Author: Barry Parr

I’m spending time every day deleting spam from MediaSavvy. The worst part is having to rebuild the entire site every time I delete a single spam. Perhaps I need to switch to a dynamic CMS like Expression Engine, so that I don’t have to do this.

Meanwhile, my email spam filter managed to move a couple of important messages into my spam folder. I don’t blame SpamSieve, which I love. I blame the spammers for making this whole process necessary.

Finally, I get a dozen spam faxes for every fax I want. This has been going on ever since I used my fax to send a message to efax.com. You can’t tell me they didn’t sell my number to the spammers.

Detecting spam has become an entire industry.

Technology is insufficient and the law is insufficient. You need both to solve this mess. What really galls me is that the direct marketing industry has fought every step to help solve the problem and given us a law that is completely inadequate to the task. Apparently, they can’t tell the difference between what they do and what the spammers are doing. If they can’t, how are we supposed to?

[This note was rewritten. The original post was eaten by a Safari bug exacerbated by the way Movable Type posts messages.]

Steve Outing asks why so many Web sites put their (fixed-width) content flush-left in the browser window, and not in the center.

I’m not a designer, but I know why I do it. When you have centered pages and go from one site to another, the location of the top left of the page in the browser window is going to move every time you go from one site to another.

I find news.com’s centered presentation very disorienting when I move there from pretty much any other site.

Eliminating fixed-width layouts is also better because they give designers the illusion of more control of presentation than they really have. Fixed width for navigation and advertising columns is more justifiable, but content should be set free and more control over presentation should be ceded to the user.

I never thought I’d program a Web page.

I’ve done a little programming, but on the Web I have been happy to either create static pages or to use a content management system. My Web programming has been limited to using includes to modularize my pages and tossing CMS macros onto my pages.

But since I’ve started building Coastsider [password required for the next week or so], I’ve started using PHP. I started using it against my will, just to get some of the “templates” that came with my classified software to look like the rest of the site. Then, I started using it to create forms for my users.

Yesterday, I created a page that interfaced to the Terraserver’s satellite photo database, including panning and zooming. It’s empowering to be able to build software into your web pages. It’s barrier-shattering to be able to mix code with HTML to do simple tasks and to interface to your content management system. The architecture of PHP encourages experimentation.

No more than a small number of individual Web site publishers will ever do much coding. On the other hand, I now think that increasingly powerful content management systems will encourage programming, rather than make it unnecessary.

I spent some time today cleaning up my stylesheets on a big site I’m working on. I was stunned by how messy they were. I was able to reduce the size of my main stylesheet by 20%. Not only that, but this top-down review showed me a lot of structural problems that had accumulated over the months: styles that didn’t cascade properly, orphaned styles, redundant specifications, and even a couple of missing styles.

The result is that I’ve managed to resolve a few formating quirks that were due to stylesheet problems, improve the consistency of my site, and improve its performance.

I’ve once again been exploring employment web sites and it’s stunning how little they’ve changed since I put together Free Agent for the Mercury News back in 1996. Looking for jobs on the Web is a miserable experience. And it doesn’t have to be.

• Some sites still don’t work with all browsers.
• There an astonishing amount of sheer garbage in these sites, such as Work at Home “opportunities”, mainly because the sites are unwilling to police their advertisers. The newspaper-owned CareerBuilder is the worst offender here.
• It’s too difficult to sort through large volumes of results.
• The search tools are inconsistent and underpowered. For example, not all sites allow you to use boolean operators. There are severe limits on the number of criteria that you can set. It’s difficult to edit predefined searches.
• Some sites make it impossible to launch listings in tabs.
• All sites are job-centric and don’t allow ongoing meaningful searches based on, for example, companies.
• Most sites do not have meaningful metadata about jobs in their databases.
• Everyone does everything differently, and because all listings are paid, you have to search every site.
• Employers’ web sites are even worse. Ebay’s site is browser-dependent and you can’t capture a unique URL for any listing. Most sites unbelievably difficult structured resume builders that simply do not work.

I think it is easier just to read the print classifieds. The tragedy is that this is a solvable problem, or there is a solution that would work for a large number of users.

Why not encourage employers to post their ads in XML format with clearly defined declarations of things like company, location, salary, title, responsibilities, qualifications, posting date, requisition number, and so on? All employers need is the knowledge that someone would read their listings and a few modest tools to create the pages.

If these pages were created, they could be crawled and indexed by anyone who chose to do so. Search tools could be provided by portals and search engines, who would compete on the quality of their search and management tools and not on the size of their listings databases. Listings could be delivered by RSS to interested applicants.

Meanwhile, applicants would start getting tools that would let them look for a particular, well-defined job in a large, but strictly-defined set of companies. Over time, they could refine their searches with tools like Bayesian filters, looking for more jobs like the ones they want and eliminating whole classes of listings.

Finally, anyone who could establish this kind of standard could create a standard XML format for resumes that would allow job seekers to enter their resume into a single form (or have it parsed) and output it in a format that any employer would be able to read.

The incentive for most large employers to do this is huge. They only need someone to set and promote the standard. There would be a market for providing tools for this to employers large and small.

Who can do this? Yahoo can, but their fortunes are tied to HotJobs. Similarly Monster and CareerBuilder have no reason to mess up their revenues stream or those of their corporate parents. The remaining obvious candidates are Microsoft and Google. It seems like a perfect fit with Google’s current strategy.

The latest edition of Adam Engst’s outstanding Mac newsletter Tidbits reviews how their various efforts to increase revenue have been progressing. (Adam: thank you for sharing your results with us.)

The bottom line is that Google’s Adsense probably makes more sense for sites without a loyal audience, sponsorships are still their biggest revenue source, and ancillary products (ebooks) are their best new revenue source.

That sounds about right. I had high hopes for Google-style targeted ads, but they don’t appear to the be panacea for sites that their increasing share of Internet advertising implies. They demand a huge volume of readers who are looking for something. This doesn’t describe the audience of most online publishers.

As Adam points out, most the AdSense advertisers were offering products (Mac systems) that Tidbits was already recommending readers buy from their primary sponsor.

There is still no substitute for a strong relationship with you sponsors.

I heard an interesting example of politically correct editing on Morning Edition today.

In a story about how an LA synagogue used a federal law that exempts religious groups from most land-use regulations, reporter Alex Cohen referred to one opponent of the synagogue “who happens to be Jewish”. That mildly annoying “happens to be” is a clumsy PC synonym for “is”, intended to connote…well, I’m not exactly certain what it’s intended to connote. After all, the guy’s Jewish and it’s relevant to the story.

What’s interesting is that KQED ran the same story ten minutes later as part of the “California Report”. In that, virtually identical, version Cohen says that the neighborhood opponent “is Jewish”. I wonder what editorial process that led to this difference, and what the Morning Edition stylebook says about this usage.