More Than You Ever Wanted To Know About...

From MSDN via Jeff Atwood's Coding Horror.

The Turkish-I Problem

These new recommendations and APIs exist to alleviate misguided assumptions about the behavior of default string APIs. The canonical example of bugs emerging where non-linguistic string data is interpreted linguistically is the "Turkish-I" problem.

For nearly all Latin alphabets, including U.S. English, the character i (\u0069) is the lowercase version of the character I (\u0049). This casing rule quickly becomes the default for someone programming in such a culture. However, in Turkish ("tr-TR"), there exists a capital "i with a dot," character (\u0130), which is the capital version of i. Similarly, in Turkish, there is a lowercase "i without a dot," or (\u0131), which capitalizes to I. This behavior occurs in the Azeri culture ("az") as well.

o.O

Back at the beginning of the month, I shared with you all a post by one of the blogosphere's "big names", regarding the development of a programmer. Well, now he brings us another gem on programming languages...it revolves around Perl, but there are some excellent general principles discussed.

At this junction in human history, the best thing you can do as a programmer (other than perhaps write your own language) is not to get too attached to any particular one, and choose them as appropriate for the task at hand, since each language has its niche.

Warning to Radar: It's another long one, you might want to span it around your nap-time.

Radar asked me this morning for some questions that he could ask a person to gauge their understanding of PHP. Here's a short list I came up with off the top of my head. It occurs to me that perhaps MG (and maybe Sideshow? I dunno how things go around there) could stand to brush up on these topics. Also, I'd be interested in a similar list for C# (Johnny?) Continue reading "PHP: A quiz"

Almost 5 years ago, after reading a paper on Latent Semantic Analysis (one of the key methods Google uses for helping a machine understand the "meaning" of a document), I looked up and said to someone (Johnny Elbows?) "Hmm - I wonder why they don't use a Hidden Markov Model for that?"

Well, a student doing his graduate internship at Google has done just that. You can find the original paper here, or download a C++ implementation from Google Code here

[Ed: I've fixed the fact that I originally forgot to fill out the detail link under "XKCD" - there's an interesting blog post behind it all on xkcd's site.]
A lengthy discussion about signal-to-noise ratios in IRC channels, started by our hero xkcd, led me to discover Dunbar's Number, which has a fascinating level of detail for being a Wikipedia entry. Which reminds me, has anyone heard anything new about Google's Knols project?

Picking up the "MTYEVWTKA" series, we're going to look a little bit into being an (Pronounced "Em Tie Vwetka") we're going to start with a short article on how to most effectively read non-fiction books.

Johnny Elbows wrote a post a while back about a Javascript Color Fader he had created. In it, he had the following block of code:


var red = getRgbComponent(color,0);
var green = getRgbComponent(color, 1);
var blue = getRgbComponent(color, 2);



Applying "currying" to this block, we can give ourselves functions that will get the color components for us:


var getRed     = function (color) { return getRgbComponent(color,0); };
var getGreen   = function (color) { return getRgbComponent(color,1); };
var getBlue    = function (color) { return getRgbComponent(color,2); };


Then:
var red = getRed(color);

Not the most natural fit for currying, but should be an effective example...

[Ed: Link fixed, thanks Johnny...] Humble Beginnings:
I wanted to use an interesting web service I found, but found myself unable to sign up for it because their email regex failed to respect '+' in the local-part of my email address. Figuring I'd give them a hand, I tossed out a quick email:

Subject: signup email validation fails

  '+' is a valid username component in email addresses, one which I use to pre-filter incoming messages. Additionally, when I went to report it as a bug on your bugs forum, I was unable to find any way to create new posts...maybe I'm just not seeing it?

Then the fun begins... Continue reading "Email Addresses"

This paper talks about methods for letting the user tell a search engine what they consider important (called a "user-defined dynamic weight" in IR terms).

There's some really exciting stuff going on over in the land of Python 3000. Guido brings us the news.

I'm especially fond of the new "ABC's". I think they're going to have a big impact on the world of frameworks.

As we evolve our concepts of data and meta-data, the issue of how to query temporal data ("when stuff happened") has lagged behind the more traditional relational data. To this point, we have had to rely on vendor extensions to SQL such as PgSQL's "interval" type. A 1995 paper proposes "a clean way of defining temporal databases based on logic...views temporal databases as multi-sorted, finitely representable first-order structures. Query languages then became formulas in suitable logics over the vocabulary of such structures."

Starting on page 94 they go through a particular temporal transform called the "Magic Set Transformation". It gives a little bit of discussion on how temporal data can affect decision systems.

I spent some time walking through this code yesterday and found his point very enlightening.

[Edited: Johnny E. pointed out that s9y was choking on the wikipedia link to Erlang. It's been fixed. Thanks, Johnny! ]

Aiming to replace/obselete/prevent technologies such as JMS, MQ, and TIBCO, AMQP. Think of it like SMTP, but for programs to exchange data rather than for people to exchange emails.

One of the leading implementations appears to be Apache's qpid (say it out loud) project. Another interesting one is RabbitMQ, an independent implementation in Erlang, a language that's getting a lot of attention at the higher levels of the software industry (sort of like Ruby did about 5 years ago, or Python 10 years ago...or Lisp, every year since 1958).

What could be better than Google or Bittorrent? How about a little of each?

There's actually a more detailed paper than this one (which is from 2003), but you'd have to be an ACM member to read it.