Internet WorkShop

The D3 javascript library looks awesome – clean, extensible, and powerful.

Checkout this example of mashing US Census boundaries with unemployment stats…

http://mbostock.github.com/d3/ex/choropleth.html

All I wanted was an updated copy of berkeley db…

Oracle has it, its free, but you have to register. Ok, I’ll register. The form has lots of required fields…two of which are broken select drop-downs with no options! This is embarassing for multi-million dollar company with a supposed commitment to open source, no?

Here is the culprit:

https://myprofile.oracle.com/EndUser/faces/profile/createUser.jspx

See the drop-downs for Job Title and State/Province.

Oh, and I tried the live chat. The rep told me to call the help line :
“No I can do nothing, please call one of those numbers to log and start a ticket sir”

Argh.

tags are ok but they are missing a namespace; maybe there is implicit one from the blogger but for global discovery what about

tech:lein
finance:lien
gvelez:chickens (means my chickens, not chickens in general)

quick to write without requiring much forethought or looking up namespaces, but more precise and less prone to overlap than flat tags

Looking for local? (ie, an API to get local data specific to a subject). Tried http://outside.in, the API looked promising and their site had good data, but after a week of 403 errors using their example code snippets and no response from the forums or support, was about to give up for awhile. Then I noticed the Weather Underground folks using fwix. They have an extremely simple API, decent quality data (at least if you filter for News or Places), and although its not as complete as I’d like, it does have more than just twitter posts. For ease of use they get 5 stars. Will see how they do on updating, relevency and completeness…

Spell checking and typeahead are two of my top gripes with modern software. URL or bookmark completion is ok, that is when I’m in the ‘trying to remember’ mode. But when I’m in the flow of writing, having the computer guess what I’m trying to say is incredibly distracting and annoying.

Originally, I thought gmail was running auto-spellcheck for me, but it was the browser, in this case Google Chrome.

In Chrome, you turn off spellcheck under chrome://settings/language ; uncheck the ‘Enable spell checking’ box underneath the list of languages.

(You can also get to this screen advanced settings screen by clicking on the wrench in the upper-right, select ‘Preferences’ and ‘Under the Hood’, then click ‘Languages and Spell-checker Settings’ )

Update 7/14/11 – I believe this is fixed in rturk 2.4 – thanks Mark!
http://rubygems.org/gems/rturk
***
rturk, the Ruby gem for making calls to the Amazon Mechanical Turk API, uses a REST transport layer. That’s fine, but all calls are currently performed by a GET, which has a length limitation. When making calls that include long strings of data – such as the XML for a QuestionForm structure in a qualification tests – errors may occur with the non-explanatory message ‘400 Request Error’.

Was able to patch it by making a change to lib/rturk/requester.rb :

46,47c46,50
< RTurk.logger.debug "Sending request:\n\t #{credentials.host}?#{querystring}"
< RestClient.get("#{credentials.host}?#{querystring}")
---
> # RTurk.logger.debug “Sending request:\n\t #{credentials.host}?#{querystring}”
> # RestClient.get(”#{credentials.host}?#{querystring}”)
>
> RTurk.logger.debug “Posting request to #{credentials.host}:\n\t #{params.inspect}”
> RestClient.post(credentials.host.to_s, post_params)

A more robust fix might be to use POST only for longer requests, or make it an explicit option on the RTurk object

I’m not sure where the bug is, but when saving some binary data that was generated in a Rails3 before_create callback, it kept getting truncated in the actual INSERT INTO statement (though it appeared fine even in after_save callback, it was truncated in the database). Using PostgreSQL 8.4.7 with pg (0.9.0) and activerecord 3.0.5

Seemed like it could be related to this fixed bug, but my problem is on the save itself:
https://rails.lighthouseapp.com/projects/8994/tickets/611-cannot-write-certain-binary-data-to-postgresql-bytea-columns-in-2-1-0

In any case, found a simple workaround: uuencode the data first, and uudecode on loading.

   before_create do
      ... stuff that builds my_hash ...
      self.my_hash = Base64.encode64(Marshal.dump(my_hash))
   end

Then later, to reconstitute the hash,

   loaded_hash = Marshal.load(Base64.decode64(@record.my_hash))

Trying to shoehorn some data from the back into the wp_connections database, I needed a perl equiv to the md5(uniqid(rand(), true) function. From some general suggestions at http://www.experts-exchange.com/Programming/Languages/Scripting/Perl/Q_26773756.html came up with the following snippet:

use Digest::MD5 qw (md5_hex);
use Data::Uniqid qw (suniqid uniqid luniqid );

sub gen_token {
# approx equiv to md5(uniqid(rand(), true));
return md5_hex(uniqid);
}

Will also need to use Scott Hurring’s http://hurring.com/scott/code/perl/serialize/ to get it fully into the needed format…

Perhaps this is somehow due to my config…but if it happens to me, it could happen to you:

I have a model, say table_one with a belongs_to: table_two. It happens that table_one and table_two each have a column named ‘datafile’. The two columns are intended to refer to entirely separate datafiles, and thru lack of imagination I named both the same. Now when I edit a row from table_one, it shows the associated row from table_two, and uses an identical name for the input fields for both ‘datafile’’s! record[datafile]

The result is that the datafile for table_one is wiped out on edit, as it is replaced by the blank datafile input for the associated record for table_two. Ouch!

Workaround of changing the name of the column in table_two fixes the problem.

Using activescaffold 3.0.5 for Rails 3.0

Seems to me the key is creating a language that lets you efficiently express what you need to. Well-designed dsl’s (domain specific languages) have a lot of power. But these guys are the ones that apparently know:

http://www.readwriteweb.com/hack/2011/01/secrets-of-backtypes-data-engineers.php