throttling there necessary to maintain the integrity of the network? Is Bell even entitled to throttle at the DSLAM level? No one has tried to answer this.

My inclination is “no.” All of the ISPs that built up transit capacity were able to handle the traffic without their customers complaining. The throttling has impacted all sorts of traffic, not just peer to peer, which isn’t even as big of a hog as YouTube. There are scads of problems without any transparency or even demonstration of why the throttling has to occur at the DSLAM, let alone whether Bell should be allowed.

As the CEO of TekSavvy says,

This is the exact problem and where Bell doesn’t get it. TekSavvy and all third party ISPs are paying for a “slice” of this network, so no, it’s not Bell’s at that point. They’re paid to make sure the infrastructure remains in good shape, but they’re not paid to police it! The flaw in Bell’s thought is in their not understanding that we’ve paid for the right to this space… We’ve paid for multiple Gig-E connections for the data to flow back to; we’ve paid for the DSL aggregation interface (AHSSPI) and we’re also paying on a per user basis (approx $20/month) to have the data relayed directly back to our main point of Interconnect.

So, in short, no, they don’t have rights to this network segment… An easy analogy would be a landlord, who is managing an apartment, gives himself a key to come in and out as he pleases and on top of that decide which tenants friends they let in! I’m not sure about you, but I’m fairly certain, one; the tenant would call the police, but two; you’d land up with a very big black-eye!

The idea behind Beacon is that it’ll advertise what you buy or use on the ‘net. It’s ruined Christmas surprises for a few people already, showed who plays games where, and revealed real tastes in movies. The nasty parts are that it works on an opt-out principle rather than opt-in, you can’t opt-out of the entire service, it’s so easy to overlook the “No thanks, don’t advertise my purchase” button, and the button only lasts for twenty seconds.

Once you trigger Beacon through a partner you have the option of opting out of that partner. You can’t opt-out of Beaconing for a particular partner before it’s triggered. Once it’s triggered, a screen overlay is displayed for twenty seconds. If you navigate away the overlay is gone, I’m not sure if it warns you of this, and if you don’t click “No thanks” in that twenty second window Facebook assumes it’s okay to advertise.

Though you can’t easily get out of the service you can kill Beacon for now. It looks like you don’t need to be buying under the same email address as your facebook account either.

So, if Beacon creeps you out, join the protest group after blocking Beacon, browse Facebook under a different browser than your usual browser (this is a good time to try out Prism! Make sure it runs under a different profile!), and pray that Facebook doesn’t start tracking your offline purchases.

It’s similar to PageRank in that it uses links as signifiers of importance. Unlike PageRank, it’s keenly aware of social bookmarking sites. It weighs votes in reddit, digg, del.icio.us and shows you them in next to each link in the GUI. The slides show that they use many other sites as well. Once all this data is collected, tasted, and rated each post is given a score up to 10.0. Each site has a running line graph showing postrank over time next to its name.

It’s like they’ve created a personal memetracker, allowing you to pick the pool it draws from and let you sort based upon how hot the story is.

The work in progress interface is being shown off at the screencasts section of their blog. If you’re not a big fan, that’s fine, you can subscribe to filtered and postranked feeds in your favourite reader. And yeah, I was careful to take a screenshot that showed no more than their publicly viewable screencast. I’m not under signed NDA but I thought it’d be considerate.

StartupNorth has a longer and more detailed review. I have a small wishlist, but I’ll hold off on my comments until they code freeze before opening.

I’m not talking about the splash screen ads that you have to view before proceeding to your content, as is the case with Salon and a few others. I’m talking about ads that take up a small, traditional piece of real estate like a top banner or a side skyscraper. Then, when you express interest in them by clicking or sometimes mousing over them, they expand. This is often done by positioning the ad to the top layer, making it huge to cover everything, and then putting transparency in flash.

My issue? Transparency doesn’t work in flash for linux yet. I don’t blame Adobe for this. They’ve been working with the Mozilla folks on this for a while. I do blame advertisers for their lack of consideration. For a while when I was visiting Canada.com to check my email I’d only see a big white box and a tiny ad at the top of the screen. I refreshed and was greeted with a different ad but the same problem. I was very, very annoyed. I know that people browsing the net with linux isn’t a very big segment for developers to consider, but of all users it’s only 0.5% behind Macs as of May 2007. It was on par in November 2005.

There’s still no fix for the transparency issue and I don’t know when it will be fixed. In the meantime I’m living ad-free. It’s a lot nicer this way.