If you have been doing LAMP (or anything MySQL) related for a while, you already know that the simplest way of getting a random set of rows from a table by using ORDER BY RAND() is a really Bad Idea^TM. The guts of SQL require that this solution generate a random number for each of the posts in your result set and then sort by this. Random numbers are generally expensive, and this requires one be made for each row, which as your result set grows will become quite the bottleneck.

I have seen other alternatives to this:

• Pull all of the IDs from the table and then select X number of random ones in PHP and run another query to pull those random ones – http://www.webtrenches.com/post.cfm/avoid-rand-in-mysql
• Use a slightly more complex set of SQL variables – http://www.electrictoolbox.com/msyql-alternative-order-by-rand/

I am testing out another approach that I’m calling “modulus” sorting. The ID approach is decent, but as tables get large, you end up doing a large data transfer to get all of those IDs. My solution may not be truly random, but it appears to be doing nicely so far.

First query:
$query = “SELECT COUNT(*) AS c FROM `table`”;

Second query:
$query = “SELECT * FROM `table` ORDER BY `id` % ” . mt_rand(0, $count / 2);

Essentially, we pull the number of records in the table and assume that the IDs are fairly evenly distributed and then sort them by the modulus of their ID and a random value. The modulus operator is quite a bit faster than RAND() and the results I have been getting so far are random enough for my purposes. I’ll drop in updates as we see how it goes!

During a big buy-out announcement recently within my company, one of the speakers mentioned the book Drive: The Surprising Truth About What Motivates Us. Now that I spend a lot of time as an NYC straphanger, I get a lot reading in during my daily commute and decided that coming from influential and successful people, it was a safe bet that it was a good read.

Let me just say up front: wow. It is not a smooth read, and at times the writing style is a bit repetitive, but the content of the book is amazing, and the majority of it is references to psychological studies around the world. If you have any desire to understand what motivates you or how you as a manager could motivate your employees, this is a must read. I’ve discovered things about myself that I did not effectively realize before, and beyond that discovered that I am not alone in them.

The core concept of the book is that human beings are far more complicated than can be modeled with a simple reward-driven behavior (“carrot and stick“). We have an innate desire to solve problems as well. When it comes to salary, as long as we are receiving fair pay or are not having money problems, more money is a very poor motivator. In fact, if you dangle the carrot of a (especially monetary) reward as a reason to do something, you can severely damage long term motivation and the desire to excel on that project. Essentially you can turn a fun project into drudgery by associating money with it.

They actually brought up Open Source web development as an example of how great things can come out of volunteer work. I take a bit of issue with that since most successful open source projects have some of the core pieces in place and are only supported because big money-spending companies back them and pay people within their own companies to make contributions. However, the idea is still reasonably sound.

Bringing this back to my previous post on the “good vs bad” sides of achieving “flow“, the book noted that it becomes much more difficult to reach flow when you are working on a project for a monetary reward. This helps to explain why when you are writing code for fun and solving problems on your own time, you can end up glancing bleary-eyed out of the window and realize it’s 5AM. If you are doing the same project for work because somebody is pushing a deadline on you, you can be checking the clock every five minutes until it’s 6PM and you can leave without feeling too guilty.

In independent analysis of the results of work in both situations, the people building for fun very consistently create higher quality code.

I say score ++ for flow.

So I work in social media for a large advertising-centric company, and one of the things that I do is test integration with Facebook Pages. I run a simple test Facebook Page to try out different things on, and I recently started having issues with including any goo.gl URL in a post to the page:

If I have a simple link to XKCD, shortened with Goo.GL: http://goo.gl/b4UzX and shortened with Bitly: http://bit.ly/1pB1Uk, I get the following results when I try to post them onto the wall of the page:

Original XKCD link: posts and embeds just fine
Bitly link: posts and embeds just fine
Goo.gl link: always asks to solve a CAPTCHA

I’ve tried it with a bunch of other sites and some other URL shorteners and have had no issues. Is anyone else having this problem? Any other Google-related services that are being tampered with? It’s not blocking it, it just feels like harassment!

If you’ve done any mind intensive task, even taking a very challenging but doable exam, you probably understand the concept of “flow”. Flow (aka getting “in the zone”) happens when you are hyper-focused on a particular problem or task, and you are able to provide contiguous output on the problem, pretty much non-stop.

Personally, coffee helps me get in the flow a lot. If I have some audio distractions around me, then Trance music can help too (preferably very repetitive and with no words). I know some people who compare the feeling of this “flow” to what happens when you’re on Adderall. The rest of the world and the seconds that it ticks with just disappear, and as a programmer, your fingers never stop typing.

It’s an amazing feeling, and I have within the past month read two different authoritative reviews on how it really affects your productivity. According to Robert C Martin in The Clean Coder, flow is patently bad. He compares it to being hypnotized (I’m not arguing there), but he says code produced in the flow is code that is unaware of the bigger picture, and not flexible. A lot of code produced in this way has to be revisited and re-architected at a later date because it will be single-minded and not jive with the rest of the program.

Just today, I read one of the old Joel on Software posts by Joel Spolsky entitled “Where do These People Get Their (Unoriginal) Ideas?”  where he says that achieving flow is one of the basic requirements of programming in general as focusing on the problem very closely helps you solve it quickly and more thoroughly.

What are your experiences with this?

Personally, I fall more on Joel’s side, but with some reservations. You need to make sure you have your environment set up correctly and you know proper programming standards. My most solid code has been written in the flow while following TDD. I was able to easily refactor it for later additions and it was solid, usable and well documented (I write all of my PHP code with PHPDoc / JavaDoc style comments). Putting all of these together and still focusing hard on the problem gives me excellent progress while still generating maintainable code.