If you need some suggestions right now, cut to the chase and jump down to the bottom of this page for my quick UK Web host recommendations.

While simple static sites can be hosted free, there is a vast array of options available for hosting scripted or data-driven Web applications such as CMS, eCommerce, blogs, etc. Cost varies from £1.50/month right up to £350/month and beyond.

Dedicated, VPS, shared?

The four main categories of Web hosting (with typical monthly prices in £UK) are:

1. Shared hosting, free-£30
   A physical server is divided up and shared amongst hundreds of customers. Because partitioning is within the operating system, there is potential for some sites to hog system resources, breach security boundaries or even crash the server. In short, you’re relying on a host’s ring-fencing technology and the responsible use by other customers on the same server.
2. Virtual Private Server (VPS), £15-£80
   By partitioning a physical server, each VPS runs its own operating system and behaves like a dedicated server with all of the security and some of the performance benefits without the cost of a dedicated server, or the potential problems of shared hosting. Resources are shared evenly between each VPS and this makes for consistent performance.
3. Co-located server, £30-£300 (1U rack space)
   Your own server located at an ISP data centre. Only necessary if you have very unique or specific hardware or software requirements. Do what you want with it, but you’re responsible.
4. Dedicated server, £50-£350
   Your server, only your own sites are hosted, so nobody else can hog resources or crash your server. More flexibility in terms of customising setup and installing and upgrading software.

Many Web sites will appear to perform almost identically whether they are sitting on a cheap shared host or a top-end dedicated server. So why is there such a vast difference in price?

Web hosting is now a highly competitive commodity service. The main factors determining price are:

1. resilience and reliability – uptime; commonly 99.99%+ for HTTP, but database servers can have problems and may not be measured in uptime stats.
2. performance – general web site response time and data throughput.
3. capacity – ability to cope with high volumes of traffic or spikes.
4. ‘over selling’ – the number of other sites hosted on each server and controls placed on customer accounts. Beware of ‘unlimited bandwidth’ – although bandwidth is rarely unlimited, this often translates to ‘busy server’.
5. features – backup, control panels, SSL certificates, auto installers, storage and data transfer allocation.
6. support – what are you responsible for yourself and how quickly can you get help when you need it?

So, start by determining your present and future needs – the number and scale of Web sites to be hosted, mission-critical status, support required, etc.

If you already use shared hosting, try ipneighbour to find out how many other sites are hosted on your server. The results can be quite interesting, and alarming.

Platform

The platform of choice for hosting Web sites is still predominantly Linux/Unix for ubiquitous LAMP applications, CMS and blogs or JAVA, Rails, or Python apps.

Most shared hosting runs on a solid Linux platform and some hosts use grid or clustering technology which is supposed to offer performance, resilience and scalability benefits. In practice, these can under perform compared to a decent VPS or dedicated server, but for many people it’s a good enough alternative.

If you particularly need to run .ASP scripts or .NET applications, you will have to go down the proprietary Microsoft route. Just make sure the server is fully patched and set up correctly – these servers are constantly under attack and blighted by security problems, despite recent improvements.

PHP 4 / 5

As one of the most widely used scripting languages for Web applications, good PHP support is essential. PHP5 was released way back in 2004 and PHP4 has now reached end of life. Unfortunately, many Web hosts still only support PHP4 and some offer only outdated releases of PHP5.

There are many significant benefits – performance, security, functionality – to using recent stable versions of PHP5 and many of today’s increasingly demanding Web applications require it. PHP5 also has much better support for Web services, XML, JSON etc.

So, choose a good Web host with up to date PHP5 support (5.2+ ideally). Hosting companies tend not to upgrade PHP on their servers, so start with a recent version or you could find yourself lumbered with a server which won’t run a lot of modern Web applications.

Reliability

So what about uptime? It actually varies little between the average hosts and the best ones. Decide if your Web sites really are mission critical. If they’re not, the rare possibility of a short outage might be worth what could be a considerable cost saving.

Side-note: if you think your hosting needs to be mission critical, then your Web site usability definitely should be! Bad usability can lose you far more visitors/customers than an unreliable Web host.

Shared hosting – getting better

Recent years have seen increasing popularity of VPS as a more cost-effective alternative to dedicated servers.

Shared hosting, traditionally unreliable, has also seen the emergence of premium level shared services which have made significant improvements in performance and reliability. This blurs the boundaries between shared and VPS hosting, just as VPS has with dedicated.

Multi-domain and reseller hosting

Most hosting packages on offer these days allow customers to further sub-let their own web space; several domains (in the 10’s or 100’s) can to be hosted on a single shared or VPS account. Sometimes this is pitched as a ‘reseller’ option, although most server appliance control panels, such as Plesk, support reseller activity and some have direct billing facilities.

You can also set up client areas with their own control panel access, quotas and allocations. You can even brand your control panel to make it look like you provide and run hosting services yourself.

Domain names and DNS management

Another important element of Web hosting and Internet services is domain name management. If you look after multiple domains, it is important to use a good domain host with a decent full-feature control panel and a sensible policy on moving and re-tagging domains.

Most Web hosts offer DNS management, but it’s probably worth managing this elsewhere, with a well established and stable company. Moving domains can be more difficult and disruptive than moving web sites. Updating DNS records to point to a new host or server is relatively straightforward compared to re-tagging your domains to a new IPS.

Finding a good hosting company

Assuming you have identified your hosting requirements, you then have the difficult task of finding a good company to fulfill those needs.

Ask around for some recommendations, but treat these with caution. The experience one lucky, or unlucky, customer has with a hosting company may not be an accurate reflection of the real situation.

Search the Internet for hosting reviews. Most hosting companies get slated on forums and review sites, so be wary of single reviews from disgruntled individuals or suspiciously glowing reviews. Look for a consensus and trust your instinct.

Web Hosting Talk is one of the main hosting forums,

Server location

If your Web sites target users in a particular country, it is best to host on a server which is located in that country. The main benefits of serving your sites from the same country as your users are:

1. SEO: search engine rankings can be affected by which country your web site’s IP number is registered to. Because search results are commonly tuned to the country each user is located in, hosting on a server in a different country can reduce search result rankings, often significantly. Hosting companies know this and often conceal the location of their data centres.
2. resilience: reduced points of failure if network traffic has fewer hops to make.
3. performance: slightly better if your local host is a good one.

My Web host recommendations

I have the dubious honour of looking after a variety of servers and hosting accounts for my UK clients as well as myself. This gives me plenty of hands-on experience and the unique ability to compare and contrast.

I will be posting a more detailed review of UK-oriented Web hosting companies in due course. But for now, here’s a taster based on my experiences so far…

• Rackspace: one of the best – great performance and resilience. Premium service, premium prices.
• United Hosting – UK and US data centres, premium shared and dedicated hosting. Very focused on server/network performance and reliability, with truly outstanding customer support.
• 1and1: good value packages, reliable but mediocre performance despite the massive network. 1and1’s atrocious support is legendary. GeoIP warning: data centres in Germany.
• Servage: feature-packed ‘value’ shared hosting, big cluster setup promises much but delivers shocking unreliability. GeoIP warning: data centres might be lurking in Germany, Sweden or Denmark.
• 34SP: good range of budget hosting options, but cheaper shared hosting has had some reliability problems. Reasonable support. Still not shaken off the ‘hobbyist’ tag. Now offers VPS. Certainly worth considering.

Other promising Web hosts; I can’t claim any first hand experience with these, but research indicates these companies are very well regarded:

• InnoHosting – UK and US data centres, well-priced shared, VPS and dedicated hosting. One to watch?
• MediaLayer – US data centres, LAMP optimised shared or dedicated hosting.
• Media Temple – US data centres, grid/cluster shared, VPS and dedicated hosting. Does it live up to the hype?
• 5QuidHost – great value US/UK LAMP hosting. Looks rather ‘back-bedroom’, but enjoys an increasingly good reputation.
• Slicehost – US data centres, high performance bare bones Linux VPS for hardcore techies.

Look out for my detailed Web hosting reviews, coming soon. All will revealed, in my no holds barred review. In the mean time, if you have any good or bad experiences with a Web hosting company, please let me know.

This can have a negative impact on project time-scales, costs and user experience. It can also compromise the design when problems which could have been avoided need to be corrected further down the line.

The term ‘design’ is often misused and misunderstood; in this context, I’m talking about the agency-style creative and graphic design stage which is all too often annexed from the peculiarities of the medium and the complexities of the development process.

Here are some ways to avoid the top six problems we encounter with non-Web designs:

1. Design for screen resolution
   Printed output has the benefit of significantly higher resolution than the screens Web sites are viewed on. As a result, designs created at higher resolution, and proofed/approved on printed hardcopy usually do not translate to a 72dpi screen. Text usually suffers first, becoming unreadably small once the layout has been scaled down to fit within a 1024x page width, or worse if you are constrained to 800x. Other design elements suffer too and the whole balance of the layout can get seriously compromised once the design has been shoe-horned into a medium for which it was not originally designed.

   Recommendation: for non-Flash page designs, create mockups or visuals in pixels at 72dpi. Don’t use a vector-based design application; Adobe Illustrator and QuarkXPress are the wrong tools for this job. Use something like FireWorks or PhotoShop, work with a browser viewport template at 100% size and use browser fonts.

2. Account for variable content
   Most successful Web sites are content rich, growing rapidly and organically with contributions from numerous sources. Problems arise when a design assumes that content doesn’t vary in length and is always displayed in a fixed size text box. This never happens. Even in cases where content won’t change, different browsers on different platforms will render text differently no matter how well your CSS reset tries to level the cross-browser playing field.

   Recommendation: expect content to change, particularly in length. Make sure your design gives it room to grow and contract. Don’t expect text to fit snugly within a fixed-height box – it won’t.

3. Don’t treat navigation as a visual design element
   Navigation design is an incredibly important part of the whole Web design process. Information architecture, content classification and user interaction must all be taken into account when designing navigation.

   Visual styling obviously needs to compliment the rest of the site design but it is essential to get the navigation working first.

   Recommendation: fully map out the site’s content first and get a handle on the big picture. From the user’s point of view, navigation needs to be understood and interacted with, not just looked at!

4. Providing a style guide
   It is not practical to expect visuals for every page on a Web site. Because there can be so many different layout variations, particularly with dynamic pages, developers need a style guide to help them get the aesthetics right where no specific visual exists.

   Recommendation: supplement design visuals with a style guide covering design elements and specific details which can be applied in situations where no page visual exists. Also describe how user interactions should look and work – for example, rollovers and ’selected’ states; developers will happily improvise, but designers will always have something in mind, so they should share it!

5. Work with – rather than just provide assets for – the development team
   Projects where concept and design are determined, then handed over to the dev team to build, will have problems unless the designers have a thorough understanding of content, technical and user interaction issues, such as navigation.
   Recommendation: It’s a team effort – people working on concept, design and build all need to communicate, work together and respect what each other is contributing.
6. Don’t sell clients non-workable visuals
   This is often where the problems start. All of the mistakes described above are rolled into a set of print-friendly visuals which clients buy into. When the time comes to start making the design work on-screen and on-line, changes have to be made to get a workable result. “Can you make it look like the visual we approved?”

   Recommendation: if you have to sell a visual concept to a client, make it very clear that it is not an accurate design and is likely to change. If you need to get accurate Web visuals approved, take these important issues into account and present something workable. There is no need to see costs escalate or to fall out with a client over the difference between what is expected and what was delivered.

The T5 shambles is a high-profile and very damaging embarrassment. It’s not the only problem of course: Britain’s Heathrow airport has been a major time-bomb for quite some time. In fact, you might question why such a large and expensive facility has been built around an airport which is so chronically short of runway capacity.

The big story this time of course is the fact that despite considerable testing and vast expense, on launch day, Terminal 5 failed spectacularly due to a number of unexpected circumstances which all combined to break the whole facility, and in some style.

This is also what can, and routinely does, happen to complex Web sites on launch day. So what lessons can we learn from this?

1. Testing is crucial, but it won’t guarantee smooth running from day one.
2. Usability of all systems, computerised or not, is extremely important. Poor signage at T5 caused confusion and staff, who had received some training, struggled with systems when this proved to be inadequate.
3. Fix known problems. Issues raised during testing may not have been acted upon. This could be down to poor communication and lax Management. Listen to staff, document all known problems, prioritise issues and deal with them.
4. Don’t overlook peripheral issues. Many of the T5 problems were compounded by difficulties with side issues such as staff car parking and security bottlenecks. The knock-on effect of these may have been underestimated.
5. Conduct full-cycle and high load testing. Unit testing is not enough – how everything hangs together is just as important. At T5, there were problems when three different teams of baggage handlers failed to synchronise their activities.
6. Anticipate potential problems and build-in capacity to deal with them. When the baggage handlers at T5 struggled to cope with the number of items passing through the system, everything backed up and ground to a halt. Ask some ‘what if?’ questions, and work out practical answers.
7. The consequences of not testing can cost far more than the testing which could have helped avoid the problems in the first place.
8. When problems develop, communicate with the people who are affected. If you can’t provide any detail, at least reassure people that problems are being dealt with.
   
9. Launch with reduced load, ramp it up, monitor the results and make improvements from what you learn as things develop. On launch day, T5 experienced large passenger numbers and a high level of load. Problems escalated quickly to a level beyond which they could not be dealt with. The facility broke and a many people were affected.

This last point is possibly the biggest lesson to learn for Web projects at launch.

Live testing is inevitable

Like it or not, real users will be testing your Web site when it goes live, regardless of how prepared you think you are. Accept this, deal with it and make it a part of your launch process.

Stage traffic buildup and never mailshot large numbers of people in one hit inviting them all to visit a newly launched Web site. Stage your traffic generating activities, monitor the affects on load, errors, user behaviour, etc., then identify areas for improvement and fix the big ones before ramping up your traffic.

With the benefit of hindsight it is easy to criticise those responsible for the Heathrow Terminal 5 shambles. Of course, T5 is a large and complex facility which was only a few different decisions away from working and being hailed as an incredible achievement.

We can learn from this and do more to avoid complex Web projects from suffering a major “T5″ at launch time.

New Web sites, particularly newly registered Internet domain names, can take many months, sometimes 6 to 12 months to become fully visible on search engines. Get something out there early, even if you are not ready for a full-on Web site.

Some businesses are in the fortunate position of not needing a Web site to help them generate business. Indeed, there is something very cool about being so good, you just don’t need a Web site.

Other businesses may have plans for a new Web site, but need to delay for budgeting reasons, or have a large project in the pipeline which could take many months to complete.

In all these cases, the new Web site could take a long time to show up in search results, particularly on Google which sandboxes new Web sites. Google considers established Web sites to be more credible and trustworthy, which directly affects how highly that site will rank in search results.

Here’s a good tip: even if you don’t need a Web site, or have plans to launch one in the future, get at least one search friendly Web page out there as soon as possible. This helps to get you established with the search engines and will pay dividends when the real Web site comes along, helping you to generate traffic sooner and achieve results quicker.

Make sure your holding page is search friendly and contains some content relevant to the business you are in. Don’t just sling up a ‘coming soon’ or ‘under construction’ page. Create something useful and relevant even if it doesn’t reveal the true identity or purpose of the site which will replace it.

Usability is becoming an increasingly important element of Web design and development, yet why is it such a tough sell?

1. Consequences are invisible – users might be shaking their heads in despair and bailing out of a site within seconds of arriving, yet most usability problems go unnoticed and unreported; if you can’t see a problem, there’s nothing to fix.
2. Clients want click-click-wow! – Web sites typically get judged on very superficial criteria. A great concept and slick design are enough to get a thumbs up and the invoice paid. Like any business, agencies must satisfy their clients.
3. Fun factor for the project team – creativity, innovation and pushing boundaries are undoubtedly more fun than dry boring user-centred design practices.
4. Usability perceived to stifle design – if a project starts with and is driven by a creative process, there will be resistance to user-centred design practices interfering with aesthetics. There is no reason why this should be such a problem. Teams need to work in harmony.
5. Overstretched budgets – what gets dropped when the budget starts to run out? Usability, and documentation are often the first to go. Core activities gobble up budgets which are already eaten in to by client/project acquisition costs.
6. No measurement of success - remarkably, many Web sites are not evaluated for how successful they are. With most other forms of marketing or promotional activity, measuring results is essential. Web sites seem to escape this scrutiny.

What is missing here? Users.

Clients quite correctly assume that when their appointed design/Web agency creates a Web site, the job will be done well, and this includes whatever needs to be done to deliver a successful project.

Taking responsibility

So, if important work like usability (which includes accessibility) and basic SEO get neglected, is the client or the agency responsible?

Few clients will insist upon specific technicalities, never mind conduct due diligence checks. It is therefore up to agencies to ensure that project scope includes work like usability and that budget allocation covers it.

Herein lies the problem: Web sites are more time (and therefore, cost) intensive than most people realise and it is the less visible work which typically gives way when tight budgets are stretched, particularly by expansive agency overheads.

The solution, therefore, is not only to raise awareness amongst both clients and agencies, but also for budgets to be more accommodating. For this to happen, clients need to appreciate the value, importance and scale of their online initiatives, and agencies need to control costs and manage budgets more carefully.

It is not unusual for large proportions of a Web site budget to get consumed by project acquisition costs and creative work, even before the developers have fired up their code editors. This leaves little room for the likes of usability, copywriting, content development, QA testing, SEO, security audits, online marketing…

It is quite understandable that agencies are driven not only by their own creative values, but also by a desire – and business need – to satisfy their clients. Until users, customers and true results start to count more, usability will struggle to attain the level of importance it needs and deserves.

There is no doubt the pace of development in mobile Internet on phones and handheld devices is moving ahead fast. Recent announcements from Apple and Google only serve to highlight this trend.

It’s all happening, and with Apple, it’s not only the excessively hyped iPhone, but the new iPod touch – the phoneless iPhone with its WiFi Web browsing capability – which will help to push mobile Internet applications forward.

Archos also offers some WiFi enabled handheld media players with Web browsing capability, and improvements in mobile phone software and connection speeds are starting to make Web browsing only a thumb tip away wherever you are.

I predict that convenient, usable devices like iPod touch, will bring handheld Internet access to the masses. Even in the home, where many people already have WiFi broadband connections, accessing the Web anywhere in the house on a handheld without needing to fire up a personal computer – even a notebook – will be a big winner.

I have been using mobile Internet services on my phone for a while now and almost take it for granted. The Gmail mobile applet is great, Google Maps for mobile is super useful and Opera Mini makes non-mobile Web sites usable on a small handset. We are also seeing more Web sites starting to serve pared-down mobile pages for speedy access to existing content on the move.

Handset developments are seeing bigger better screens, a choice of connectivity options, some of which are very fast, and built-in GPS which will might have a good effect on mapping and location-based services. Even though it is possible to determine the location of a mobile handset without GPS, this is rarely used due to privacy and accuracy issues.

The mobile networks are doing their bit as well: affordable flat rate data tariffs are becoming more common and fenced-in network access is a thing of a the past.

Developers and agencies need to get on board with this sooner rather than later. Get into the habit of making your Web sites work on mobile devices, at least ensuring that navigation is usable and essential content is available.

Standards based coding will go a long way to achieving this, but it may still be necessary to employ transparent content negotiation techniques. Browser testing will become a bigger issue and mobile browsers, particularly variations of Opera and Safari, will need to be tested.

It makes sense for a Web site to reach the most number of users as possible, and with many more people experiencing mobile Web access, this is another source of eyeballs which should not be ignored.

Search Engine Optimisation can seem like a bit of a black art, but take care of the basics, and you can get 90% of the job done without consulting the rune stones.

Here is my top ten of essential SEO:

1. Domains: register your domain early or use an established domain; search engines can sandbox (effectively hold back) web sites with young or recently registered domains. It also helps to use a domain which contains a few prime keywords that relate to your area of activity.
2. Page titles: summarise the page, include three or four prime keywords which also feature in the page content. Also make page titles human-readable as you want users to see and understand the title on the search engine results page, and be interested enough to click through. Page titles are extremely important, so get them right.
3. Keyword and phrase research: put yourself in the mind of a potential user. Think about the words and phrases they are likely to use when you want them to find your site. Use these keywords and phrases in page content and metadata…
4. Metadata: write appropriate description and keywords tags which are tuned into the content on each page. Generic keyword and description tags are of limited value. Using relevant keywords in page URLs also makes a difference.
5. Web standards page coding: there are many reasons why web pages should be constructed in modern standards compliant coding, not least of which is that it helps search engine visibility no end. Good coding also helps overall usability and will make pages work better on a variety of different browsers and Internet connected devices.
6. Page content: not only should you give people a reason to visit your web site, but having good quality relevant text content is also important for search engines and PPC ads, which look for and analyse text on each landing page.
7. Keywords and phrases in URLs: formulate URLs with relevant text containing primary keywords. Avoid lengthy URLs loaded with query string parameters and ID numbers.
8. Sitemaps protocol: this is a file or dynamic XML feed which summarises every page on your site; search engines read and process these files to help them index your site more completely and often, more quickly, it’s like a short-cut around the search engine’s link discovery crawlers. SE sitemaps are particularly important for large web sites with a lot of content.
9. Inbound links: the more quality, highly ranked sites which link to your site, the higher your page rank will be. Consider creating good content specifically to bait inbound links, but write for people, not for search engines. Avoid driving the site from one container page – frames and single object Flash are particularly damaging. If you prevent deep-linking, this kills your traffic generating potential. Avoid link farms and don’t rely on reciprocal linking.
10. Analytics: don’t just look at the basic top-line figures, use your web analytics to mine deeper for information about how your site is being found, what works and what doesn’t. You can also experiment with content and design variations, track the results and continue a cycle of fine tuning to improve results over time. Creative use of analytics can also be used to examine user behaviour and test UI variations. It’s a powerful tool, use it.
11. Don’t cheat: your site can get penalised, or blacklisted altogether, if you attempt to beat the system. Don’t use link farms, autosubmit services, text cloaking or anything like that. Google is particularly hot on this and has a webspam team dedicated to the task of weeding out the cheats. Generally speaking, honesty pays in SEO.

This only scratches the surface of what you need to do in order to achieve good natural search rankings and generate traffic for your web site. Of course, just ticking these items off the list isn’t enough – to be truly effective, the work has to be done right.

Even getting 90% of the way there with your search marketing efforts will take some experience and expertise. Breaking into that last 10% will take a considerable amount of effort, which in many cases could be hard to justify.

Get the basics right and your web site will be found. Just make sure there is something compelling and usable when people arrive. Think beyond visitor arrival.

Finally, paid search. This is an enormous revolution in advertising, just look at the rapid growth and huge revenues of Google. This is also one area where thorough setup and incremental fine-tuning is essential, particularly when there is a lot of money at stake.

If you don’t take the necessary steps to make your Web site visible, you are pretty much wasting your time creating it in the first place.

Copywriters need to be Web savvy; studies have shown that users read Web pages in a particular way: they scan text for highlighted keywords, paragraphs are skipped over, attention spans are notoriously short and they won’t hesitate to click and leave if they get bogged down be text or can’t find the information they are looking for.

Here are some basic tips for Web writing:

• Consider SEO requirements when composing copy
• Use the journalistic style inverted pyramid; open with a conclusion, then drill down to more detail
• Be super concise and keep the word count down
• Highlight key words and phrases
• Use one subject or discussion point per paragraph
• Include meaningful sub-headings to break up the text
• Make good use of bullet-point lists to distill information out of verbose paragraphs, into summary lists
• Write meaningful link text that makes sense when user scan the page
• Use hyperlinks to cross-reference pages of related information on the same or other sites
• Tone down the marketing hype; people are particularly suspicious of this on Web sites

Writing for the Web now has to include important interactive and search marketing influenced details such as: link text, page titles, META tags and pay per click ad copy. Some specialised Web writing skills have therefore emerged:

SEO copywriting involves tuning content, page titles and META tags to match, while achieving optimum keyword densities and making sure everything – particularly page titles – are human readable as well as SEO friendly. Page titles (the HTML <title> tag) are particularly important due to the weight they carry in search rankings and the influence they have on user click-through rates.

Pay per Click ad copy writing is even more of a challenge – both keyword tuning and human reading factors need to be absolutely right. You also have to plan copy variations for split testing and ensure keyword relevance so ad copy and landing page match user expectations.

Good copywriters are worth their weight in gold and appropriate Web copywriting skills are no less important.

It is also interesting to note that – for better or worse – the reading habits of emerging generations of Web users may be affected to the point that print-based writing will start to adopt the characteristics of Web writing and this could, in time, affect how print copy needs to be written.

This might not be a good thing. If it does happen, blame the Web.

There are of course numerous frameworks for PHP, right from monolithic bloaters like CakePHP, through to presentation layer frameworks like Smarty, which is not a full application framework, but does a lot more that a basic templating engine.

Frameworks help to give structure to your PHP applications, which speeds up development and makes code more consistent and maintainable. You also benefit from separation of application logic from presentation and data handling. Mixing HTML/XHTML with PHP is undoubtedly bad for anything other than the odd dynamic page or two.

The are many other benefits, including: database abstraction, a full-range of function libraries, improved security, presentation layer templating, clean search engine friendly URLs, and so on.

What makes CodeIgniter my framework of choice is the small footprint – lean and efficient, CodeIgniter minimises the extra layers of code that inevitably get introduced with a framework, so performance is good. It is also extensible, flexible and does not impose strict controls on how you organise your code.

Finally, CodeIgniter has really good documentation and as the friendly CI community often attest, it brings greater levels of enjoyment to the process of Web application development, and this can only be a further boost to your productivity and the quality of your work.

I often get asked this question, so here’s an answer:

Although .com officially means global ‘commercial’, many UK users perceive .com to mean US/American or more specifically, NOT domestic/UK.

This becomes an issue when users are scanning through lists of Web addresses or search results – .coms can be dismissed immediately if users are looking for a home-based business. In this respect, .co.uk has a more comfortable and trustworthy feel to it.

So, for a genuinely global business, .com is the boss, although localised domains should still be used.

For predominently UK-oriented businesses, .co.uk is the most appropriate choice and can actually enhance your credibility on-line.