There is a practise called sharding that allows the database content to be split across servers.

If you were to take a very simple example, take a customer database and look at the main table, “customers” – this is going to be an ever growing table of customers details. If the business in question is very successful, the database may grow very quickly indeed. Historicaly you would have all the details in one table. Customer names, addresses, phone numbers, etc. This isn’t the most efficient way of storing a lot of data.

We have two methods of making it more efficient. Horizontal and Vertical partitioning. You may have heard of partitioning before in relation to disk drives, and its loosely based on the same principle. Putting data in different places to keep things tidy and more efficient.

1. Horizontal Partitioning is known as sharding. This involves splitting the data and tables and putting some data in one and some in others. In our example, we could have all customers starting with the letters A-M in one table, and the N-Z customers in another. Now potentially the tables are halved in size. And then you just have to change your lookup service to distinguish which table to get the data from. You could split it even further, down to Country, or Post Code, etc.
2. Vertical Partitioning is where you take the customers table, and split the fields out into separate tables. So you might have all the static details for the customer (stuff that doesn’t change very often, like name, address, email, etc) in one table, and data that changes often in another table. The static tables get cached and therefore become quicker, and the table size shrinks as it hasn’t got all the data in it. This is also sometimes called database normalisation.

That was just a small explanation about Database Sharding for you.

http://www.mysite.com/blog/index.php?category=news&year=2012&month=1

This is a rather simple request, but it is not very nice looking or easy on the eye, and it can be improved, for the user end experience.

Take this for instance:

http://www.mysite.com/blog/news/2012/1

That’s much nicer, and easier to read and understand even for a non technical person. They may understand the category, year and month without any input from you,

How do you convert your site? Well you don’t need to, but you can use a .htaccess mod_rewrite to change the url in the background.

RewriteRule blog/([a-z0-9]+)/([0-9]{4})/([0-9]) /blog/index.php?category=$1&year=$2&month=$3 [L]

This takes the first quantifier after blog (a-z0-9) and says any characters, put into $1. The second is a 4 digit number, and the third, any number.

The [L] tells the re-write engine that if this rule matches, don’t do any more rule processing (last one)

If you implement this way of displaying your links, then your site should be better off in the big wide world of search engines. A lot of search engine spiders ignore URL’s with & in them as it can cause recursive looping and its easier to just ignore them.

Also, if you come to rewrite your blog in a new language, ASP for instance (shudder at the thought) then you can change the rule simply:

RewriteRule blog/([a-z0-9]+)/([0-9]{4})/([0-9]) /blog/index.asp?category=$1&year=$2&month=$3 [L]

Thinking about your user experience, even if you are considering the search engines, is important.

If you want to read more about the Apache re-write module, you can go here

    /**
     * @ORM\Column(type='integer', name='row_number')
     * @var int
     */
    protected $rowNumber;

Your ORM Annotations must be double quotes ” to allow the read ahead to read ahead…

VPNetMon is an application that montors your VPN connection and if it drops, the application can be configured to close certain apps.

So if you are using a BitTorrent client and using an anonymous VPN connection to download content this will make you feel much more secure indeed.

You can get it at http://vpnetmon.webs.com/

Upon returning to my desk (Ubuntu 11.10), I decided to merge the branch with the major trunk. I checked out a brand new copy of the branch, merged it (for the first time in several months) and then re-committed the pages.

Guess what, he’s still checking out the original!!!

Do yourselves a favour guys, make it Linux this Christmas…

padding: 5px 10px 5px 10px;

in a CSS file. Do you ever wonder which of the properties are which?

Just start at the top and move clockwise until you reach the left side.

Top Right Bottom Left

It saves doing this:-

padding-top: 5px;
padding-right: 10px;
padding-bottom: 5px;
padding-left: 10px;

Just a simple one that I’m sure 99.9% of people knew.

One major problem was Unity, but I got round this by installing Xfce and use that as a shell instead. Much better.

Now, my Thunderbird is proper maximised. Like full screen, so I can’t switch to any applications unless I use ALT TAB. Found help on the net.

1) Close Thunderbird

2) Locate your localstore.rdf file, mine was in:

~/.thunderbird/ftfxbq7f.default$

3) Edit the file, and find the section that looks a little like this:

<RDF:Description RDF:about="chrome://messenger/content/messenger.xul#messengerWindow"
                   width="1024"
                   height="1024"
                   screenX="344"
                   screenY="113"
                   sizemode="fullscreen" />

And change it to:

<rdf:Description RDF:about="chrome://messenger/content/messenger.xul#messengerWindow"
                   width="800"
                   height="600"
                   sizemode="normal"
                   screenX="5"
                   screenY="5" />

Then, once restarted, Thunderbird should be all nice and accessible again. Thanks to Greg for this one!

https://developer.mozilla.org/en/CSS/CSS_Reference

var myvariable = "a value";

These variables are globally accessible unless they are defined within a function. Javascript performs an operation known as “hoisting” when it compiles which draws all the declarations to the top of the page.

Here is a piece of code:

var test = "my test";
alert(test);

When run, the above will display “my test” in an alert, as you would expect.

Consider this:

var test = "my test";
alert(test);

function testMe() {
    alert(test);
    var test = "another test";
    alert(test);
}

testMe();

The above will display “my test” and then an “undefined” error and then “another test” in an alert. This is because we have declared a function variable within the function testMe function and this overrides the global state.

If we refactor the code as so:

var test = "my test";
alert(test);

function testMe() {
    alert(test);
    test = "another test";
    alert(test);
}

testMe();

Because we have not used the var keyword, the variable exists throughout, and the alerts you get will be “my test” and then “my test” and finally, “another test”

Its very important to remember this, as when you have some large functions, you could be getting obscure results you could be looking for a long time to find the issue.