Thursday, December 08, 2005


Today I tried leaving a comment at Yahoo's search blog. I filled out the info fields, entering the address of this blog in the weblog field ( and hit post. I was quite surprised to be greeted by the following screen:

If you look at the line of text starting with "Your comment could not be submitted..." you'll notice that Yahoo is objecting to the fact that my blog is hosted at and I'm told to correct the error and repost. I suppose I could transfer my blog to another company's weblob servers to satisfy Yahoo's request and to have my comment posted there, but I opted against this.

What are the alternatives? Yahoo's Search Blog could institute a slightly more demanding authentication scheme. Currently, a commenter is asked to type the word "Yahoo" in an input field. And they wonder why they're getting spam? How about a captcha or something?

Another alternative is to use Yahoo Mail's spam filtering tools. Anything's better than blacklisting an entire domain!

When these kinds of things happen, I just get depressed about the Internet and technology in general. Can't someone spend a little bit more time thinking things through?

(I ended up posting, but entered my website instead of my blog address instead.)

Saturday, December 03, 2005


Another little experiment:

I've set up a blog that anyone can (anonymously) post to. Why? To see what the YubNub community (and hopefully many others) are thinking about, worrying about, hoping for, love, hate, etc., etc.

This is currently done via YubNub using the tell (or netsend) and look (or netread) commands. For example:

tell I'm listening to some old David Bowie and I'm loving it.

This will (almost) immediately post the "I'm listening..." message to the MultiBlog.


This will forward you to the MultiBlog website to read the posts. If you're not familiar with YubNub, see my YubNub post. (These commands must be typed in to a YubNub command field.) Also, if you subscribe to the RSS feed you can avoid the "look" command altogether.

If anyone wants to help spruce up the site's graphics or template, please contact me.

Have fun!

Cellphone Dreams

Last night I had a strange dream. Someone asked me to do a web search on my cellphone. I slowly went through some tiny screens and finally got to Google. I entered a search term (I wish I could remember what it was) and selected "Search".

Cut to the next scene, where all of a sudden the someone and I were now huddled about some kind of very old micro-computer looking at a screen which seemed to be displaying Gopher or Lynx.

Does this say anything about the future of web access via mobile phones?

I hesitate to jump into either of the two futurist camps. One which says that the PC is dead. And the other which says that the cellphone is the future of web access. I have no doubt that handheld devices will play an integral part in our lives, but I can't see the PC leaving us any time soon. Evolving, yes. But dying? No.

For the future of mobile web access, I prefer to imagine a device made of electronic paper which unfolds/unrolls to the page-size of a standard hardcover book (or larger), which hooks up to the web wirelessly, is touch-sensitive, and is very easy on the eyes, both in terms of design and display quality.

I think that was my first tech dream.

My attic

I think I use the social bookmarking service as much as the average user. "I gotta read that later..." "I definitely gotta check this site out..." "That looks kinda neat..." I bookmark all this stuff, and what do I do with it? Nothing!

Well, that's a mild overstatement. I do occasionally search for a site within my tags, but this typically occurs only when I remember that I had previously bookmarked it. So, this is like putting a bunch of paraphernalia up into the attic, and afterwards, never rummaging around up there. I need someone to climb up to the attic on a regular basis, dust off a few boxes, and present me with something of interest. Why did I put them up there in the first place?

With this in mind, I'm working on a randomizer. Some kind of tool that will alert me in some way about the neat links I've bookmarked there.

Codename: randel

More info once it's close to being done...

Sunday, November 20, 2005

dogear: in-page bookmarking

One of my very early posts was about creating bookmarks within a web document. Most, if not all, of the current bookmarking web applications allow you to bookmark a page, but not a sentence.

As I was bookmarking articles to my account, I became frustrated by the fact that I would probably never get to them, as I was being forced to read them in a linear way. Read one article, then move to the next. I don't work like that, so I decided to something about it.

Ta da! dogear.

After dogear parses through a web page, the reader can click on almost any sentence to bookmark it. The sentence is highlighted and its location is sent to the dogear database.

Now you can take breaks while you read, read out of sequence, survive browser or computer crashes, and bring your readings with you wherever you go.

dogear is in its alpha stages at the moment, but I'm trying to get everything to work properly (some sites just won't go through the dogear engine; others don't display properly), and to add some "social" features: Who else is reading this? What else are people reading who are reading this? I want to review this. Et cetera.

dogear is looking for some alpha users to test the concept and provide feedback.

Contact me at

unblokt: 2nd George W. Bush speech begins

An unblokt user has requested another go at a collaborative George W. Bush speech. The first Bush speech was completed a few weeks ago and can be read here. Some quotes:
"We should put our faith in God and in whatever our leaders say."
"Anyone who disagrees with us is a terrorist."
"The terrorists hate babies. They hate apple pie. They want to destroy the American dream. Not only do they want to destroy it, they want to squash it and mash it up like a bug."

I think Bush's speechwriters must be getting in on the unblokt action too.

Sunday, November 06, 2005

unblokt Volume 2 complete!

Well, another classic has been written. Thanks to all of the authors who contributed on this second full-length novel at unblokt. Volume 2 can be read at the bookshelf. A third book has just been started, and this coincides nicely with National Novel Writing Month. Let's see if we can pump out another 200-pager by the end November, shall we?

Tuesday, November 01, 2005


Driving back to my apartment tonight, I had an exciting thought. (As a math guy, this is exciting to me; probably won't be very exciting to you.)

Google Maps + Mandelbrot Set

You remember the Mandelbrot Set right? Mandelbrot was an IBM researcher working on visualizing iterative functions in the complex plane (I think) and he discovered this incredibly beautiful mathematical world which is now named after him.

Anyhow, it would be so cool (and useless) to see Google implement this. What better interface to roam around this infinitely-detailed bug-shaped world than the zoom and drag interface of Google Maps?

Saturday, October 29, 2005

unblokt vs. Yeats is Dead! (Part 2)

Another very interesting connection between unblokt and Yeats is Dead! is a bizarre reference to a melon-baller. From the aforementioned article, two of the authors discuss a scene:
"I didn't kill anyone," says Marion Keyes, sweetly.

"No," says O'Connor, "you just threatened to gouge someone's eyes out with a melon baller."
In volume one of unblokt, the following:
"The thought of killing her, slowly, with a melonballer, aroused me in the mornings."
"I brandished the melonballer."
"And if you have ever brandished a melonballer, you know how painful it is."
Kind of an interesting coincidence, unless melon ballers play an important role in all novels.

Friday, October 28, 2005

unblokt has many, many authors

I read an article recently about a murder mystery written collaboratively by a group of 15 Irish novelists. The novel is entitled Yeats is Dead! and two of the authors whose names I recognized are Roddy Doyle and Frank McCourt. One of the other authors, Joseph O'Connor, claimed in the article:
"It is the greatest number of authors to have been involved in a collaborative novel. Or at least, to have owned up to their chapters."
Of course, he must have meant in a published collaborative novel.

Why do I say this? Because the first volume of unblokt has at least 500 authors. Of course, some of these authors have written complete and utter drivel (for example: "asdf asdf asdf asdf "), while others have written sentences that I'm sure Yeats himself would have been proud of.

I would love to get the first 300-page volume written at unblokt published somewhere. Then O'Connor's statement would be decimated by a book with 30 times as many contributors. Not the most flowing prose, but certainly the most varied.

If you know (or are) a publisher interested in being part of this truly monumental human achievement (ha ha), please get in touch.

Friday, October 21, 2005

socialight II

Some corrections to my previous post. Through socialight's website, users can specify latitudes and longitudes, click a Google map, as well as pick the size of a shadow, although there are only 3 sizes: exact, general, and city.

In any case, socialight (or a quicker competitor) is going to be big.

socialight: drop shadows everywhere

I'm quite excited about socialight. I have recently posted a few comments on geo-tagging, and it seems like soon I'll actually be able to do it.

socialight lets you leave "shadows" at specific locations by tagging a location with a message or a photograph using your cellphone. Shadows can be made public, private, or accessible by a select group of people. Their website gives the following examples as to how the service can be used:
+ A woman shows all her close friends the tree under which she had her first kiss.
+ An entire neighborhood gets together and documents all the unwanted litter they find in an effort to share ownership of a community problem.
+ I leave a note for my friends at the pub to let them know I'm running late.
+ A food-lover uses Socialight to share her thoughts on the amazing vanilla milkshakes at a new restaurant.
+ The neighborhood historian creates her own walking tour for others to follow.
+ A group of friends create their own scavenger hunt.
+ A tourist takes place-based notes about stores in the shopping district of a foreign city, only for himself and his Contacts, for a future time when he or his friends are in that city.
+ A small business places StickyShadows that its customers would be interested in finding.
+ A band promotes an upcoming show by leaving a StickyShadow outside the venue.
Reading the FAQ at socialight's website, I was struck by a few things:

1. The service is available on only one model of phone and on only one company's network. I realize it's sometimes good to release a product early, but only one phone?! What's the current market size?

2. Only addresses and intersections are used for tagging. Can I set a fuzzy area? How focused is the shadow? What about GPS? I guess most phones don't have this technology yet, but it would be so much simpler to not have to enter any addresses at all; simply type your message and socialight tags your current location.

3. There was no mention of positional alarms in their examples, but these are equivalent to personal shadows. The next time I'm at my favourite coffee shop, I want to remember to try the new lychee latte, so I set a shadow for myself.

If socialight can roll out quickly, I think this is going to be a killer application. I hope that the word is spread, that all of the Motorola i860 users on Nextel's network start using it, and that all other cellphone owners start clamouring for access.

eFrames II

I was in the process of reading Cory Doctorow's cool new novel Someone Comes to Town, Someone Leaves Town via RSS feed, when I was struck by an idea (not original of course) after reading this paragraph:
Alan's father was a mountain, and his mother was a washing machine -- he kept a roof over their heads and she kept their clothes clean. His brothers were: a dead man, a trio of nesting dolls, a fortune teller, and an island. He only had two or three family portraits, but he treasured them, even if outsiders who saw them often mistook them for landscapes. There was one where his family stood on his father's slopes, Mom out in the open for a rare exception, a long tail of extension cords snaking away from her to the cave and the diesel generator's three-prong outlet. He hung it over the mantel, using two hooks and a level to make sure that it came out perfectly even.

So the idea is this. Frames should be self-levelling. Well, eFrames should be, at the very least. Of course, using little bubbles to determine levelness isn't all that "e", but maybe some kind of sensor could be used instead. Forget scratching up your wall with the big aluminum level from your garage; simply flip open a little tab at the front of the picture frame, and adjust its position accordingly.

I couldn't find any pictures of this type of frame, but I read something about mounting brackets with levels built-in (not very instructive link here). However, I'd prefer to directly monitor the level of my picture frames, through the frame itself.

Sunday, October 16, 2005


I recently set up AdSense on this blog, but was disappointed to find that Firefox couldn't sense that it should display them. Blogger trumpets Firefox all over its help pages, so I was quite surprised to find ads being displayed in IE but not in FF.

To add to the frustration (and this happened in both browsers), I couldn't report the problem. Each time I tried to submit the issue, I got the following error:

(The first two blurred areas were actually series of 8 numbers. I've removed them in case they reveal anything about my account.)

When I clicked the back button, I noticed that indeed my email address was now prefixed be this U/B string. Blogger was using Javascript to add this text, then reporting it as an error.

I'm not exactly sure what to do now. Hope that more users choose Internet Explorer? Or that someone at Blogger reads my blog? The latter is unlikely, and the former is undesirable.

Saturday, October 15, 2005

Positional Alarms

I was eating at one of my favourite fast-food curry (カレー) restaurants in Japan recently (Coco Curry) and after ordering, a friend reminded me that on a previous visit I had said that next time I would order such-and-such. Well, the next time had just arrived and I had forgotten to order the meal I had been so excited about last time.

It occurred to me then, that what I needed was a positional alarm. Currently, all of my alarms are time-based: pay this bill by this date; play squash on that date; etc. The technology for positional alarms is becoming more and more commonplace. GPS receivers are found in cellphones, PDAs, MP3 players, and cars, to name a few. (Other technologies exist as well: beacons, etc.)

Regardless of how it's done, I want to be able to set the following geo-alarm on my cellphone: "Order the pork cutlet curry." The name of the restaurant should be optional, depending on the fuzziness of my alarm. If I create a "tight" alarm, it will go off when I'm back in the restaurant. If I create a fuzzy one the alarm would go off when I'm driving near the restaurant allowing me to decide whether I want to eat some pork curry or not. In this scenario, the "snooze" button delays the alarm until the next time I'm in the area. (Snooze could allow for alarm tightening if I drive by the restaurant everyday and the alarm becomes annoying.)

This is a trivial example of a positional alarm, and makes no use of the web, or of social networks. If I can manage and share my personal geo-alarms on the web, and import or subscribe to other users' alarms, then things get more interesting. I could subscribe to alarms for "pasta Toronto" and get buzzed around dinner time when I'm in the proximity of a great pasta restaurant. I could subscribe to "art global" and get notified about art galleries and museums within walking distance wherever I am on Earth. (These examples are awfully close in concept to my previous post on Geo RSS.)

I see a lot of utility simply in the ability to set personal positional alarms. The cooler social applications would present themselves once this service was more widespread.

Friday, October 14, 2005

e-paper recycled

I've written a few posts which touch on e-paper. In none of them did I talk about the magazine, book, or newspaper industry switching from regular paper to e-paper to add new-fangled features to their publications. Why? Because I don't think they should be printing on anything any longer; on a mass-scale that is.

However, when I read about e-paper, people always seem to be talking about printing newspapers and magazines on this new medium. I suppose the established paper companies are concerned about going the way of the dodo, and are locking away certain ideas. Even Siemens, in the blurb accompanying their new wafer-thin display product, fails to mention any breathtaking applications.

The vision should be this: I buy a "container" for my content. The container is an e-paper product, designed to look like a book, a magazine, or a newspaper. Content is wired or beamed into the container via some mechanism from the publisher. When I buy content, I own it forever, and can reload it at any time.

Wired could sell a sturdy e-paper container for its issues. It has the same dimensions, the same look-and-feel of a Wired magazine, yet it can be uploaded with the current issue, any back-issue, a composite of issues based on a tag search, etc. For the amount of money they'd save on printing costs, they could give it away.

Each publishing company could sell these containers. I wouldn't mind owning a ton of 'em: one for Wired; one for your standard-sized magazine; one for the tiny Reader's Digest-sized magazines; a paperback novel-sized one; a newspaper-sized one; a comic book-sized one; the list goes on, and on, and on...

Transcribing Books

Google's and the Archive's and all of the other groups' attempts to scan a bunch of books is a wonderful idea.

For those of us who feel left out of these Alexandrian projects, there are a few options: volunteer to help Project Gutenberg (or other similar group), start emailing scans of your books to Google (ha ha), or do something new.

Gather a trusted group of people who own a particular edition of a non-digitized book. Set up a writing project at one of the new collaborative wikis or word processor web apps or create your own. Take turns entering sentences according to some well-defined rules. If the writing group is large enough, the endeavour could be completed in a not unreasonable amount of time.

A perfect group of people for this project would be a gaggle of library workers. However, it's an odd input mechanism and would need some kind of pay-off to catch on. Any ideas?

Oh sir... it's only wafer thin.

Siemens last week, at the Plastics Electronics trade fair in Frankfurt, exhibited wafer-thin colour displays.

Flexible, colour, relatively cheap. I want some. And an engineering degree.

Perfect for the eMag.

Web 2.0 in the Wide World

A common thread in my recent thinking about the web, is the extraction of all this wonderful information (shared pics, blogs, bookmarks, etc.) out of the web and into the real world (see my eMag and eFrame post).

It's great that web users have all these cool tools to share information and collaborate and tag and blog and feed and... But all (or most) of this is done from the keyboard. I constantly get chastised for the amount of time I spend at this machine. I think I'd prefer accessing the information of Web 2.0 from my armchair or bathtub.

And I don't mean accessing the web from a tiny cellphone or PDA. I want my eyes when I'm older. I mean: embed hooks to the web into the physical world.

Some cool products I can't wait for:

*a book, normal in appearance, has in fact just contacted Project Gutenberg, and has had its pages filled with a literary classic I should be reading.

*a magazine on my coffee table contains news, blogs, and images, which have been automatically uploaded from the web.

*a to-do list on my fridge updates itself with entries I've made at work, from my cell, or that my family members have made (if the list is shared, that is.)

Of course, if people could teach themselves how to engineer things (as they can easily teach themselves to create web apps), these products would have arrived yesterday.

Here's hoping they arrive tomorrow, or the day after.

Thursday, October 13, 2005

Google Bookmarks

A short 10 days after my Google-icious post, Google has implemented the idea, with nary a nod in my direction. Isn't 10 days just enough time for someone to read my post, organize a project team, develop the feature, test and debug it, and release it to the public?

In any case, it's clear that someone at Google thinks that merging bookmarking with search is a natural idea. The implementation is not perfect (bookmarking is buried within Google's Search History which I've never actually used until today), but the feature's presence is a good sign.

* float bookmarks up to the main results page (possibly the plan after beta)
* default the Labels field to the search terms used in finding the item
* just call them Tags already!
* not sure if I like having to click twice to add "Labels"
* add "remove" links for both bookmarked and unbookmarked history results

Tuesday, October 11, 2005

Publishing "unblokt: volume 1"

I've begun an attempt to have the first volume collectively written at unblokt published. I've placed a "Make a Donation" button (PayPal) at the bottom of the unblokt site to try and raise about US$500 to publish it at Xlibris.

If you thought writing at unblokt was neat, imagine ordering a paperback copy from Amazon and reading this masterpiece in the subway, on the couch, or as you go to bed.

Help out by clicking here:

Monday, October 10, 2005


I had a wonderful idea about a year ago, but like most of my wonderful ideas, someone had already thought of it and had already implemented it. The idea was eFrames: wireless picture frames which receive images broadcast from a PC or other device.

I thought I'd hit the killer product with this one, but when I found the current offerings out there, I was disappointed by the lack of competition and the high price. (Flickr or Slide needs to partner with someone to get this product on everyone's mind.)

Web photo sites and slideshows and loops are great, but extracting these into your physical home setting would be wonderful. I would love to see the family picture frames on my grandmother's wall cycle from one photo of me to another (only the frame that is tagged with my name of course.) She might think her house was haunted, but it'd be cool nonetheless.

Frames tagged with one or more keywords would give them certain themes: family, nature, animals, Japan, random, etc. Pictures could be beamed from any digital photo device as well as from web-connected devices. Personal or public web photos or images based on tags could be grabbed from Flickr, from your local drive, your website, or your Slide or Loop album. Custom RSS image feeds would also be a neat feature.

This is the future of framed images. Once ePaper and/or eInk have come down in price and can display a full range of colours, there will be many more very cool products. Imagine a large eFrame perched on your wall, cycling through prints of remarkable art from all periods of history from cave paintings to Michelangelo to Munch.

Friday, October 07, 2005

RSS Messaging

A few days ago, I wanted to send a quick thank you to a friend. I opened up a new Firefox tab, clicked my email shortcut link, logged in, clicked compose, typed my friend's email address, entered "thank you" as the subject, a very brief thank you message in the body, and clicked "Send".

What I really wanted to do was type the following in my YubNub command field:

rss jsmith2005 hey thx for the tip! -sean

rss (a not-yet-created YubNub command) communicates with an RSS message server (not yet created, as far as I know) and posts my short message to my friend's personal RSS message feed. This example uses YubNub (because it's so sweet and simple to integrate with an exposed API) but other standard methods would work too (web app, desktop app, wireless app, etc.)

John Smith could set up his personal message feed to accept anonymous messages like the one above (my name is simply part of the message text, not a true identifier,) or could require some form of sender authentication. This would be done on the sender/server side. Jane Smith would choose a feed name which she could make public or tell only to her friends and associates.

Personal RSS feeds might be easy to abuse. A bot could flood a public message feed with spam quite easily. If John Smith so chooses, he could require a form of password to send him RSS messages:

rss jsmith2005:kyoto hey thx for the tip! -sean

If the password is missing, the message would be dropped or forwarded to Jane Smith's spam message feed.

Why waste an email on "hey thx for the tip!" I certainly don't want to.

Thursday, October 06, 2005


The previous post talked about marrying any Web 2.0 service which provides directional information to your car. Now, ignore the "appliance's" form (whether it be a car, cellphone, glasses, etc.) and consider the data.

Locations (addresses, intersections, lat/longs) could be tagged with keywords just like a website is tagged using Shops could tag themselves with the products they offer, current sales, job postings.

As you walk, drive, or hop through the streets of a city, your "appliance" automatically subscribes (and unsubscribes) to data feeds based on your current position. If a feed item jives with your interests or needs, you are alerted (text message, email, cellphone call, etc.) Of course, these alerts can be micro-managed to allow for complete privacy, full-alert-onslaught, or anywhere in between.

Notice I used the word "shops" above. Here, I think you should be thinking "long tail". What shouldn't happen is huge department stores dumping their entire inventory as tags into the system. Tags should be limited in quantity and used to broadcast interesting products, sales, and shop specialties. This allows small, eclectic, out-of-the way shops to advertise their wares to consumers who are interested.

If the appliance receiving the feeds is your car and its windshield, then more cool things happen (see previous post). The magic green line leads you to a shop selling an alligator skin sofa you've been looking for. A coloured overlay causes a bookshop to "blink" a few times, with a virtual "Help Wanted" sign above it, because you're looking for a job.

Latitute-longitude pairs could also be tagged by anyone. Using my cellphone, I could tag any position that I'm at with various keywords for my group of friends or to share with the public. I could tag all of my favourite hang-outs in the city for my friends or others with similar interests to check out. You could have your cellphone autotag your geoposition at regular intervals and have a "diary" of your whereabouts during the day. Lost and found items could be tagged. Places could be tagged with historical or cultural information ("Glenn Gould used to eat breakfast here," "A battle occurred here", etc.)

The possibilities are truly endless.

Tuesday, October 04, 2005

Google Maps + Car Windshield Mashup

Admittedly, this idea needs a bit more than Google Maps, but it definitely needs your car's windshield.

Any web service today which offers any kind of geo-spatial data becomes useless if you can't get "there" easily. Sure, my cellphone can probably show me a tiny map with directions to my destination, but a hands-free method would be much better. And much safer when I'm driving.

Scenario: A wireless device in your car is sent a message from your favourite geo-app. This device illuminates the interior surface of your windshield with a street overlay. From your perspective it looks like your route has been painted onto the street with your favourite colour. Choose a green line, a red dashed line, a thin yellow-bricked line, and follow it, left, right, and straight to your destination.

Safety issues? Sure, there are probably many. But I'd rather be looking straight ahead at a cool transparent green line, then constantly referring to a small unreadable map terminal to my right (or left, if you're in Japan, England, and all those other cool countries.) These issues can surely be dealt with.

This could be extended to (and I think some engineering groups have already done work in this direction; again, I wish I had become an engineer!) glasses, sunglasses, contacts, motorcycle helmets, planes, trains, etc.

Where's the Web 2.0 element? Coming in the next post. Stay tuned.

Sunday, October 02, 2005

Coffee Table RSS Reading

If only I had become an engineer. Then I could manufacture the eMag. This idea has been floating around since before the advent of electronic ink and paper, and flexible displays. Should these not be ubiquitously present in every household, on every coffee table, on every night-stand, and on every bathroom floor right now?

Bind some high-quality ePaper together, with a tiny wireless receiver (and other relevant electronics and software) in the spine. Done.

RSS feeds are wired from PC to eMag. Flikr pics and headlines are beamed to the cover. A table of contents is created on the fly, and articles are loaded into the remaining pages. Simple.

The web is "infinite". Every minute, a new blog or news item is made available. This can become overwhelming to even the most seasoned web traveller. However, limiting access to information in the digital world is unthinkable. In the physical world, though, it's a necessity. The eMag is a finite reading object. Once you turn the last page, you're finished. Of course, it can be reloaded with a new cover image and new articles at any time. But, the reader feels like a solid block of information has been consumed, rather than juggling numerous meandering streams of non-stop data.

A message to RSS aggregators: focus heavily on data manipulation, tools and APIs, and not so heavily on your web interface. It will soon be replaced.

(magazine cover thanks to the cool tools at


Google is personalizing the Google experience: personalized search, personalized home page, blogging, etc. How long before they jump into personalized bookmarking?

I think Google could become a formidable opponent to and other social bookmarking services (or more aptly, it could become a formidable assassin) by adding a tiny button or checkbox beside each of its search result items.

Using as an example, each of its bookmarks' four data fields could be filled automatically: 1) URL, 2) description, 3) extended (filled by the short snippet of text Google displays for each search result), and 4) tags (defaulting to the search terms you used to google your target).

I click the checkbox, Google ajaxes a database or sets a cookie. A thin, unobtrusive tag field pops below the search result with the default tag(s) present, allowing the user to add or modify these tags. As I type these changes, Google again does a bit of ajaxing and/or cookie-setting (a kind of reverse Google Suggest - updating a database rather than selecting from one.)

Whereas and company are slowly building enormous databases of links and tags, Google already has the mother database sitting on its servers. All they need to do is create a new table, containing a user_id field, a URI_id field, and a tag field (user johnsmith has bookmarked with tags "common names").

Of course, it may not be as simple as this, but it doesn't appear all that difficult.

Google, on your bookmarks, get tagged, GO!

Saturday, October 01, 2005


Whereas many of the new web 2.0 applications are rehashing desktop ones (although this is not meant to disrespect them in any way), YubNub is wonderfully fresh and new, yet at the same time harkens back to the early days of the black screen and white flashing cursor of the command prompt.

Many will claim "Oh, it's just like the such-and-such toolbar for the whatzit browser", but in fact it's not. It's a browser- and operating system-independent interface to anything and everything on the web. A site where anyone can create simple or powerful commands to squeeze and extract information from the web.

For example, typing "gg giant squid" performs a Google search for giant squids. Want to search for images instead? "yim giant squid" performs a Yahoo! search for giant squid images. Want to buy a giant squid from Amazon? Just type "amzn giant squid" (no promises that your purchase will be successful.)

But these simple examples are just the tip of the iceberg. You can combine multiple commands and use some simple programming control structures to create incredibly useful and powerful commands. Or incredibly silly ones, it's up to you. YubNub is a social command line, which means that anyone can create commands, and anyone can use anyone else's commands.

Here's an example of a more complex command (which I have yet to create, but will soon.) At movie rental shops, I often see an unfamiliar movie on the rack which looks interesting. However, I've rented rotten interesting-looking movies in the past. I can use my cellphone to access IMDb and browse to the movie's page and check out its rating - but this would use up a lot of bytes, hence money. I'm only interested in one number - the rating (or maybe two, the rating and the number of people who have voted.) In any case, here's what I can do. Create a command which searches for a movie on IMDb and "scrapes" its IMDb title id number. Create another command which goes directly to a movie's page by referencing this id number. Create a third command which scrapes the rating from a movie's main page. By linking these three commands together I create one command (which I could call "imdbr") which takes as input the title of a movie, and which outputs a number, the movie's IMDb rating. At the rental shop, I access on my cellphone, type "imdbr Beany and Cecil Meet Billy the Squid" (yes, this is a real movie title) and in a few moments I see (if I'm lucky) a number greater than 7, in which case I head to the counter.

In just a short time, YubNub has grown incredibly quickly, and has the potential to turn into something enormous. I use it many, many times, every single day. I've actually stopped thinking about it. It's become like breathing.

What's more incredible is that YubNub is the result of my friend Jonathan Aquino's entry into a 24-hour programming contest. YubNub fared very well and was awarded second place. Soon after, heads began to turn in the tech world, and Jonathan was promptly invited to Foo Camp.

YubNub - no longer just an Ewok exclamation.

Friday, September 30, 2005

Better Bookmarking

One of my next projects (unless someone tells me that this already exists) is to create a better bookmarking app, one which remembers exactly where I am within an article, blog, or other such web text that I'm in the process of reading.

In the real world, written words are typically confined to a physical form with finite dimensions. If I dog-ear a page in a book, or invert a magazine to the article I'm reading, it's quite easy to pick up where I left off. I can also leave permanent markings on the page to make this easier. On the web, this is a bit more difficult.

Bookmarking a page to a website like or to your Favourites menu in your browser is more like tossing a book or magazine onto the coffee table. You really don't know what page you're on. All you know is that you've been reading it, or want to read it someday.

Occasionally, a web text is quite long, and stretches way below the boundary of your browser's window. If I'm reading such a text (or any text for that matter) and bookmark it, there's no easy way to recall where I was at the next reading. Sure, I can leave the page open in my browser along with umpteen other articles and blogs, but this can backfire (someone presses PageUp or Home; someone closes your browser; your computer dies.)

So what's the solution? A few things come to mind:

1. Browsers should make bookmarking better.
2. should make bookmarking better.
3. A new service needs to be created.
4. I should stop reading things on the web.

It should be a fairly simple matter for browsers to improve their bookmarking functionality to include "geo-positional" bookmarks within a web document. I click a bookmark button, I highlight a word or sentence, I click the same (or another) bookmark button, and voila. The browser can use HTML tags, or surrounding text as context clues.

I'm not going to directly give any ideas to improve their service. I'm sure they have plenty of brilliant minds working on this problem. Of course, they could simply read the next two paragraphs.

What kind of new service is needed? Perhaps the simplest idea is a browser-based plugin or extension, like a Greasemonkey script. The script creates hidden links for each word or sentence, and when the user clicks one, a bookmark is created either locally, or much better, at some centralized bookmarking site. This is a nice idea although I don't like having to install new things into my browser, nor do I like the fact that this is browser-based and would have to be installed at each and every computer I intend to use.

Better, but more difficult to implement, would be a new service which allows you to read a website through another website's bookmarking interface. For example, I go to my favourite social bookmarking site (which has the new service), and I click on an article that I've started to read. The bookmarking site loads the article into some kind of frame, parsing all of the text and adding hidden links to it, and leaving only the smallest vestiges of the bookmarking site in view (a tiny control menu, for example.) The article is advanced to a highlighted sentence, the one I had previously bookmarked. I continue reading, but am interrupted, so click the current sentence to bookmark it. The browser sends a message (AJAX) to the bookmarking site, updating the bookmark for this article, and the new sentence gets highlighted. This idea can be broken, however, by sites which kill surrounding frame elements. (There should be a way around this.)

Is such a product available? If yes, please comment. If not, I'll have to start working on one soon. (Or implement idea number 4.)

Thursday, September 29, 2005

Collaborative Writing: unblokt

unblokt is my first attempt at social software. To get an idea as to how unblokt works, consider the 200-page novel that is currently being written. A visitor to the site is presented with a random sentence from the novel and the one following it. The task is to write a sentence between these two. However, these 3 sentences are not "locked" together. At random, any of them may be presented to other users, who may then stick more and more sentences between them. Kind of like a "Big Bang" for writing. I am informed by Matt Obert of the Verbatim writing group that the writing method implemented by unblokt is in fact quite old, and is termed larding.

At this time, random people on the web have written

a 300-page novel,
a "Dear John" letter,
a "George W. Bush" speech,
a "Letter from Camp",
about 100 pages of a second novel,
and a few scattered sentences in a number of other genres

These can be read on the unblokt bookshelf. A word of warning: some authors have chosen to write some fairly racy sentences, so be prepared. I hope to soon make unblokt both social AND democratic by allowing the community to vote offensive sentences or spam "off the island", so to speak.