Random Thoughts

Tech, words and musings from an Englishman in Seattle

David Harrison Reinvents The Internet

In a particularly well reasoned essay, David reinvents the control of the DNS infrastructure.

How did I stumble upon this? Well, David and I went to school together at Downlands County Secondary School (when did it change it's name to “Community School”?) in Hassocks (where I grew up) and Haywards Heath Sixth Form College (what's with the “Central Sussex College” thing?) in the south of England - both of us pretty much geeks! In fact we sold games that we'd written for the ZX Spectrum games and shipped them out on little plastic bags amongst other geeky activities.

We lost touch after we both went to our respective Universities, but happily he stumbled upon my blog a couple of years ago and we reconnected. Every now and again he sends me another great idea. Most of these ideas are utterly fascinating, if very off the wall. And sometimes they just make a lot of sense, even if potentially impossible to implement due to infrastructure inertia.

David's 'The Independent Network - An Alternative to the Internet' is one such missive. If you have any interest in the current domain name morass, it's a must-read.

And apparently I'm not the only one who's been getting David's semi-regular essays.

In his latest column, PBS' Bob Cringely talks about David's idea and, quite frankly, agrees with him.

Hopefully David's website won't explode under the traffic sent his way…

David hadn't seen the article before I sent some congratulations his way, to which he jokingly responded (and I hope he doesn't mind me quoting him): “If I'd known it was going to go public I'd have spent more than 30 mins writing it! :-)”

I think it took him longer than that…

Memeflow Is Still Up And Running

My server was unreachable today for about six hours due to some hassle at the colo, but I was reminded that Memflow, a service I put together in 2005 as an experiment, is still being happily used by quite a number of people.

I was reminded because I got a bunch of email from quite a number of people wondering why their favourite “start” page was down!


If you're looking for a fast, image-free (ok, there's an RSS icon) start page, may I suggest you give Memeflow a try?

Here are some posts about the site from a while ago.

Late To The Party - Web 2.0 Video

I know I'm extremely late to the party with this video as everyone and his uncle twice removed has already blogged about it, but some of my dear readers don't follow ye olde blogosphere as avidly as you and yours truly.

So this is for them.

This truly is a wonderful piece of video and above all a great introduction to whatever this thing called “Web 2.0” might be all about.

DayLife - Dig Deep

I've been playing around with DayLife today, and to be quite honest, I really like it. DayLife, for those who have somehow managed to miss all the coverage is a new aggregation site with a number of twists. There has been a bunch of commentary since it came out, so I'll just add a few of my impressions.

I like the design, the feel of the site. I love the use of images and callouts.

But most of all I love the ability to dig deep. Interested in that story on infant car seats Well, just dig a little deeper. Try clicking on the link on the right hand side to Graco Inc. Now you can dig deeper into the people and organizations related to the company.

I love it!

As someone who wants to know as much as possible about a story that interests me, DayLife might be a keeper.

Definitely one to follow.

Now I'm Mad

When I bought the domain steve-lacey.com, I really wanted stevelacey.com but it was owned by a traveling, musician, minister type.

Ho hum, never mind.

Well a couple of months ago the domain lapsed and came up for renewal, I entered the auction for it and got sniped at the last minute.

What's up there now?

Garbage, that's what. A fracking ad-farm. Don't click on anything.

Bastards. (Say that with an English accent please.)

IE7 Annoyance

If you visit this blog with IE7, you'll get a topbar thing that tells you that the site wants to run the “Windows Media Player Extender” ActiveX control.

Since I did no such thing, I have no idea what is causing this. Of course I could voodoo debug the site and disable everything one by one until the culprit reveals itself, but I'd rather have you, dear readers (especially you webdevs out there) tell me what the problem is.

Prize for the winner. Well, maybe not, but you'll get my thanks anyhow :-)

Tags: , , , .

GeoTagging In Flickr

I've been waiting a long time for this, and finally it's here. You can now add location information to your photos in Flickr via a handy new tool that uses Yahoo Maps.

A new “Map” button in the Organize tool and a new “Map” button on everyone's Flickr page results in pretty easy access to the functionality.

I've geotagged a few recent sets - you can check out the resultant map on my Flickr page.

So, what does this mean? Well, for a start, a lot of people are going to now start geotagging their images, which means that you'll be able to search for images by location and there will actually be some results. Image location search just hasn't been interesting before as there really hasn't been any data.

Expect to see a lot of location based image mashups appearing very soon…

CAPTCHA - A “Hot” Web Service?

Due to the deluge of spam comments and trackbacks that I regularly receive at this blog, I'm always on the lookout to find ways to reduce it. Most recently with Automattic's Askimet service. Askimet has been pretty successful for me, but that doesn't stop me from thinking about adding some level of CAPTCHA functionality to the blog.

A lot of services now use CAPTCHA. These include Blogger and most signups for various web services, but apart from the fact that they're easily broken, like most people I find them highly annoying - the letters that you're supposed to pick out of the images and sometimes just downright impossible to discern:


I've often thought that there must be better ways to do this, when out of left field, Tim O'Reilly points us to HotCaptcha from the folks that brought you HotOrNot?. Here you're presented with nine pictures of individual people, only three of which are judged “hot”. Pick the right three and you're a human.

Now, ignoring the potentially NSFW content, this is pretty cool. They have provided a webservice that uses the collective intelligence of humans to put together a question set that is almost impossible for a non-human to answer. Pretty cool.

Now, where could this go? How about something like presenting a bunch of pictures from Flickr and a set of tags, where only one of the tags applies to all of the pictures?

I'm sure that there are a bunch of potential solutions in the same vein that can get us better results than standard CAPTCHA and with none of the annoyances.

Hopefully we can stop squinting at bad CAPTCHA soon.

Dorkbot: Web 2.0 Event

The August Seattle Dorkbot (hey really need to update their website…) is going to be focussing on the programmable web.

From the call for submissions:

The dorkbot overlord committee met this week and after a bit of discussion decided for August (Aug 2nd) we should examine cutting edge art that exploits new Internet tools: think the programmable web, web 2.0, mashups, and open source APIs.

For this Dorkbot we are soliticiting submissions/nominations for presentations and demos (we'll have a demo table). Priority will be given to artists with their own original art, but we're also taking submissions for folks who want to give “best of the web” overviews.

Sounds like a good opportunity for all those Seattle Web 2.0 startups…

Bad Web Form Design

I'm seeing this pattern appear more and more on consumer web sites that I have to interact with (e.g. my cellular company and the bank) and it's driving me insane.

I talking about input boxes that automatically move focus to the next box when you have entered enough letters and/or digits into them.

The most common use of this pattern is when entering phone numbers.

Assume (for the audience in the US) that there are three boxes - the first two accept three digits and the final box accepts four. Focus initially starts in the first box and then once you have entered three digits (e.g. 425) the focus automatically moves to the next box.

The problem occurs if you enter something incorrectly.

Say I incorrectly entered 426. Focus is now in the second box, and being a keyboard kinda guy I hit Shift-Tab to move to the previous box.

Ping pong!

As there where enough digits in the first box, focus moves forward again to the second box without giving me a chance to correct my mistake. On some sites, if I use the mouse and click back in the first box, the same ping pong problem occurs.

The only way out of the problem seems to be to double-click or drag-select to select all the text in the first box, and then the focus stays there allowing me to reenter everything.

I understand that the point of this pattern is to allow the user to type '4254669305' without having to click in each box, but it seems to me that the website designer just wanted a pretty page. Why not just have a single input box?

Does anyone else have this issue? Or am I just being pendantic :-)

Btw, Here's a website that shows you how to implement this bad pattern and exhibits the shift-tab problem that I describe above, though not the click problem.

An SEO Experiment

Although I'm the #1 “Steve Lacey” on the popular search sites, as far as “Stephen Lacey” goes, in rate a #8 on the front page of Google, and I'm nowhere with the others.

Stephen is my real first name, and my mother hates it when I sign a birthday/christmas/etc… card as “Steve”.

I learnt that the hard way about twenty-five years ago.

Also, I use “Stephen” in any pseudo-legal, registration for conferences, etc… context. So it would be good if people who searched for me, well, found me.

Anyhow, in experiment number one, I've just added “Stephen Lacey” to the meta keywords. Lets see what effect that has. Yes I know I could employ the full SEO techniques, but I thought I'd see what actually works.

That means taking it one step at a time.

Btw, don't you think that the search engines would consider “Steve”, “Steven” and “Stephen” as synonyms? Even if they gave greater weighting to exact matches, treating them as completely seperate semantic entities seems odd.

Sir Tim is Blogging

Sir Tim Berners-Lee, father of the World Wide Web now has a blog. Only one post so far, but what an interesting post it is.

It turns out that the first browser was actually an editor also - he envisaged the web being read/write from day one, it just so happened that everyone got hooked on the read side of things.

So it took until mid-1999 for us to see the original vision when Dave Winer brought us back to our senses with his “Edit This Page” insight.

Holy Cr*p!

Google reports third quarter results

Google reported record revenues of $1.578 billion for the quarter ended September 30, 2005, up 96% compared to the third quarter of 2004…

Ninety. Six. Percent. Warp factor eleven Mr Sulu!

Web 2.02 - Interactive

An excellent post by Nick Bradbury, citing the current Web 2.0 hype and stating that the desktop still has a place. Steve Makofsky also weighs in here, as does Robert Scoble. But at this point, I'd like to disagree with Robert.

For me, streaming HDTV is irrelevant. I've yet to find a good definition of "Web 2.0", but I'm pretty sure that fast, reliable interactivity is a key component, which rules out streaming as an interesting feature.

I'd like to add another rule to Nick's:

Certain classes of applications cannot, by definition, be instantiated as a web service.

Well, for the foreseable future anyhow.

Of course, by this I mean a rather large subset of interactive entertainment. E.g. Games.

Yes, you can download a flash app, but at that point it's a local application. Yes, you can run something from Wild Tangent. Again, at that point it's effectively a local application streaming content.

For the smarts of the application to be remoted we need might higher bandwidth than is currently available today, and more to the point exponentially lower latencies.

I don't know about you, but I like my interactivity to happen at least at 30 hertz.

Memeflow GoTo - Fixes and features

Time for an update on Memeflow GoTo, my little bookmarking web service. It's been a while...

The big new feature is RSS feeds. You've always had an OPML feed of your links, but now you can get that as a full RSS feed, with each block as a separate item, or a changes RSS feed with an item per day detailing all the additions to each block.

The changes RSS feed is pretty nice as you could have a block set up for a particular topic (or, dare I say it, tag) and you and others can subscribe to the feed and see your changes. I'll probably add feeds for the specific blocks too.

The layout of the site has been tweaked too, so that all the informational blocks on the right-hand side are now placed on tabs. Take a look at the changes tab for a list of bug fixes. Most of the bug fixes have been to do with UI problems:

  • Make OPML compliant to v1.1.
  • Fix a bug where clicking another mouse button while left-click dragging a block or item would cause weird things to happen.
  • Fix a bug which could cause multiple edit boxes to be open for edit and out of sync in page.
  • Small pages adjustments. Separated system blocks into tabs.

In the queue for new features we have:

  • RSS feeds for specific blocks.
  • Provide a public facing page, so you can provide other people with read only access to your page.
  • Allowing re-ordering of items within blocks.
  • Import of bookmarks from IE or Firefox.
  • A place to enter comments or tags for links.
  • Some form of collapsible UI for the actions at the end of each link.
  • The ability to collapse blocks.
  • Pages of blocks.

Please let me know of any other ideas or bugs...

Automatic OPML feed

Scoble is looking for a blogging platform that automatically produces an OPML feed based on tagged data.

Well, I just knocked this index template up for MovableType which seems to do the job. It loads fine in Dave's OPML editor too:

<?xml version="1.0" encoding="<$MTPublishCharset$>"?>
<opml version="1.0">
  <dateCreated><$MTDate format_name="rfc822"$></dateCreated>
<outline text="<$MTCategoryLabel$>">
  <outline type="link"
           created="<$MTEntryDate format_name="rfc822"$>" />

Just add it as an index template, and you're golden. This is working in my install of MovableType 3.2 - I expect it will work fine with TypePad.

The autogenerated opml for my blog is now appearing over here.

The Scoble Effect

I was just thinking that I couldn't have written anything interesting recently as I haven't been linked to by Robert Scoble for a while, when low and behold a new post from him shows up in my PubSub feeds. Not only for my Ajax service GoTo, but also for yesterday's podcasting article.

Cue a nice spike in traffic, both for this site and for GoTo. Cool.

The really interesting thing for me was that the big influx in trial users for GoTo showed off an oversite in my recent changes to support lazy registration, aka trial mode. When I added the trial mode, I also added a default set of blocks and links so that new users aren't presented with what amounts to a blank page.

Of course, this meant that the statistics for those sites went through the roof, so I've excluded links from trial users from the "Top Ranked" statistics.

Need help testing edge cases of your code under load? Just figure out a way to get Robert to link to you...

Tracking Clicks

Being a bit of a neophyte (well, maybe I'm a journeyman now) to javascript, I have been tracking the links that users click on in my AJAX homepage site by having the URL for the link be something like http://www.memeflow.com/goto/goto.php?hit=234, which after noting the hit would redirect to the requested page. Rather cryptic and problematic.

The problem is that this url is what shows up in the status bar, not the url for the site that the user is expecting to navigate to. The other problem is that you can't right-click on it to copy the url to the clipboard.

So, along comes various threads on message boards about google tracking clicks by rewriting the url on the 'onmousedown' event. Very sneaky and very clever.

I've told users that I'm tracking the clicks since day one, so this isn't a problem for me. So now, the url for each link is exactly the correct link for the site the user is expecting to visit, but the element has this attached to it for the 'onmousedown' event.

this.nodeItem.onmousedown = function(evnt)


var left = true;

if ((evnt) && evnt.which)


if (evnt.which != 1)

left = false;


else if (event && event.button)


if (event.button != 1)

left = false;


if (left)

this.setAttribute('href', 'goto.php?hit=' + id);


Most of that code is actually checking to make sure the click was from the left button, so that a right mouse-click has the desired behaviour.

I hope someone finds this useful...

Recent Updates To GoTo

Following on from the release of GoTo, my ajax powered home page, I've just released a bunch of updates, including the most requested feature - support for adding a page to GoTo via a button link when at that page. You can now also move links around between blocks, just like you could move blocks around.


  • By dragging the link "Add to GoTo" to your browser links area in Firefox, or right-clicking the link in Internet Explorer and adding it to your favourites, you can automatically add the current web site to your GoTo subscriptions. They will be added to a new block under the "Top Links" section, and you can drag the items to your normal blocks from there.
  • You can now delete a block by clicking on the "Delete block" link.
  • If using Internet Explorer, you can now hit "Enter" in edit boxes to complete the operation.
  • Blocks and items are now sorted by creation time.
  • Fixed issues regarding slashes being displayed after text has been escaped for storage.
  • Editing an item now selects the contents rather than just giving focus.
  • Fixed bug in ranking (links linked to by the same user multiple times now only counted once).
  • Now supports all URI schemes, not just http.
  • Added the ability to reset your password.
  • Deletion of blocks now requires confirmation.
  • You can move items around between blocks by clicking on the "#" icon next to the item

Please send any comments or suggestions to me, or goto@memeflow.com. Cheers!

Mucking Around With AJAX Website Design

Over the past couple of weeks I've been mucking around on and off with a website idea. You can try it out on my "other" site over here.


The basic idea is that the site acts as a very fast homepage. You create blocks that you can move around, and in them you place links. It's all in an AJAX style using XMLHttpRequest stuff, so there's no page reload - everything is edited in-page and hopefully everything happens very quickly.

So there you go. If you give it a try, please drop me a line with your experiences.

I've got a bunch of other features in the pipeline for it, and I hope to keep everything stable.



© 2001 to present, Steve Lacey.