Sociable released a WordPress plugin to link up WordPress blogs to Facebook with Facbeook Connect. I just installed it, took about 5 minutes, and it works great. It allows you to link your FB account to this blog, so you can post comments with your FB identity. Your comments are automatically approved if you posted with fbconnect. You also get to use the familiar fb:multi-friend-selector to invite friends to this blog.
Not bad!
I’ve been using jRails in my recent Rails projects, the original Rails in-place editing plugin uses script.aculo.us, there is a jRails version of it, but neither of them is RESTful – they both create extra actions to update the in-place edit fields.
I found janv’s rest_in_place plugin and it uses the default update action to update the field, so no routes modifications are necessary. I had some problems with the plugin at first but after a pull-request correspondence the plugin now works well. Here are the highlights on how to use it, keep in mind that I use HAML.
The plugin’s init.rb doesn’t load anything for you, so you have to go into your application layout and include the js file:
= javascript_include_tag 'jquery.rest_in_place.js'
If you have CSRF protection on, this plugin also requires you to set a javascript var. If you have jRails it automatically append the token in ajax requests, but then you would have to modify the plugin a bit to get it to work.
:javascript
rails_authenticity_token = '#{form_authenticity_token}'
In your controller’s show action, handle the javascript response:
def show
respond_to do |format|
format.html # show.html.erb
format.js { render :json => @model }
end
end
then you can render the helper in the views:
- div_for @model do
%p
= label_tag "Name"
%br
%span.rest_in_place{ :attribute =>'name' }
=h @model.name
%p
= label_tag "Location"
%br
%span.rest_in_place{ :attribute => 'location' }
=h @model.location
37signals moved TaDaList to run on pretty much the same things I’m using: EC2, EBS, Elastic IPs, Apache, Passenger. They also started from the same Ubuntu Intrepid images. 🙂
Joshua posted more info on their setup in the comments section of the blog post:
Joshua Sierles 28 Nov 08
Matt,
Our custom image is based on the Ubuntu Intrepid images from Alestic. We install useful EC2 gems and base packages, bundle, then provision each instance by role. Working with EC2 is so easy, we don?t see much value in using a third-party provider for our scale.
Yaroslav,
EBS and Elastic IPs were the primary motivation for moving to EC2 . We use EBS extensively: for MySQL data, logs and local repository mirrors. Performance so far is excellent. Snapshotting volumes is a breeze and makes setting up MySQL slaves and staging environments really easy.
[From Ta-da List on Rails 2.2, Passenger And EC2 – (37signals)]
Treasure Island is a nice spot to shoot the SF skyline, I went there yesterday with the SF Photography Meetup group. That was my first event with them, got there half an hour late and just setup and started shooting, didn’t talk to other members. I think one of the organizers walked by to see if anyone needed help, but I guess I didn’t look like I needed any.
I also shot a few shot of Sherry with 580EX2 tripped with Pocket Wizards, but none of them turned out well enough to post-process. Anyway, some shots:
I tried the included HDR workflow in CS4, but the resulting images didn’t look that great, maybe I just don’t know how to use it. I did auto bracketing on most of the shots so I was hoping I could generate some dope HDR shots. Most of these shots were merged manually with layers and masks with more than 1 shots, but I could’ve easily done the same with a single raw file.
Found this photoset by ttstam from the Flickr 5D Mark II – group. The noise control at high ISOs on the 5D Mark II looks impressive from these images. You really have to go to Flickr and look at the full resolution versions to see the noise performance. He also included reference shots with the 5D and 40D. (not the exact same shots though and they’ve been post-processed and HDR’ed)
Here is the shot at ISO 3200:
ISO 6400, slightly noisy:
ISO 12800:
ISO 25600, it starts to get noisy even in the downsampled version:
I’ve had a dedicated server at ServerBeach for about 5 years, overall they were pretty solid, one time the hard drive died but they were able to mount the dead drive in read-only mode so I could recover all my data. After they were acquired by Peer 1, there were some connectivity issues, but after a few tickets it was pretty solid again.
My ServerBeach server was so old that it was one of the EOL (end-of-life) servers, which means they no longer stocked spare parts for them, which is kindda not good. Ray and I have been hosting our Ruby on Rails projects on EC2 for quite a while, OnMyList is also hosted on EC2 with software load-balancing, multiple app and database instances. I had always planned to consolidate things and move my ServerBeach setup to AWS/EC2, but I never got around to it. About a month ago ServerBeach sent me an email saying that they had to reassign the IP addresses on the server, that is a bit of a pain coz I would have to change the DNS on all the domains, so that pretty much gave me enough incentive to move.
Ray and I started with the Intrepid Ubuntu AMI by Eric Hammond. To get a static IP, I allocated an Elastic IP Address, you get one Elastic IP for free per instance, and you can assign it to any instance. We also use Elastic Block Store (EBS) for persistence storage, the entire /home, the MySQL database, and most of the important configuration files in /etc are symlinked to the EBS volume. With EBS you can easily create snapshots for backups by using the ec-create-snapshot command included in the EC2 tools.
To get an Elastic IP, just do:
% ec2-allocate-address ADDRESS 75.101.xxx.xxx
To associate that address to an instance:
% ec2-associate-address -i i-xxxxxxx 75.101.xxx.xxx ADDRESS 75.101.xxx.xxx i-xxxxxxx
To create an EBS volume (make sure the zone is the same as that of the instance):
% ec2-create-volume --size 50 -z us-east-1c VOLUME vol-xxxxxx 50 us-east-1c creating 2008-10-31T18:01:29+0000
To attach the volume to an instance:
% ec2-attach-volume vol-xxxxxxx -i i-xxxxxxx -d /dev/sdh ATTACHMENT vol-xxxxxxx i-xxxxxxx /dev/sdh attaching 2008-10-31T18:01:46+0000
Now, ssh into the instance and run mkfs, I’m using ext3 but you can do whatever you want:
% yes | mkfs -t ext3 /dev/sdh mke2fs 1.41.3 (12-Oct-2008) /dev/sdh is entire device, not just one partition! Proceed anyway? (y,n) Filesystem label= OS type: Linux Block size=4096 (log=2) Fragment size=4096 (log=2) 3276800 inodes, 13107200 blocks 655360 blocks (5.00%) reserved for the super user First data block=0 Maximum filesystem blocks=0 400 block groups 32768 blocks per group, 32768 fragments per group 8192 inodes per group Superblock backups stored on blocks: 32768, 98304, 163840, 229376, 294912, 819200, 884736, 1605632, 2654208, 4096000, 7962624, 11239424 Writing inode tables: done Creating journal (32768 blocks): done Writing superblocks and filesystem accounting information: done This filesystem will be automatically checked every 29 mounts or 180 days, whichever comes first. Use tune2fs -c or -i to override.
After this you can add a line to /etc/fstab to mount the EBS volume to a mountpoint.
Everything on the server was pretty much installed from APT, we went with BIND9 for DNS. Rails apps are hosted with Passenger as usual. God is used for monitoring. We host the primary DNS for all our domains on the instance, and secondary DNS’ are hosted at EditDNS. I chose EditDNS because they support SRV records, and we need that for Jabber. All our domains use Google Apps, there is no reason to run my own mail server when Google provides superb reliability, spam filtering, secure IMAP and SMTP services, the gmail web interface is also way better than any open source webmail package I’ve seen. Pretty much all the hosted domains share the same zone file, which makes DNS management super easy. If you got BIND9 setup but the server ain’t answering remote queries, check to see that the port (53) is enabled, with EC2 you have to manually open up ports, most ec2 setup instructions open up port 22 for SSH and 80 for HTTP, if you need anything else you need to authorize them yourself:
ec2-authorize default -P udp -p 53
For Rails apps we setup a deployer user to make things really simple, the Capistrano recipe is pretty straight forward. We use remote_cache with Git to make deploys very very fast.
To migrate user data over from my old server, I first created the users on the new system manually, I didn’t feel comfortable copying over the /etc/passwd, shadow files, and what not, also I had to get rid of quite a few old accounts. After the users and home directories have been created, I used rsync via SSH to copy everything over, I ran rsync with quite a few exclusions, like I wanted to exclude all the Maildir’s and the spamassassin bayesian database files coz I didn’t need them anymore, and a lot of them were pretty huge. If you run rsync in archive mode (rsync -a) the permissions and ownerships are sync’ed automatically to the users and groups with the same names. I also had to migrate Apache1 virtual host configs to Apache2 ones, but that was pretty straight forward. SSL setup was also pretty trivial in Apache2, I moved the hostname over and was able to reuse my InstantSSL certificate.
To migrate MySQL databases, we just did mysqldump’s and loaded them on the new server. I realize this probably wouldn’t work if you’re migrating a high traffic site, but fortunately we could tolerate a few hours of downtime.
After everything was running smoothly, some of our domains had longer TTL values and people were still going to the old IP. So I setup Pound on the old server to forward all HTTP and HTTPS traffic over to the new IP, worked like a charm. I went with Pound because that was the only one available in APT on the old Debian sarge server. Next time I do a migration like this, I will make sure to lower the TTL of the domains a couple of days pre-migration, but even then I am assuming DNS’ obey the specified TTL’s, and most probably don’t. I use OpenDNS and it was pretty easy to expire their cache manually.
It’s been a few weeks and the instance has been rock solid, amazingly it is much higher performance than the ServerBeach dedicated server, but since the server was over 5 years old, that is not saying much. Cost-wise, the EC2 instance is about $20 cheaper than a dedicated host (I was paying about $90 a month for the server + 2 additional IPs). Ray and I have been using EC2 heavily for over a year, so to scale horizontally with more instances is pretty trivial, and it’s far quicker than getting another dedicated server setup with a hosting provider.
I spent most morning messing with color profiles of my MBP. The “late 2008” MBPs come with 2 different panel models so far: 9C84 and 9C85. I have the 9C84, probably because I got mine on the release day. I’m not sure which model is better, but it’s more important to have it calibrated correctly than which model number it is.
I really should get a hardware calibration device to calibrate my screens and printer, well, I already have an ICC profile for printing with my printer (Canon i9900) and my photo papers (Ilford Gallerie Pearl), if you use Ilford papers you can download profiles here. I spent more than an hour looking at comparisons and reviews of the different hardware calibration devices, they are pretty confusing as it’s not just the hardware that matters, the software makes a huge difference too, obviously. Doesn’t seem like I can get anything I will be completely satisfied with without spending over $1k. If I were to get something now I’ll probably go with the Eye-One Display 2 by X-Rite. If you’re interested in that here’s a link to get it from Amazon.
I googled to see if I can find profiles calibrated with different devices and software, and I found this thread on MacRumors forums that is exactly what I needed. I tried pretty much all the uploaded profiles there and these 2 looked the best to me. I am using the D65 one, but most people will probably like the native whitepoint one better. (I’ve shared these 2 profiles here and here).
btw, for my photography, I shoot in Adobe RGB, RAW, and open them at 16-bit in Adobe Camera Raw (CS4). I never convert them to sRGB before I publish my work online. I do embed the Adobe RGB profile in the JPEGs, so if you view them in applications that support color management they should look fine. If you don’t, the colors will look really messed up. Safari in Leopard uses ColorSync, so if you use Safari you’ll be fine. If you use Firefox 3 in Leopard, you need to enable color management yourself, as it is disabled by default. To do that, you can either edit the settings in about:config, or install this add-on. If you’re one of the unfortunate few who are still on Windows, I believe Vista has color management built in, if Vista sucks too much and you’re still on XP (good choice!), you can try to download Microsoft Color Control Panel.
Now my profiles from display to prints are pretty close, so I’m okay as long as I shoot and sell my prints. But I upload a lot of images to Flickr, I think maybe I would start converting the profiles to sRGB so they won’t look ridiculous on browsers that don’t support embedded color profiles or color management at all. Any color management or digital workflow tips you’d like to share? Please post them in the comments.
If you’re new to color management, RenĂ© Damkot wrote a great post about the topic at Canon Digital Photography Forums.
The San Jose Mercury News reports that a 47 year old man named Jing Hua Wu, after being laid off from his job at a fabless digital radio semiconductor startup in Santa Clara called SiPort, came back to the office on Friday afternoon with a gun and killed three people. Sid Agrawal, the company?s CEO, and Brian Pugh, VP Operations, were killed along with one other as yet unidentified woman. Wu is still at large according to the latest reports.
[From A Sad Day In Silicon Valley]