Create a Great Personal Home Site

About this Document

This document explains why and how to create a great personal homesite. The text can be found below.

Document Information

Written By:

Shlomi Fish

Finish Date:

28-November-2005

Last Updated:

28-March-2006

Licence

This work is licensed under a Creative Commons Attribution 2.5 License.

The Article Itself

Introduction

We’ve all seen it - the bad home site: flashy or non-standard colours, annoying animations, large text all over the place; too little content, obscure pages that no one knows how to get to from the reachable ones; no navigation menus, no site map, and no breadcrumbs trail; horrible markup and too little content to show for. If you want to have a good laugh, then DivisionTwo magazine has an insanely funny parody of the bad homepage (warning: contains a lot of explicit content).

However, occasionally in one’s infinite browsing, one can find some good personal home sites. They have a consistent style, valid markup, a common look and feel, are attractive to the eye, load quickly and have a lot of good content for hours of experiencing. The purpose of this essay is to explain what the elements that make a good home site are, for everybody there who have a home site or want to have one. It will also touch on some of the technicalities of creating and setting up one.

Why have a Home Site?

So why should you bother with having a personal home site? There are several advantages to have a great one. Here are some:

1. More people will know about you.
2. You’ll receive a lot of feedback from other people, and often engage in interesting conversations with them.
3. You can bring in a lot of publicity to yourself, and your business.
4. You can make money from advertisements and donations.
5. You’ll be able to practise and improve your writing skills, artistic skills, etc.
6. You can link to your favourite pages and resources on the Web.

Elements of a Good Site

The elements that make a good home site are divided into two: content and presentation. Let’s tackle them one at a time.

Great Content

Filling your home site with a lot of content takes a lot of time and requires quite a lot of work. It is also a process that requires one to invest more time working on the site into the future. It is, however, very fun.

Here is a list of items you can put on your web-site in its early incarnations if you don’t have any good ideas of your own:

1. Information about yourself: resumé, biography, etc.
2. Hyperlinks to pages and resources you like or are related to.
3. Photographs you took.
4. List of recommended books, movies, artists, etc. Note that you can benefit from such items using the Amazon.com associate program and similar associate programs.
5. Fan sites for Music artists, T.V. shows, Movies, etc.
6. Archives of favourite quotations and collections of jokes.
7. A weblog (= online diary), or at least a link to it.
8. A Paypal donate link and Google ads (which bring income from the site).

Once you’ve set up a home site you’ll probably find that you have more and more ideas on what to put there. But you first need to overcome the first obstacle and actually have a web-site.

Good Presentation

Now for the second aspect of a great site: a good presentation. This involves a very large number of different elements. The list below aims to be as encompassing as possible, but it is possible some points are missing or some will need to be added in the future.

Security

A website needs to be secure. The most possibly secure web site is a static HTML one where the HTML is served directly from the server, without the server processing it. This is adequate for the needs of most web-sites, especially personal home sites.

Alternatively, if you want to have a server-side generated web-site you need to make sure it doesn’t have SQL injection attacks, Cross-Site Scripting (or HTML injection or XSS) attacks, privilege escalation, comment or wiki spam (see the rel="nofollow" attribute for instance), or any other security problem. This requires much more conscious effort and discipline than a plain HTML site.

A server-side Content Management System (CMS) of some sort allows one to edit the pages using a web browser, and also allows web page visitors to add comments or even extra content to the pages. (With all the other implications of such interactivity). One possible way to have such a CMS is CMS hosting. Essentially, a CMS hosting provider manages many instances of the same CMS on its servers, allows one to register his own instances, and maintains and upgrades them all at once. That way, you can have your own CMS without much of the maintenance headaches.

Of course, if the CMS host neglects to install an update in time, then you still have a security problem. But it is still probably a better idea than deploying your own CMS on a normal hosting account.

Navigation Aids

In my previous article “Building Navigation Menus” I gave a list of common patterns in web-site navigation. They appear under the heading “Common Patterns in Navigation Menus and Site Flow” in the article. Here is a list of them without too much explanation:

1. A tree of items
2. Next/Previous/Up links to traverse a site
3. Breadcrumb trails
4. hidden pages and skipped pages
5. A site map

A good site should have a navigation menu, and possibly the other navigation aids mentioned here.

One thing that was omitted from the article (due to being too new to be included) is a Google Sitemap. This provides a list of interesting pages in the site for Google and other search engines, and is also recommended. One can see it in action in the Google search for KDE, where the top result displays more pages in the site.

Visual Appeal

A good site has good visual appeal. It is pleasing to the eye, interesting to look at, and isn’t intrusive. A central element to such a design is a common look and feel of the site. Generally all pages should contain a navigation bar or two, some common icons, a footer, etc. If the general display of the page changes from page to page, the site’s visitor will feel confused. If the pages are plain HTML pages, with no dedicated blocks, then the site will be harder to navigate and also look unprofessional.

Valid and Semantic Markup

There are definitive standards for HTML and related technologies as set by the World-Wide Web Consortium. By following the standards, one can make sure that present and future browsers will know how to handle the page correctly. See the designing for compatibility section in my “Web Publishing Using LAMP” presentation for more information.

A good site master makes sure the markup of all or most of his pages validates. It leaves a better impression, is much easier to maintain this way, and would make sure your web page is displayed correctly in the future (barring some surfacing web browser bugs). Note that the valid markup does not affect how the page is presented and you can easily style it in very attractive ways, while still adhering to web standards.

Refer to the Web Standards Project for more information.

Aside from being markup, your markup should be semantically correct. So, for example if you want to make a text larger, you should use CSS instead of designating it as a heading (<h1>, <h2>, etc.).

Colour Choice

Aside from looking attractive, the choice of colours in a site needs to be non-intrusive enough, and correspond to how people read the page. For example, flashy colours are too hard to see through, and should not be used as the backgrounds of titles. The body of the page itself should be distinguished from the main page layout somehow.

I admit that I am not an expert in this area, and some of my designs are bad in this regard. Someone once pointed me to several problems with one of my designs, which he and I thought were relatively OK. I must say, what he said simply made a lot of sense, and was perfectly obvious, yet I thought the design was OK beforehand.

A different issue with colour choice is making sure one’s page is accessible to people with the various types of colour-blindness. The problem is that are three such types (Red-impaired, Green-impaired, and Blue-impaired) and making a good colour choice for that is hard. Most colour blind people can be expected to customise the page layout using custom CSS stylesheets, so it may not be much of an issue. And if all the important content is in text, it will always be accessible enough.

Responsiveness

The pages of a good web site load quickly and the web site as a whole feels responsive. Putting an excessive amount of images (especially large and poorly compressed ones) on a web site is guaranteed to cause it to load more slowly. Other culprits are a bloated markup, an excessive amount of JavaScript, and a slow connectivity of the host.

Good web-sites heavily optimise for limiting the bandwidth, because it affects broadband users, and people with a slow connectivity much more so.

Various Annoyances

There are many common annoyances in web-sites. Some of the most common ones are:

1. A horizontal scroll bar, that prevents the entire width of the site from being visible. A webmaster should realise that it’s not enough to make sure there is none when the browser is maximised in a large screen. By all means, many people use much lower widths, and the web designer should accommodate for them.
2. Frames are evil, and should be avoided. If you want a common look and feel, you should generate your pages from templates, either off-line or on the fly.
3. Most JavaScript is incredibly annoying and often useless. It also many times makes the site much less portable across different browsers and accessible. It is often added either because people want to demonstrate their JavaScript skills, or because broken site-generation programs insert it there.
4. Animation on web pages is incredibly annoying and distracting. Some people claim that a web page should only have at most one animated image. I think that usually even one animated image is too much. (If you want to display a movie demonstrating some action, then an animation on an isolated page is naturally an option.)
5. Pop-ups are annoying and should be avoided. Plus, most browsers give mechanisms to prevent them from appearing anyway.
6. There shouldn’t be any excessive dependency on Macromedia Flash, and no Flash animations on the front page. Flash is not open-source, often causes accessibility problems, can be very annoying, and most people quickly grow to hate it. Some people completely disable Flash on their default browsers.

Nice URLs

Imagine the URL http://www.myhost.tld/index.php?section=about&subsection=personal&page=bio. Such URLs that contain an excessive amount of CGI parameters are hideous. A good site is served on the root, and uses nice URLs with slashes. Something like http://www.myhost.tld/about/personal/bio/. Assuming you’re using a server-side scripting technology, you can easily implement it using the PATH_INFO environment variable. (You should avoid using Apache’s mod_rewrite and friends for that, because they cause a lot of trouble for such things).

The sections in the site should be properly nested. If something belongs under something else it should follow its URL with a slash and with the new component. So for example, graphical art created using Inkscape should be under “/art/graphics/inkscape/”, etc. It will be a good idea to browse other people’s homepages and see what kind of things they have there to see how to appropriately section and sub-section your home site.

Absolute Links

A good site shouldn’t have any links like http://www.mysite.tld/some/place/ within the site, as they will break if the site moves to a different place. Preferably, no links that begin with slash, should be present, either, because they will break if more components are added or removed to the beginning of the site. Coding a logic that will calculate the relative link to an absolute path within the site, is not very hard using with some rudimentary programming skills.

E-mail for Giving Feedback

A web-site should have a meaningful E-mail at the footer for giving feedback. Many web-sites have a form for submitting feedback. This is less comfortable for many users (including most power users, which you should really cater for), and often also breaks in many browsers.

Some people believe one should not mention his or her E-mail on the web so he won’t receive any spam. So let me assure you of this: spammers will eventually learn of your E-mail address whether you like it or not. The best way to battle spam is to make sure you filter it on your side using a good spam filter. (I’m using SpamAssassin and am very happy with it, but there are many others).

On the other hand, preventing legitimate users from sending you mail just by pressing a link at the bottom is going to make everyone annoyed. So make sure you have a link like <a href="mailto:webmaster@myhost.tld">webmaster@myhost.tld</a> at the bottom of all your pages.

It would be a good idea to also have a contact form, so people can contact you just by browsing the web. But it’s not a replacement for an E-mail address.

Accessibility

Accessibility is a very important aspect of a good web-site. There are several aspects of accessibility:

• Accessibility for People with Disabilities - not everyone can view the web site in with full colours, in full page and clearly. Some people are blind, or even both blind and deaf. Some people are colour-blind. Some people can’t see rapid flickering properly. And so forth. A good web-site uses semantic markup that accommodates for all of them, and enables viewing the web-site in any medium.

• Accessibility for Different Browsers and Platforms - not everyone are using Microsoft Internet Explorer 6.0 on a Windows machine. Some people are using Linux, Mac OS, or a different version of UNIX. Some people are using Mozilla Firefox or a different browser, and their percentage is significant (about 10% as of November 2005) and growing. Plus, they are a very important minority, because they tend to know better than the masses who use Explorer.

  A site should be compatible with as many browsers as possible. This should be done by adhering to web standards, and using portable markup.

• User Interface Accessibility - a web site should have a familiar user interface, which people would be able to know their way around very easily.

Site News

Visitors who frequent the site would probably like to know what has changed in the site since their last visit. Unfortunately, many web-sites don’t have an accessible feed of news items. Generally, the front page of the site should display the most recent news, with all the old news items archived somewhere for posterity.

Recently, news syndication technologies such as RSS or the more modern Atom have emerged that allow users to concentrate the news items from various sites. A good web-site should have such feeds for what has changed there.

Site Search

A search engine for your site is a wonderful enhancement for your site. You can easily set up such a search engine using the Google Search Code and similar services from other online search engines.

How to Build such a site?

By now, you probably have a good idea about what to put and not to put in a good personal home site. However, you may not know how to do it. So here’s a step by step explanation of how to build a web-site for people who are completely new to programming and web-design.

1. Learn HTML According to the Standard. The first thing you need to do is learn standard XHTML (an XML-based dialect of HTML, which is cleaner and more powerful than standard HTML). I recommend the HTML Dog HTML and CSS Tutorials, which are very good.

2. Learn to Use a Web Site Generation System - complex and good sites are impossible to maintain by editing a large number of separate HTML pages. For example, you want all the page to have a navigation menu and a common layout and you’d better maintain it all in one place without keeping a lot of duplicate markup. Here is an overview of several alternatives:

   • Latemp - what I’m personally using for most of my sites. Based on Website Meta Language which has a relatively steep learning curve. Latemp and Website Meta Lang are very powerful and capable, and one can create very good sites with them, with very little redundant markup.

   • A Templating System - Perl’s Template Toolkit is an extremely powerful and capable templating system. HTML::Template is not as powerful, but easier (and, in my opinion, much lamer). The Python programming language has a templating system called Cheetah which I did not look into yet and cannot review.

     Clearsilver is a cross-language templating system, that can be used from several common languages, and combine code from any of them.

   • Other Web-site Generation Systems - I’m aware of WebMake, and you can find others in the list of Open Source Content Management Systems.

   • Rolling your own - once you learn programming, you can naturally roll your own framework for generating your site. This will not be hard at first, but you may find that you may need to constantly enhance it, fix bugs, etc. As a result, it will be a better idea to use an open-source framework, that is already quite usable, and will be maintained into the future.

   • A Note about WYSIWYG Site Creators - some web workers are using WYSIWYG (= “What you see is what you get”) site creators to maintain the content of their sites. Among the most-prominent ones are Microsoft FrontPage, Macromedia (now Adobe) Dreamweaver and the open source Mozilla-based Nvu. There are a few facts that you need to know about them.

     The first is that they often have generated markup or server-side code that is often non-standard, non-semantic, or non portable across browsers. Hopefully, it’s better with more recent versions of them, but possibly still not perfect. The second fact is that they often have a “lock-in” behaviour, in which you only get the final output, and not the actual internal templates used to generate them. Reverse-engineering them is possible, but less trivial than with the programmatic site generation frameworks that I described.

     Finally, another problem with them is that when writing HTML one has to write a semantic markup for the pages, and then decide how to style it appropriately, rather than design a page visually, and then generate markup.

     Whether you use such tools or not, you still have to learn HTML and CSS beforehand, to have a good understanding of what the tool does behind the scenes. I personally had little experience with such tools, as I find writing markup by hand to be very quick, fun, and like the fact that it produces optimal results.

3. Write the Code - now comes the fun part: write the markup. You can start from a very minimal site, and gradually expand it, and, if necessary, clean it up. Don’t worry if you don’t have enough content at first - you can always add more later, and you can refer to my earlier ideas for content you can always put on your site.

   You can use a web server running on your local machine to serve the files on your local host and test that everything works. Apache is a very nice cross-platform and flexible web-server, and it is probably included in your Linux or Mac OS X system assuming you are using one of them. If you’re using Windows you can always install it separately - it’s open-source and gratis.

4. Set up a Good Hosting for your Site - once you have a basic site ready, the next step is to set up a good hosting for your site. A good hosting has a good connectivity to the Internet backbone and the rest of the world, does not display any ads or pop-ups, and allows one to have an unlimited bandwidth (possibly while paying for more bandwidth as it is generated). It will also probably cost you money, but not too much.

   One can use various mechanisms to update the remote copy from your local copy. Covering them all is out of scope here, but it’s possible you can find some documentation on your hosting provider’s site.

5. Use a Version Control System for Maintaining your Site - as you work on your site and revise it, it is possible you’ll break something. A version control system, allows you to keep the entire history of the pages, view the difference between two versions, and easily perform rollbacks.

   The various ad-hoc ways of version control like keeping .bak, .bak2, etc. files or keeping periodic archives, are not as robust, or failsafe as using a true version control system.

   The Better SCM site contains a lot of links and information regarding the various version control system. I am personally using the Subversion version control system, which I can wholeheartedly recommend.

   In any case, note that you must eventually use a version control system, because otherwise the integrity of your work would be at risk.

Now you should remember that your homepage is a process. Make sure you add more content as time goes by, refactor it, adapt it for more modern styles, fix problems, and in general give it some love. The more you care for your home site, the more popular it will become, the more visitors it will attract and the more fun you’ll have.

Critique of some Home Sites

The purpose of this section is to review some prominent home sites which I’m familiar with, and see what their originators have done well, and what they have not. All these web-sites have these things in common: they all belong to people I don’t know very well, and they all have great content. I will conclude by reviewing my own personal site.

Paul Graham’s Home Site

Paul Graham’s site is beautifully designed, and contains a lot of interesting essays and articles. It has a navigation menu to the left, made out of images, but it does not expand or change from page to page. What’s worse is that all the pages are present on the same top directory - /arc.html, /faq.html, /books.html, /nerds.html. Let’s suppose you end up at this page, called “zero.html”, what page does it belong to? Even the Google back-links feature is no help today. (The spoiler is that it belongs to this page).

Graham routinely adds new essays to the site but he does not maintain an RSS feed by himself. Instead, the RSS feed is maintained by someone else, and often lags behind the dates in which the actual essays appear.

The URLs of the various pages are often not semantic enough. For example: /hp.html. Without accessing the URL, what do you think of when you see “hp”? My immediate guess would be Hewlett-Packard and Google agrees with me. I might also think of “Harry Potter”, but in the case of Graham’s site it stands for “Hackers and Painters”.

The Joel on Software Site

The Joel on Software site hosts Joel Spolsky’s weblog about the software industry, as well as articles he routinely publishes. The web-site I’m talking about is its slightly older incarnation, as Spolsky heavily modified the style and layout of his homesite recently (for better or for worse). The front page used to be a tag soup, that did not even validate as HTML 4.01 Transitional. Now it’s much better, but there are still many problems. Some of the other pages also do not validate. The web pages do look fine in the lynx browser, and so they are most probably very accessible.

The site contains a navigation menu and some links to the left, which is good. The book reviews section contains only the books Joel thinks are a must, and not others that he read and reviewed in time, or mentioned. (That’s not good.)

Some of the URLs contain long numbers: http://www.joelonsoftware.com/articles/fog0000000069.html. This is quite user-un-friendly. It has been fixed in newer articles, but there are still some leftover articles like that.

The site looks great, and is very aesthetic and pleasing.

Eric S. Raymond’s Site

Eric S. Raymond’s Home site contains a lot of software Raymond has written, documents he wrote, and some information. Most of the pages on the site validate as XHTML 1.0. Most of the pages contain a rudimentary navigation menu to the left, but it remains static, and doesn’t change. The pages are quite plain looking and boring from this regard. The site contains a site map which is a good thing.

The pages in the site are arranged hierarchically. So for example the essay “The Cathedral and the Bazaar” is below the Writings directory. This is also good. Sometimes this hierarchy as far as the URLs is concerned has been broken, due to historical or other reasons.

Daniel J. Bernstein’s Personal Site

Daniel J. Bernstein’s Personal Site is probably the worst best site on the Internet. It uses plain HTML pages without any style or eye candy. Similarly to Paul Graham’s site, almost all the pages are found under the root directory - “/qmail.html”, “/djbdns.html”, etc. There is no navigation menu, or any other organisational or navigation aids.

Critique of my Own Personal Home Site

My home site has a common look and feel, a navigation menu, a breadcrumbs trail, and a site map. Most of the pages there validate as XHTML 1.1, It has a news feed that mentions what has been updated in the site, and old News items are preserved in their page.

I heard ambivalent opinions about the design of the page. Some people said it was attractive enough, and other claimed it still looks plain. It has some colours, some images, etc. but I otherwise didn’t try to push HTML to its limits in its design. Instead, I opted to create a familiar home site layout that the vast majority of visitors will feel comfortable with, right away.

Some of the sections of the homepage, contain their own navigation menus, in order to make the main navigation menu less crowded.

One thing bad about the site is that it stretches across two domains: http://www.shlomifish.org/ and http://vipe.technion.ac.il/~shlomif/. What’s worse is that the number of components in the URL of each domain, is different by 1, and so may be confusing. There are cross-links from pages of each domain to those of the different one.

The layout of the site is mostly hierarchical, but often breaks, due to the fact pages were placed at certain places, and I’d rather not move them, so they won’t break in the results pages of search engines.

Conclusion

Having a good web-site is a lot of hard work, but it’s also very fun and rewarding. It is my hope that this article explained how to get on your way in creating a wonderful web-site. Alternatively, if you already have a home site, then this essay may have given you many ideas on how to perfect it.

I’d be happy to hear about what you think of this article, your experience with setting up a web site of your own, and may be able to evaluate your present and future web sites.

So until then, happy web-authoring!

Links and More Information

Eric S. Raymond has written the HTML Hell page, which makes a pretty good read. Jakob Nielsen’s site about web usability has a lot of good advice. Especially of note is his the worst Web design mistakes series.

The Joel on Software site has an excellent online book called “User Interface Design for Programmers”. It’s a very recommended read and also incredibly amusing. Finally, Michael Crawford has written a very good essay about White Hat Search Engine Optimisation.

Acknowledgements

Thanks to Sagiv Barhoom, to the Freenode IRC network people b0at, cochi, intrr, Pete_I, Windrose and kaitlyn, and to the OFTC IRC network’s JasonF for commenting on earlier drafts of this essay.

Coverage and Comments

• Whatsup (in Hebrew)
• Linmagazine (in Hebrew)
• The Joel on Software forum