Josh (the blog)

I’ve delivered simple, clear and easy-to-use services for 20 years, for startups, scaleups and government. I write about the nerdy bits here.


@joahua

Something about backwards search engines

No, I’m not talking about elgooG.

The Sydney Morning Herald published an article entitled “New Australian search engine launched” today, the first paragraph of which reads “Australia’s newest search engine Ansearch opens for business today with a novel twist, demographic searching.” It’s not a particularly well written article, but the article vendor is AAP, not the SMH itself, so we’ll leave that alone, at least for the minute.

It goes on to laud the search engine for their innovation, both in this feature of demographic searching, and in other areas:

Ansearch says it cuts down search clutter by displaying the main search results as single websites and not the individual pages of websites.

What, like the Google [More results from domainname] feature? You know, the one that actually works properly? I say “works properly”, because a quick search of Ansearch reveals that their “cutting search clutter” feature is a tad broken — not to mention their character encoding.

Proof that it's broken, demonstrated by duplicate entries and incorrectly encoded characters

Speaking of broken character encoding, let’s take a look at their source… well, they get some marks — at least they bothered with a Content-Type. Never mind if the content is broken when displayed with that Content-Type — it’s not like a search engine could actually do any useful data processing to make things display correctly when using a slightly redundant Content-Type… oh, wait, disregard that comment: they’re not using a doctype, either.

See, what gets me is that this search engine has just been launched. Which means the climate in which it’s been developed isn’t the same as 5 years ago, when accessibility was just on the very edges of the radar — you’d (wishfully) imagine that at least a doctype wouldn’t be too much to ask for, even if they still insisted on using table-based layouts. Interestingly enough, that’s what one of their software providers, Omniture, have done. Which leaves something of a foul taste in the mouth, too, because they’re reselling that garbage to people — including, if you believe their website, three of the five top Fortune 500 companies (aside: doesn’t that make them three of the top Fortune 5?).

Perhaps that criticism is unfair — their latest version (assuming that’s what powers their own website, although possibly not… maybe their internal web team accepts that their product would be overkill, and coded it in Dreamweaver, instead… some of the JavaScript certainly looks Dreamweaver-esque, and, if Ansearch’s website is any example, the doctype probably doesn’t come from the Overture system!) seems to handle much better than what Ansearch are running: I say this, because apparently they’re using a version which was written back in 2003. Hey, if it works… but we’ve already established it doesn’t.

And there concludes my rave review of yet another quite-some-way-from innovative and fresh search engine, this time in Australian waters.

Note: I’m not saying it’s any worse in terms of accessibility, usability, and semantics than most other search engines are — only that it has less excuse, being launched now, as opposed to 5 or 10 years ago. It’s easier to make something work first time than it is to haphazardly patch over it later, especially something as gargantuan as I’d imagine a search engine would be.

Oh, and now for something that’s just plain amusing — the number 1 search terms on this brand new search engine, from befuddled users wondering why it sucks so much:

Top search queries are Google for both Weekly and Monthly statistics

Yes indeed, the first thing users did was try to escape… how’s that for telling?

About the new theme, and a bunch of numbers

Okay, so maybe throwing a new theme up like that straight away wasn’t the greatest of ideas… but you’ve gotta admit it looks cool. I’ve just patched up the contact page (n.b. the address has changed, coz it’s now managed through WordPress Pages… I’ll figure out a way around this, and both the old form and the new themed one work, but for the time being the address is <a href="/blog/contact/">/blog/contact/</a> rather than the old <a href="/contact/">/contact/</a>), as well as a few posts that were being dodgy… especially with graphics, which’d been designed for a fixed-width site of 600px, as opposed to… whatever this is. I haven’t got time to go slashing at CSS just yet, but it’ll happen, unavoidably! (I’ve heard and completely agree with readability comments… that’ll be one of the first things to be fixed, because it’s a real problem…)

In other exciting news, I just cracked 1GB of transfers this year.. which doesn’t sound like a lot, but I did around 470MB last month, and this month (albeit 4 days in) is looking to be similar in terms of transfers and hits if the trend continues. Last month there were 44,444 hits (isn’t that a cool number?) exactly, probably in no small way as a result of the moderately successful DashLite, which has got a ridiculous number of trackbacks or comments in the last 24 hours (well, I thought it was… shrug!), taking the total number of comments on that post to 15, probably the most comments on any single post on this site.

Now I’m gonna wander off cram for a Business exam I’ve got in a few hours, so I can pack it full of buzzwords and be awarded marks accordingly… *sighs at folly of that subject*

404: Post not found

Ben pointed out to me that last post was the four hundred and third (Forbidden) post, so I thought I’d actually announce this one as the post which was not found, for no reason other than that it’s the four hundred and fourth post on this website.

The post you were looking for does not exist. Please try using the search facility provided on this website, or, alternatively, go back to reading the weblog normally and just pretend you never read this, because it’s utterly meaningless…

WordPress 1.5

I’ve finally shifted this weblog across to the most excellent WordPress 1.5, which means I can now use my own DashLite, as well as improved themes management (which I’ve unceremoniously trampled upon to keep this current behemoth working quickly), spam combatting features, and slightly-more-than-rudimentary content management capabilities.

I was hoping to release a new theme with this upgrade, but whilst I’m keen on procrastinating, I’d really rather not fail my exams tomorrow and the subsequent day… maybe later this week.

Ten minutes later: Stuff it. I’ve banged in the theme I was planning to base the new one on, coz it’s just so cool I couldn’t leave it a moment longer! Full credit, of course, to Chris Lin. I’ll be making bits and pieces different so it’s not just a stock theme, but I felt like a change and hadn’t had time to whip up something of my own, so this’ll do for the time being (plus it’s sexy!)

Graveyard retired

Some more attentive regulars (who don’t just peruse this website by means of syndication) may have noticed the disappearance of a link in the top bar in the last several hours. This is because I’ve finally got all the old content into WordPress, with no small amount of assistance from Michael, under a category called “Before WordPress” (this post is categorised similarly, and shall likely be the last ever entry into that category).

Practically, this means that that content is using semantically better markup, has better meta information for search engines, and is internally searchable, using the WordPress search function (it wasn’t before).

For most regulars, this probably doesn’t mean much, but the old articles attract the most search engine traffic, so this’ll be of benefit to people finding relevant content, at least, because the old script could be somewhat retarded in the way it was indexed, as there was no formal permalink structure, just a bunch of loose query strings, which search engines didn’t like.

<\/geek off>