tenpastmidnight blog

One of the clients I do linking work for has a links page which means they can exchange links with other websites.

Recently they've been dropping down the rankings in a couple of search engines for no reason I could find. Their general linking was going well, their new website was all good semantic HTML, built by Nathan, and they definitely weren't using any spammy techniques.

I think I found the problem when I ran through the websites they linked to: one of them had been taken over by a link affiliate network. This is a website that is full of links to other websites, all with affiliate codes in so if people follow them, the site owner gets paid for the traffic coming through, or a cut of any purchases made, depending on the network. The links on this particular site mainly seemed to be pointing through the Overture PPC system, so it may well have been someone trying to fool people in to clicking on links they did not realise were adverts.

The Google Page Rank of the site had been reduced to zero, which was a bad sign as it had previously been four. This is a sign that Google may have banned a site, which means if you link to it, you get a negative effect on your own rankings. So hopefully, removing the link to it means my client will get their own high position back.

The site they had linked to was legitimate, at the time. What's presumably happened was the domain came up for renewal and the original company didn't renew because they closed, or forgot to renew in time and the new owner swooped in and scooped it up. If you run this sort of link farm, getting a domain which has plenty of links to it is perfect, you don't have to do any of the linking work yourself, as long as no-one notices that the site has changed.

If no-one notices the site has changed, the link farm can do well - they'll get traffic through the links, and potentially a good ranking in the search engines for a few terms, at least for a while. However, if they trip an automated filter, or someone reports them via an abuse form, then they plummet in the rankings, and everyone who linked to the original site gets dragged down with them.

The lesson for me? Always insist on checking sites for my clients, even if I'm not dealing with their link exchange work. As links they set up can negatively affect my work, I have to make sure I can at least keep track of it.

It's one year since I went up to Leamington Spa to work on a contract for Big Picture Interactive, my first contracting experience rather than my now normal freelance work. I needed the money (rather desperately) and was rather nervous about going 140-odd miles just to work with a bunch of people who hadn't even given me a phone interview, but it worked out well and everyone up there turned out to be very nice and wanted me for far longer than originally planned.

The difference between freelancing and contracting is a bit of a fine line, especially for a one-man-band like myself. Basically being a contractor usually means working through a recruitment agency, working for one client at a time on a particular job. Being a freelancer means having work agreements direct with clients that you've found yourself (through your own marketing, referrals, whatever) and that you generally have a few things on the go at the same time, or at least you do if you want work to keep coming in.

Doing pure contracting would probably be more lucrative, but also means working away from home a lot more. Personally, I like being near my friends and family, and I prefer working with companies where I can have a long-term relationship with them as we try to improve the on-line side of their business. To me contracting means being plunked down inside a company to fix a particular problem, then usually never being seen again unless you happen to keep in contact with people you met there. Freelancing is potentially a long term deal where you're used over time and trust is built up in what you do.

All in all, it's still sitting in front of a screen writing near-gibberish that happens to make sense to computers somewhere. But I'd rather be doing that as a freelancer than a contractor, even if all the companies I'd be placed in where as nice to work in as Big Pic.

There's a common misconception that web developers also know everything about PCs and other computers. It's understandable, we tend to know more about the way things run than most people, and are the sort of people who look after their own PCs, and have probably built one or two in the past.

However, like lots of developers, I avoid trying to run the server the website I program run on. Systems administration is a whole other art, and I'm happy to leave that up to the experts. One of my clients is having trouble with their MySQL server at the moment, and although I'm frustrated because it's not working, I'm happy that it's not my problem to find out what the heck is wrong with it and try to keep the thing running. Although it's my responsibility not to run anything that's going to cause the server to crash, it's a completely different skill to keep servers running 24 hours a day, every day of the year.

Fortunately, I know the systems guys who look after the servers on most of the sites I'm heavily involved with. Paul runs the server with all my personal projects on, including my work site and The Farm, which we're both part of. Another Paul runs WPC, and although I've only talked to him by e-mail, David knows him well, and Olly (who I'm sure has a site, but I can't remember the address) looks after the servers where several of my client's sites and my much neglected tenpastmidnight site lives.

Several of the servers I use have had unusual glitches recently, although not as bad as the MySQL server I've been trying to use. A few days ago I had a bit of a panic when I realised TPM was down, and so were several other sites, one of which I needed to do a quick bit of work on. Then I remembered Olly had told me he was moving the servers. This wasn't moving the data, but physically moving the servers between offices. I hope he used his car (or borrowed a trolley from the nearby Somerfield) as I remember the servers being ruddy heavy when they arrived and taking two of us to lift, and the UPS made them feel light.

In all it could only have been 2-3 hours downtime, which surprised me as I expected it to take a lot longer as the IP addresses were moving at the same time. This is exactly why I'd rather have good systems people - they know how to swap this stuff over easily without any long problems, or having to re-set DNS to new IPs, which would have been a real pain.

(I should mention Paul no. 1 is also a web developer, which confuses the matter more - it's not that web developers can't also be good sysadmins, it's just that usually we aren't.)

There's a new version of Google Desktop Search, which reads files on your hard disk and lets you search on them, rather than searching the web.

Last time I used Google Desktop I uninstalled it because it didn't serve my needs - it didn't read pages Firefox has looked at, and it couldn't search in OpenOffice, PHP or ColdFusion files. Now it can.

It indexes the Firefox cache as standard, you need one of the new plugins to handle OO, PHP and CF. To handle the latter required configuring the Any Text File Indexer when you first install it, which is very easy if you can write the stuff inside the PHP files.

It's indexing my PC at the moment, I'm hoping it's good enough to keep around this time.

The Register reports that Yahoo is rebranding Overture, but more interesting is the comment that they are "hoping to tempt more developers to its search products by providing easier access to its Web service APIs." This fits in with trying to catch up with Google, who have had a search API for a long time, and also with the rumour of an upcoming O'Reilly Yahoo Hacks book (fitting in with the rest of their Hacks range.)

I'm interested in a good Yahoo API, the one Google provides didn't have the same index as the one available on the web, which made it a bit useless for what I wanted to do. I'd like to experiment further with the Google one, but as well as finding the time I also need to be sure it's going to give up to date results, otherwise it's really just a little toy.

Yahoo all ready give RSS feeds for some of their services, so an API could be seen as a next sensible step and I look forward to seeing what it can do.