A word to the wise about virus protection.
URL stands for uniform resource locator. It’s more politically correct, and perhaps logical to call it URI — the I is for ‘indicator’ — now, but nobody outside the W3 Consortium does so.
I occasionally get complaints that the URL (see box) is ungainly and unmemorable. I could have called this site something like lescroupiers.co.uk or lescroupierscardiff.com, but both of these would have invited confusion with Gordon MacIlroy’s casino (see our history for more on Gordon). Using the whole name was intended to make our nature clear to everyone. The .org was just to hammer our non-profit basis home, and leave www.lescroupiers.com and .co.uk uncontested for Gordon’s businesses. (After all, he did help us get off the ground — and still signs a small annual cheque.) The final .uk, which many seem to see as adding insult to injury, is there to offset the ‘Les’ part which could be taken to be French, and to reinforce our geographical location — because we are a very real, and not at all cyber, organisation.
We registered through, and were originally hosted by, Ghoulnet. I bought some internet magazine in WH Smiths and just went through the ads looking for registering and hosting companies. Ghoulnet was my third choice, but the first (and therefore only) one whose shopping basket worked at the first go.
We’re now hosted by Mythic Beasts. I’ve been using them since January, and I’ve found their support staff helpful and even supportive. They are also admirably prompt at announcing problems on their site, and do so in straightforward, demotic English, as opposed to curdled management-speak.
This site was originally hand-coded in the sense that no WYSIWYG editor was used. Most of the original HTML was written using Arachnophilia, which is really just a souped up text editor. I now use vim: one of the advantages of our new hosts is that I can edit pages directly on the server, which saves me some time, which I can then employ more usefully by staring into space with my mouth open.
Every page on this site is a composite of discrete files shuffled together by a server running Apache on Linux, which is half the reason why every page now looks pretty much alike. This particular miracle of modern technology owes everything to PHP, which is the lazy person’s all-purpose programming language. PHP handles all the functions which show when pages were last updated (I never have a clue as to what date it is). This should clear things up for anyone who’s noticed that pages suddenly gain extra navigation, or the club logo changes overnight while the page says it hasn’t been updated for months. It’s also responsible for the contact forms, the (hidden) hit counter — which doesn’t count search engines, and serving subtly different versions of each page to ‘standards-compliant’ and ‘legacy’ browsers (see CSS below).
If anyone is at all into acronyms, PHP [allegedly] once stood for ‘Personal Home Page’; it now stands for ‘PHP: Hypertext Preprocessor’. Source: www.php.net/manual/en/history.php. Think that’s bad? Try GNU: Gnu’s Not Unix.
Ace uses our MySQL database. Indeed, the whole site could be database driven except that I find database urls (which often look like “?page=lotsofgarbageand numbers0123456789”) to be hugely counter-intuitive. I’ve been saying this for a couple of years now, but next year the Castles results will be in a database so each club can print out their own runners. Of course, it’s not my fault that we haven’t done it so far: the results I get are in the wrong format. [Excuses, excuses — Ed.]
Some results pages such as the SSAFA 5k series were written in XML, and are turned into HTML by the server. This meant that I had to prepare only one set of results, play with it a bit, and the rest was done by the software. This came in very useful when Terry Caveney was discovered to be missing from one race. If I’d done the results the normal way, I’d have had to renumber every finishing position after him — most of them.
We also have RDF and RSS files, which are really more appropriate to weblogs. They’re there because I was interested in methods of generating them, rather than because they serve a useful function. Unusually, they’re not generated when a page is updated. They’re generated when the homepage is loaded, which checks the last updated date of all the major files on the site, and if and only if one of those is newer than the RDF file, then a new RDF file is written along with an RSS file. If I updated the site, but no one looked at it, nothing would happen. It’s not a perfect solution, but as people do look at the site, it works just fine.
The other reason that this site looks the way it does and is very easy to update, is its use of Cascading Style Sheets (CSS). Users with older browsers still see a semi-styled version of the site as I worked out a cheap detection hack, but the look of the site is aimed at version 5+ browsers, especially Mozilla (Mozilla is just Netscape, with the subtle difference that it works) and Opera, both of which work (more or less) on all platforms.
I used to meddle with some aspect of the overall look every week, and am permanently unhappy with the link colours and the logo as red-green colour-blindness affects something like 1 in 20 men, which is something I didn’t even consider originally. (See below.) I have done something about this, making visited links significantly fainter. As the "feel" of the site was described as ‘a bit old fashioned,’ I took away the link underlines, favoured by Jakob Nielsen, but I restored them as I felt that bold links were too intrusive and non-bold, non-underlined links too subtle. I’ve assumed that no one confuses either red or green with black or white.
All the current browsers now have similar features, but Microsoft’s Internet Explorer (IE) is by the far the most popular browser — more visitors to this site use the recently released IE6 than all the versions of Netscape added together. (This is still true, but Firefox is catching up. The differences are subtle — Internet Explorer doesn’t understand a rule which says that if paragraph follows a paragraph, it is indented. (Look in any book to see what I mean: the first paragraph after a heading or a quotation, or a new chapter is not indented, while all the following ones are.) It’s a trivial CSS layout thing, but one that matters to me, and one that Microsoft will get right soon. The stylesheet includes this:
p {padding: 0; margin:10px 20px 0; border: 0; line-height: 1.3; }
p + p {text-indent: 1em; }
IE, as I’ve said, doesn’t understand p + p, so, as I prefer paragraphs to be indented, I get around this using Conditional Comments, a nice little Easter Egg which only works with the Microsoft browser. As they’re hidden in comment tags, other user-agents ignore them. So no iffy browser-detection (which can go wrong, now that there’s a plethora of new browsers out there), and the W3 are against the practice of banning or serving different pages to different user-agents. (User-agent only means the software which downloads the page which includes spiders, validators, and browsers.)
<!--[if gte IE 5]>
<style>
p { text-indent: 1em; }
</style>
<![endif]-->
This tells IE to indent all paragraphs, while not spoiling the ‘correct’ layout in others. There are other ways around this, such as Tantek Çelik’s Box Model Hack.
(And speaking of logfiles, I am obsessive about visitor behaviour, and spend an unhealthy amount of time reading my weblogs and fretting about the search terms people use to find us. Reading logs can be difficult and tedious, and I find the best tool is analog, available from www.analog.cx, which is free, and easy to set up. It gives you the basics, how many hits came from searches, how many 404s (file not founds) were generated, but there’s nothing like trying to follow users surfing your site, or reading their often strange search queries. I thought that I was a little weird in checking out the pages they came from, but I’ve just discovered searchrequests.weblogs.com. Suddenly, I’m not alone any more.)
On the plus side for Microsoft, most of you will be reading this page in a large serif typeface called Georgia, which was invented for Microsoft along with its sans-serif sibling, Verdana (which is used for headings and captions here), in the mid-90s by Matthew Carter. I’d always wanted a serif font for this site, just to go against the grain. In print, serif fonts are easier to read; they just render less well on screen as each letter is made up of fewer dots. Georgia — with miniscule serifs — gets over this hurdle smartly, as well as being more ‘organic-looking’ than the slightly futuristic Verdana.
Georgia ... is now enjoying a vogue that borders on trendiness, says the mighty Jeffrey Zeldman. I disagree with the article on Arial he links to, preferring webactivism.org’s take on the Arial vs Helvetica debate. And while it’s nice to be in with the in crowd, webactivism.org go on to say At 13 pixels, Georgia starts to exhibit some glimmers of an entirely new character, gently rounding, which I entirely agree with, so why two of Zeldman’s examples bastardize the font by forcing it to 11px, I’ll never know... The good news is that a slender majority of Unix users now have Georgia installed, and as it comes with IE on the Mac, that means almost everyone now has it.
However, I do go with the grain sometimes, and I’m aware that this involves alienating some users. For those of you still using Netscape 4 and feeling pissed off, this is pretty much the party line. It makes sense to me. All our pages validate with the W3 Consortium, and should therefore work everywhere (but see above), or, where I’ve been careless, are supposed to.
The logo is mostly my own work. The club’s proper logo was designed by Rob Evans, and it looks just fine on T-shirts and notepaper. The version I used to use was a scanned printed copy, and ever-so-slightly squint. It was also nearly square and didn’t ever seem to fit with the rest of the site. If I’d thought about it, I’d have adapted the site to fit our existing, and known, logo, but I arrogantly took it upon myself to come up with something better.
The first change was adopting a logo designed by Matt Townsend, which did for a while, until I decided that the way forward meant having a thin strip across the top to act as logo and title, and I came up with stealing the red man from Matt’s image and writing the name in most Welsh way possible, which meant a celtic looking script (American Uncial, which came with Windows ME), and a conscious parody of the flag, with the runner taking the place of the dragon. (The dragon, I’m told, faces west so he can shit on the English, but running to page right seemed the more natural option. Track runners in the home straight move from a spectator’s left to right, and the reading eye also tracks right.) Rob chose the red and green colours, visually punning on the roulette table and the Welsh flag. His original work for headed notepaper favoured the ‘casino intrepretation’: I’ve chosen to emphasise the Welsh link.
The ‘Running Club’ part is in Nano Sans, a very useful and beautiful techno-looking font (only $15) designed by Joe Gillespie in 1982 for computer games. If your arms are too short for your pockets, there are always free fonts, like Jason Kottke’s Silkscreen.
What’s wrong with the logo is that some people are colour blind. Most of these are men (perhaps 8% of the population, although Europeans are the most common sufferers), and most can distinguish red and blue, but with the greatest area of confusion between what the rest of the population see as red and green. (When I was at university, the common evolutionary explanation for this was that men mostly hunted and colour perception was more important for fruit-picking — done by women. This has always seemed like weak post-hoc theorizing to me.)
The images to the right show how the club logo may appear to red-insensitive protanopes (top), green-insensitive deuteranopes (middle), and blue-insensitive tritanopes (bottom; these do see blues, just not the blues others see). Sadly for the logo, deuteranopia, or insensitivity to green is the most common form of colour blindness, and it’s hardly readable.
Credits: Anitra Pavka in Digital Web Magazine and Christine Rigden of innovate.bt.com. I knocked the images up using this colour chart; I’m pleased that they are quite close to Vischeck.com’s own guesswork.
Until I bought a digital camera in January 2004, most photographs on the site were taken with a film camera and scanned in. The uncredited ones were taken by me or with my camera if I was running at the time. Rob Evans, who is far more skilled behind a lens than I am, took many of the good ones. Most of the others were sent in, or are used with permission from other sites. (This mainly applies to photos from the Castles, where we didn’t have a good enough archive of pictures.)
The look of many would be considerably different if I’d known what I know now when I started. Many are too dark; and many have been over-compressed. I’ve come to regard the best size as being around 250–300 pixels on the long side and the ratio of the short to the long side being around the golden ratio (that’s the one paper comes in, so that A3 cut in half through the long side gives two sheets of A4; A4 cut in half gives two sheets of A5 etc.) - or around 1:1.6 (0.62:1).
It would also be nice if the pictures came in regular sizes. Again, this is an idea which didn’t even occur to me when I began. I’ve held off working toward this as I’ve thought that the images are too variable to fix their size in advance, but this hasn’t bothered the BBC. See their coverage of Paula Radcliffe in the 2002 European 10000m. (And those pictures are 300px x 180px, a ratio of 5:3 — see previous paragraph.)
Most are now in the ratio 4x3, because that’s what my camera takes. And I like to think that the quality has improved too.
I’m also currently bothered with the text — on the home and about pages. If anyone’s interested in why I list foreign marathons no one is likely to do on the calendar page, it’s because we used to tour around Europe doing the big marathons; it’s a good way to see a new city. It’s also (you knew this was coming) because I decided long ago that the word most non-runners associate with running is ‘marathon’ (as evidenced by the kids in my street who used to ask me how many marathons I’d run — every day). So the word marathon had to turn up a lot somewhere…Keyword Density Analyzer
When I changed the site over to php, we had a counter on the front page. After it passed 5000, I took it off. I wanted to know what sort of traffic we were getting, so I wrote a counter counter, which updated itself with every 1000 requests for the home page.
| Hit | Time Date |
|---|---|
| 891000 | 07:30:12 23/07/2008 |
| 890000 | 12:18:43 21/07/2008 |
| 889000 | 07:46:19 19/07/2008 |
| 888000 | 07:28:34 17/07/2008 |
| 887000 | 12:53:35 15/07/2008 |
| 886000 | 21:33:02 13/07/2008 |
| 885000 | 20:29:57 11/07/2008 |
| 884000 | 12:23:41 10/07/2008 |
| 883000 | 15:09:32 08/07/2008 |
| 882000 | 00:20:09 07/07/2008 |
There are a lot of hits from search engines, which range from the obvious, like "running clubs", through the stupid, like "croupiers" (we get a lot of hits from people clearly looking for casinos), through the perverse — we get at least four searches for ‘naked running’ per month because of Matt’s animated gif which I took off for that very reason, to the weird and surprising: we rank highly on sites about TS Eliot’s essay on Hamlet due to an unthinking remark I made somewhere.
Last updated 24 March 2007
^Top | Sitemap | About this site | Contact us | Les Croupiers Running Club Home
