Passionately opinionated » Programming

What do you do when all 4 wheels fall off the wagon at once?

Ben Fallaize — Fri, 26 Mar 2010 20:04:14 +0000

The last 2 weeks have been a nightmare.

To run a successful e-commerce website, you need a good domain name, a well built site, an easy to use payment system, and an efficient pick and pack system. Having finished 2 websites 2 weeks ago we were ready for launch. Cue website 1. Time to change the dns settings. What do you mean you don’t have control of the domain name? Who does? What do you mean you haven’t had control of it for the last 2 years…

Website 1 is beginning to look like a lost cause, cue website 2. XML files submitted to fulfillment company. Wait a week. What do you mean you still haven’t tested them? We want to launch tomorrow! Ok, we’ll launch next week then. Monday comes. We’ll test them today, promise. What do you mean the technical team have all left for the day, who’s testing my files? We’ll do it tomorrow. Yeah, sorry, there’s something wrong and we don’t know how to fix it. It’ll be fixed tomorrow.

Right, well we’re launching anyway. Payment system set to live account. Live account doesn’t work. Phone them. Sorry, it’ll be fixed tomorrow. Tomorrow comes, yes, it’s fixed, test again, no it isn’t. Phonecall number 2, fix it, yes we will, tomorrow. This pattern continues for 4 days.

Back to website number 1, domain name is looking promising, A record has changed, website still not there. ARGH, I gave them the wrong IP. Grovelling email, please change it again. Finally fixed. Test newly live site, SSL doesn’t work…

Website number 2, yes, we’ve located the problems with your files. Sorry, our instructions are erroneous. New files, tested. Yes, they look fine, let’s launch on Monday!

Tablesheets – A unique coding style, or an abomination?

Ben Fallaize — Thu, 19 Nov 2009 21:41:02 +0000

Back when I first started hand coding websites, it was all about tables. Designs were cut up into squares and laid out in rows and columns. This was ten years ago, oh how I miss those days…

While most of the world has moved on to CSS based sites, I just can’t be doing with it. Don’t get me wrong, I have tried to embrace some things in CSS, navigation for instance, and divs for e-commerce sites have made my life a huge amount easier, but I just don’t see what all the fuss is about.

Today, a colleague was having issues with a 2 column layout. To get 2 columns to stretch to fit the content, remain the same length, and have a footer that stays underneath them requires at least 4 divs, and 3-4 lines of CSS for each one. A table requires 6 lines. It does the job without any hacks, and remains the same in all browsers.

This brings me on to my second issue with CSS. Almost every CSS site I have ever seen has some degree of hack to make the site look correct in all browsers. Is this because browser makers are being lazy, or is it that CSS based design has opened the door to slapdash ‘designers’? Either way, why make life difficult? If a website needs columns, why not use the method made for it?

I use a combination of the two. I normally use a container div, with a table inside that to split it into safe, predictable columns. Then I style my TD’s and sometimes have divs within my tables. People from the tables camp and people from the CSS camp probably look at my code and cry, but it works. I don’t get issues that I have to fix with a hack. Sure, you have to think about Internet Explorer, I mean you can’t even put a body tag on a page without IE8 screwing it up, but I don’t have to deal with items being a couple of pixels out of place.

Perl’s ability to get away with murder

Ben Fallaize — Mon, 14 Sep 2009 19:19:13 +0000

Recently I’ve been using Perl to write a sort of search engine. Basically, it gets given a web address, it goes off, has a good mooch through the given web address, gathers the information I tell it to, has a little sleep, and comes back to me, all the while fooling the chosen site into thinking it’s a perfectly normal web user.

Now I am using this technology for perfectly genuine reasons, but it strikes me that it would be very easy to abuse. Also, for your inexperienced programmer, could easily get out of control. What also strikes me, is that these two scenarios almost overlap if you aren’t careful.

Say for instance, you write a basic spider to read a site, pick up any links, read them too, and record all the information in a text file. Say also, you forget to put in a checker for duplicate links. Already you have a spider that will never stop. In other words, you’ve put it into an infinite loop. ”Oops” you might say, but consider the website you just spidered. What you have essentially just done is a DoS attack. Sure, it may not have done any damage, but it has the potential.

However, chuck in a few random commands, and your basic terrible script can suddenly emulate a real web user, give it a user-agent, and voila, spider away, steal what you want, the possibilities are endless and frankly, scary…