Ben's blog

This is going to sound like blowing my own trumpet, but it's not - it's genuine scientific interest, PLUS blowing my own trumpet. So if you're a bit techy (as in "computer-literate", not "tetchy" misspelt), have you ever noticed how people ask you to do things for them? Like, "My computer doesn't work, can you please come round and spend a couple of days sorting it out for me?" Or, "I can't find some information, and apparently it might be on the Internet, can you drop everything and dig it out?"

As part of this great talent, there's the mystery of the disappearing problem. It might be something like, say, "My computer can't find the wireless network" or, "My DVD burning software won't work", but whatever it is, weeks or months of relentless dropping pf hints about how much your efforts are appreciated eventually mean you are installed in front of the computer in question, checking things out. And when you do check things out, everything works fine, no problems, nothing, just a straightforward wireless network log-on or a DVD to be burnt.

Now, even taking into account your sneaky, evil friends and relatives who do this simply because they can't be bothered to do it for themselves, that still leaves a weirdly high proportion of problems that just disappear whenever someone with a vague competence is in the vicinity. This becomes even more starkly evident when you try the "here's how you fix it for yourself" solution.

In this scenario, someone, let's call them "Sophy", says something like, "The microwave says it's in 'demo mode', and it won't cook anything." Let's also say that you bought the microwave second hand so you don't have the manual and the nearest computer where you could download the manual is a couple of rooms away. In these circumstances you would not unnaturally say something like, "Turn it off and then on again, it'll probably just reset."

The response to this is *always*, "I did that, but it didn't work." Naturally you go across to the socket, turn off the microwave, wait a couple of seconds and turn it on again. Hey presto! The microwave is reset and everything is hunky dory again.

So what has happened here? Either a) "Sophy" is lying, b) "Sophy" is having a laugh at your expense or c) your magic fingers just did it again. Being no great believer in conspiracy theories, while I believe that a proportion of these cases are "b"s, I don't think the entire world is out to get me (well, not all of the time, and when it is, I usually deserve it), and "Sophy" never lies, so it must be "c". Hey, I believe in magic.

But...

As a post script to this, can I just add that actually, wireless networks are exempt from this because they are temperamental bastards? It's a well observed phenomenon that with a network kind of near the limits of its range and power, you can never latch on to it until your techy pal is sitting in front of it, when, because he or she conducts and re-radiates wireless electro-magnetic radiation that much better then you, the connection is magically made. And it lasts until the pal has left the room. This is well known even with lovely Apple Airport Extreme wireless base stations and is actually a documented feature of all Windows-based devices.

Look, I've got nothing against Internet Explorer. Honest. Years ago, I worked on an IE4+ only web site called v3, which had some of the best gadgets and content available on the web at the time (around 1999). There was a DHTML space invaders game, which was the only way to access a series of short lie-based factual articles called Fact pie, and a fantastic splatter game featuring the Stoney Blokes, along with loads of other silly and brilliant stuff. All of the interactive stuff was written especially for the IE4 API and none of it worked in anything else, but as IE4/IE5 had more than 85% of the market share at the time, that was no hardship.

Meanwhile Netscape had released Navigator 4, which was their first DHTML-ready browser. Trouble was, to get it out to market earlier than IE4, they based its API on an older version of the web standards draft documents, so a lot of what you could do with it was based on stuff which never made it into the final set of standards, making it a bit of a developmental dead end. Oh, and it crashed all the time too. No, back around the turn of the millennium, while I was willing Netscape to get it right, I was a big fan of the Internet Explorer experience.

Since then, Netscape and then Mozilla have done their best to overcome this slow start, beginning by throwing away the Navigator 4 API and indicating to everyone they'd done this, by omitting version 5 altogether and jumping straight to version 6. This was based much more solidly on W3C web standards for CSS, HTML and JavaScript, transforming DHTML coding from a buggy, trial-and-error nightmare, into a much more reliable and rewarding process. Plus, the new browsers didn't fall over all the time.

Internet Explorer, on the other hand, was far too high and mighty to do anything as mundane as comply with web standards (okay, for the sake of accuracy, I have to point out they did comply with some web standards, but not nearly as many as Mozilla, or Opera) and continued on their merry way, adding fancy new incompatible features to the browser rather than making them render web pages in the same way as everyone else.

My favourite examples of this include IE's dimensions model which means that widths of padded boxes or columns on the page come out differently than they should when coded in standards-compliant mode, or the margin bug which meant that a floating column doubled the width of the margin on one side, or the layering system which meant you couldn't just tell an item on the page to sit above everything else, you had to pull it out of the code and put it in a special place. The list goes on and on.

Okay, so Internet Explorer did introduce the fantastic innerHTML property, which allows you to stick new stuff inside containers on the page, without having to constantly traverse nodes and elements on the page, and it has got slightly closer to being standards-compliant with each new release. IE7 is marginally less awful than IE6, and IE8 is another improvement. But with each new release, they seem more intent on adding fancy new functionality, hardly any of which anyone ever uses, instead of doing the obvious thing. MAKE IT WORK LIKE EVERYONE ELSE'S BROWSERS.

If Microsoft could do that, so you didn't have to spend hours recoding all kinds of bits and pieces of a site so they work in all the different flavours of IE, the web would be a much happier place. Web sites would be built more quickly. They'd work better and do more. And Microsoft are not going to lose market share if they do that. In fact, more non-MS developers would probably end up using it for development because it's *there* on every PC, and would probably recommend their clients do the same, so they might even gain market share.

Then they'd be able to realise their dream of crushing every other technology company and organisation into the dust, while actually doing the world a favour. But instead, they concentrate on trying to come up with some new bit of functionality that is so amazing and so unique to them that everyone else is left looking stupid and the W3C web standards organisation shuts up shop, hands the keys over the Steve Ballmer and slinks off to the pub. It's never going to happen, and in the mean time we're all left spending extra hours and days of our time making decent, standards-compliant web sites that look lovely in Firefox, Safari, Opera and Chrome work in IE6, IE7 and IE8. Thanks a bunch.

You'll have seen our slide shows I expect. There's one on the homepage with all those nice wide images fading into each other. You can have them on your web sites too, by the way. You just have to ask.

Anyway, they work beautifully, text or images or both, can fade in and out in a dance of endless loveliness. Or so we thought. Then we created a slide show for Unique Earthenware which had nice big images at the bottom, with some text to introduce them above. It looked great in Firefox and Safari and Chrome. But not in IE 6, 7 or 8.

Instead of looking great, it looked as though someone had taken the text and replaced it with their stomach contents. "Blocky text" is not sufficient to describe the horror. Back to Firefox... aaaaah. Sneak a peek in IE... aaaaaarrrrghhhh. Initial thoughts were that we had to replace the text + images with images containing the text, but that seemed a cop out.

Aha! We work on the web, and sometimes it contains pearls of wisdom. A quick search revealed that the solution to the block text is either to use a background colour in the text containers that are fading in and out, or to use MS fonts, like MS Sans Serif. We specifically wanted to use Tahoma in these headlines, so we gave the tags a lovely background of white and Bob's your uncle. No more ugly block where our lovely text should be.

In the interests of accuracy, my fictional lawyers tell me I should point out that I haven't actually tried the MS font solution, but I expect it'll work. So, having found that out AND finally got a strangely configured IBM laptop onto the office wireless network, today is complete and I go to bed happy.

It's funny, but every so often I just feel like I can't go on. You muddle along with the same old products you've been using for a while because they by and large work, and you base your new products on the old ones because, by and large, the old ones work.

And then you notice one more snag and suddenly you just can't put up with muddling through any more, adjusting code here and fixing things there. Then it's time for FIX and UPGRADE.

I've been working on a rating product to allow visitors to clients sites to rate items in a number of categories and leave comments. There's more to it than that, but you get the idea. Anyway, as I've been working away at it, bits of code I was reusing from elsewhere started getting on my nerves, and then a new site of ours started having minor administrative problems with permissions. I tried to ignore it, but in the end I had to do something.

And when I looked, I found that all the problems were coming down to the same sorts of things, so there was no putting it off any more. I had to upgrade the older products before I could safely work on my rater.

And now, four days later, it's finally all done. All the code has been upoloaded to the test server and tested, and redone and retested and redone and retested and... so on until, hey, it doesn't break anything any more.

Then it was uploaded to the development server and tested with the development versions of our existing sites. And redone and retested and on and on... THEN it was time to put the whole lot on to the live server and push the button. Plus a whole load more testing of course.

Why would you care? I don't know really, but I just had to get it off my chest.

Damn. I've just noticed I've done a whole blog entry with a single keyword, and what's the point of a blog entry if it doesn't have a single keyword to up the SEO and attract you here from Google or Yahoo, or MSN or Ask Jeeves. See? That's better already. Now, if I can just work Zope and Plone and content management and ecommerce in, I'll have done a good thing. Job done I reckon.

After three weeks, Google has indexed all the new pages of our lovely new site - and we're still on the first page for web design and hosting of google.co.uk. Hurray for us. Traffic is creeping up as more people come to the site, and our stickiness is increasing too, so they're spending more time browsing.

With the Ashes over, and England victorious over the feeble Aussies, thanks to a couple of amazing collapsible batting performances, you'd think my obsession with learning to bowl off-breaks would have diminished, but it hasn't, especially now that I've discovered that my action is apparently unique and therefore I could actually have discovered a new, ultra-effective devilish type of spin. The fact that it turns pretty inconsistently is neither here nor there. I suppose I ought to try it against a batsman at some point.

Our latest site, with a new listings product and fancy JavaScript calculator is nearing completion. It'll be finished within the week, ready for testing prior to launch on September the 1st. I particularly like the listings view/click statistics that we've built in. Very nice.

Checking in with Google Webmaster Tools this morning, I see that Google has started to index our new site, with 9 of the 87 URLs listed in the XML sitemap parsed. This ought to mean that over the next few days, the rest of the site slowly filters in. Then we'll be able to see how the new site stacks up in terms of Googe ranking.

So having got the new Beetlebrow site live and looking great, and made a few tweaks to make it run faster and work better, I thought I'd better nip along to our Google webmaster account to upload the new XML sitemap and make sure the transition went as smoothly as possible.

The sitemap is built with a product we developed ourselves from a simple Limi template and a UML diagram. It's not complicated, but it doesn't need to be - it just lists everything on the site of the types of pages we want Google to index, while allowing us to specify pages to exclude. So far so good but...

There's always a but. This time it was that, for some reason Google didn't like the sitemap - it was missing a tag. How very strange. As this was the first time we'd used the new product, and I was also including it on a couple of recent sites we'd built - www.siefersharrison.com and http://www.richardmax.co.uk - I didn't know what kind of problems we might face with it, but I had to sort it out.

Looking at the sitemap in Firefox, it looked fine, and the W3C validator said it was well-formed XML, but then in Safari all I got was an empty set urlset tags. What's going on? Well it's obvious innit? Schoolboy error. I was logged in to the site as a manager in Firefox, but I wasn't logged in in Safari, and Google, being an anonymous visitor was getting the empty urlset tag experience too. And unless you specifically give the scripts which access all the configuration settings for the sitemap the same rights as a site manager, if you're not logged in, they can't read the settings. Dur.

So, having given myself a stern telling off, I gave the scripts the necessary permissions, and hey presto, the sitemap work in Safari, and Google stopped complaining. So I submitted the three sitemaps, for Beetlebrow, Richard Max and Siefers Harrison, and waited...

Twenty four hours later and all of Siefers Harrison and all of Richard Max are indexed, but Beetlebrow hasn't been touched. Another 24 hours after that, still zero files indexed for Beetlebrow. Is Google punishing me for my stupidity? If so, they're even cleverer than I thought. I know it can take days, weeks even, for Google to get round to indexing you, but to do our two client sites so fast and leave us hanging on? It's just not fair.

On the other hand, I did find out that the old Beetlebrow site was on page 1 for "web design and hosting" (see http://www.google.com/search?q=web+design+and+hosting), which is pretty darned cool, cos it's not like there's much competition between web design firms for good Google rankings or anything. On the other hand, that'll probably disappear when it does get round to indexing us. Curse it and crush it.

At the same time Richard Max is NUMBER 1 in Google for "planning lawyer" - not blimmin' bad for such a generic term - see http://www.google.com/search?q=planning%2Blawyer. So while I'm disappointed Google hasn't seen all of our lovely new Beetlebrow content yet, I'm exceedingly delighted with our SEO efforts on Richard Max. Hurray for Google!

And here it is - our shiny new site. We've been so busy on work that pays, that we've neglected our own site. The only problem with that is - we're a web design company and in theory our site should be one of the most fantastic in the known universe, a superb endorsement of our great skill and talent.

When we first built our previous site in 2002, the web was still a "cool" place, and web sites were "funky" and "groovy", and our site reflected that, although with better jokes. But the time had come to get serious, and what you see is the result. Built on the back of our fantastic Plone CMS and a load of programming effort, we've got all kinds of bells and whistles built-in, most of which you can't see, but which makes it easy to keep the site up to date and full of exciting new content.

Some of the stuff is visible, like the zoomable images like this one in the portfolio, or the slideshow on the homepage. Okay, so it's not earth-shattering new technology, but it is pretty groovy (see what I did there?), and how are you supposed to know that we can give you these things, if we don't have them on our site?

And there's more - like this blog - entirely our own work, and easy to install on your web site, if you want one. But I think my favourite bit of stuff is this mechanism for linking to a particular part of a page, which not only takes you to the appropriate part of the page, it also highlights the relevant part and builds in a little "back" button to return you to the page you just came from. Check it out. Have a look here at the FAQ about ecommerce, and then, on the same page, follow this link to the bit about email. If you scroll up and down you'll see that only the bit you were linking to is highlighted and has the back button. Simple, but nice.

Until today, blogging was "kind of" built in to our main web site building block plug-in (what Plone developers call a product). That meant we could set up a blog and it would look okay, but control over the leaving of comments was a bit tricky, and it just wasn't a clean solution.

So yesterday I extracted all the blogging goodness from our main web site product, created a new blogging product, added one to the other and voila! Blogging to go.

That's why the first few entries of all out blogs have the same date. We had to take our first few entries, which were built with the old product, and manually copy them into the new set-up. Okay, I know I could have written a migration script to copy them, and their dates, from one to another, but for about eight entries in total, life's too short.

And besides, I've got better things to do with my time, now that I've taught myself off-spin. All I have to do it to be able to land it outside off stump on a length with some kind of regularity...