A clever way of categorizing! New Scientist: "A new system that can caption your digital photos by listening to you and your friends chat about them is being developed by Hewlett-Packard in California.

Digital photography is booming, and people are storing ever greater volumes of photos on the hard drives of their PCs. The trouble is that people rarely label their photos.

"This is the weak link for digital photo collections," says Margaret Fleck at HP's lab in Palo Alto. "In 10 years' time, finding something amongst them will be very difficult."
Fleck's answer is to tap into the wealth of information in the conversations we have when we talk about our photos with friends. She says the stories we tell do not merely describe the photo, but also talk about the events that happened before and after the picture was taken.
[...]
To harness this information, Fleck has developed software that records these conversations to hard disc, converts the speech to text using a speech-recognition program, and then extracts keywords with which the photos are captioned and indexed."

Sounds like a mistake though: the problem with our digital pictures is not finding them, it's sharing them and the stories around them. (This is a new interest of mine.) Recording those stories to then throw them away just to extract some keywords is a waste.

Gmail Tested, Reviewed

This is the kind of fresh thinking that puts a smile on my face.

Situated software - megnut.com: "This gets to something I've been thinking about for sometime now, the possibility of using personas to represent groups rather than individuals. In fact, I even proposed it as a talk for the last O'Reilly Emerging Technology Conference but it wasn't accepted.

I'm still tickled by this idea of modeling the groups, because as Clay writes, there's a power in groups that you don't find when the same individuals operate in isolation. By creating group personas (groupas? grouponas?), perhaps we could better design and hone our software to utilize the group's power."

InfoWorld: Toshiba, Sandisk develop 4Gb flash memory: "The new chip can store 4Gb (512KB) of data and the companies said they are also working on a second new chip that contains two of the 4Gb chips inside a single case for what is effectively an 8Gb (1GB) flash memory chip."

Is that "4Gb (512 KB)" just an editing mistake, or does it mean something?

WorldChanging: Another World Is Here: How to Build a Coral Reef: "This group has come up with a process for building new coral reefs. It involves wire-mesh frameworks, low-voltage currents creating mineral accretion, and the transplant of broken pieces of coral."

OK, instead of me posting every single story Worldchanging runs, why don't you add them to your newsreader. I don't really post every story, so if you like this stuff check them out.

BBC NEWS | Technology | Radio mail links Pacific islands: "A internet connection speed of 2Kbps may not sound like much, but it is providing a lifeline for the people of the Solomon Islands.
[...]
But for the past four years, the People's First Network has tried to mend fences by using high frequency radio to send and receive e-mail.
[...]
The two main ways of getting in touch with people are short-wave radios or satellite telephones. But radio offers no privacy, whereas satellite phones are too expensive for most to use regularly."

Not to speak of the other advantage of email over radio/phone: asynchronicity (sender and receiver don't need to be there at the same time).

"At the moment that are 14 e-mail stations in schools or clinics in rural areas. The stations are owned by the community, with decisions taken by a committee of village chiefs and religious leaders.

The stations are a pretty basic affair, consisting of an ageing laptop, radio and modem. The kit is powered by a car battery, which itself runs off a solar panel as in most areas there is no electricity. The total cost of the equipment runs to around $8,000."

You'd think that can be done cheaper. Then again, cheap supplies aren't as plentiful everywhere in the world as they are in the USA.

The project uses WaveMail: "WaveMail is optimized for use over relatively slow links like radio-links (Pactor / Packet radio), Inmarsat and telephone lines. In order to use the slow links in an efficient manner, all messages are compressed and a very efficient protocol is used for message transfer (compression and decompression is done automatically without user intervention)."

Next they'll try to provide internet access. See World's poor to get own search engine: "Researchers at the Massachusetts Institute of Technology (MIT) are developing a search engine designed for people with a slow net connection. Someone using the software would e-mail a query to a central server in Boston. The program would search the net, choose the most suitable webpages, compress them and e-mail the results a day later.
[...]
The thinking behind the TEK search engine is that people in poor countries are short of money but have time on their hands, whereas people in the West are cash-rich but time-poor."

Seems like solid thinking.

Percepnet: "Thus, while we would sense lights, sounds, tastes, smells and touches, we would perceive them as, say, a chocolate ice-cream or a glass of lightly chilled Burgundy.
[...]
The research that is being conducted in developmental neurobiology, neurophysiology and cognitive neuropsychology is suggesting that the processes involved in perception are so intermingled that there is little value in trying to divide them up neatly into sensation and perception."

Learn Dutch. Nice site.

(via Phil Bradley) Inside the Searcher's Mind: It's a Jungle in Here!: "There's No Such Thing as an Average Searcher." A study by a marketing agency, and therefore suspect, but still worth a read.

"For instance, women tended to scan all organic [he means non-sponsored] results and read titles and descriptions more carefully than men."

Mmm...

Other conclusions are about banner blindness ("It seems that people have already mentally divided the results page from their favorite engine into sections"), search behaviour ("A person searches, chooses a result, visits a site, and hopefully converts. In reality, we see the typical search pattern is quite different.").

Nothing you don't know and haven't seen analyzed better.

They do have an interesting section about what JaredSpool calls "the seducible moment". In short (and this is my interpretation of their results): don't pitch products during search, instead help users refine their search towards your product. Again, this is my interpretation - the article talks about the "Anonymity threshold". ("If given the choice between getting information and remaining anonymous and getting the information through registering, people will always choose the former.")

Nagging seems to work: Alex has started writing a Topic maps tutorial entitled "How to sort your CD collection with Topic Maps", following up on my post Why is it so hard to learn topicmaps.

GMail screenshot

NewsForge | Builder.com outsourcing content production to India: "Builder.com, CNET Networks' site aimed at application developers, plans to begin offshoring authoring of many of its articles to India shortly."

urlgreyhot : $10 to whomever figures out my anonymous user page not found problems.

I've started to feel that what were the hottest contenders previously, MovableType and Drupal are actually failing on the fundamentals.

So it's not just me.

Drupal is a great CMS product, and has a wonderful developer community, but something bothers me. The "remember me" login feature was broken in version 4.3, and now it's been removed from 4.4.

A question on the mailing list got me this answer:

"The reason "remember me" was removed is not because the remembering is broken per se, the problem is that it is not possible to choose the
behaviour anymore due to PHP overriding Drupal's cookie. Whether or not you are remembered depends on the session persistence and lifetime settings in php.ini / htaccess. Check PHP's documentation if you are not getting the behaviour you want."

Developer hogwash! A community CMS should not be released with such a central feature broken.

Escapable Logic: "You can get a lot done by filling out government forms, but it's a lot like writing code.
One of my projects involved 120 acres on the Denver-Boulder turnpike, but without access. All it took to increase the value of our land 20-fold was to get four layers of bureaucracy, including the Federal Highway commission, to authorize us to build the interchange by adding an assessment to our land and other interested parcels." (Via Doc Searls)

Wires - A girl, a toolkit, Iraq: "Tonight as I write this, a group of nervous wiremen are sitting at Heathrow airport, waiting to board the flight which will eventually lead them to Baghdad, to install the cameras in the studio there."

(Via Simon) Topix.net Weblog: The Secret Source of Google's Power. That's why I like this blogging thing. Thoughts like this were surely going through the (sub)conscious of thousands of people after the GoogleMail announcement. Then someone articulates it well, and we all get to read it. It gells. We all learn. Nice.

GoogleMail no Joke: "It's not like Internet search service Google can't laugh at itself, but when an April Fool's joke got out of hand on Thursday, a real business plan was rumored to be a Web hoax -- and that was no laughing matter."

03/22/04: Pot-in-Pot" href="http://hinterlands.cc/index.php?showtopic=30">Hinterlands.cc -> 03/22/04: Pot-in-Pot:

"This is Mohammed Bah Abba's Pot-in-pot invention. In northern Nigeria, where Mohammed is from, over 90% of the villages have no electricity. His invention, which he won a Rolex Award for (and $100,000), is a refrigerator than runs without electricity.

Here's how it works. You take a smaller pot and put it inside a larger pot. Fill the space in between them with wet sand, and cover the top with a wet cloth. When the water evaporates, it pulls the heat out with it, making the inside cold. It's a natural, cheap, easy-to-make refrigerator.

So, instead of perishable foods rotting after only three days, they can last up to three weeks. Obviously, this has the potential to change their lives. And it already has -- there are more girls attending school, for example, as their families no longer need them to sell food in the market."

Alex writes a rebuttal to my article on Why is it so hard to learn topicmaps?

Alex: "We are hard-working people who write specialised and clever applications, and so we don't really have the time to write "how to sort your CD collection with Topic Maps.
[...]
So, in Peter we trust, and even though I've seen most of the responses (I added some myself) he still says there are no hackable applications that you could use for easily learning about Topic maps. How does he come to that conclusion? I surely don't agree, but it is hard to fight this one back as I don't have the readership he's got."

Um, Alex, you may be overestimating my readership there :) And hopefully this link will help the people who read my blog also find your post (and your excellent blog).

The point of the post, and I stand strongly behind that, is that we need better tutorials, for the reasons I outlined.

I am a big believer in topicmaps, but as it stands, it's almost impossible to learn topicmaps without a large amount of developer background. I believe it doesn't have to be so hard. I believe you don't need to know your data model from your reification in order to get going with topicmaps. A basic understanding of XML should be all that's required. And I believe that providing tutorials like the ones I was attempting to describe is the best way to get topicmaps understood and adopted.

There are clear advantages to using topicmaps technology (for example, you can change the relationships in your data without changing your data model) that we should be able to illustrate with a basic tutorial. Developers generally learn by coding - we should work with that.

A text adenture game based on RDF. Cool. (via RDFPlanet)

SWAD-Europe Weblog: FAQ: How do I parse RDF? A list of RDF parser libraries for most programming languages.

Why is it so hard to lean topicmaps?

Some ideas on why many people hit a wall when trying to learn topicmaps.

Topicmaps are a powerful standard for managing ontology-like metadata. It's kinda hard to get your head around topicmaps, but this is not because they are especially complex or hard to grasp (Not harder than object orientation, relational databases or XML for example). They're really not that bad. It's because of the lack of easy to understand real world applications that you could hack together in order to learn topicmaps data structures and concepts.

When you learn about databases, you'll start by programming a simple script to organize your CD's or something.

Manage your contacts, organize your income - there are thousands of beginners tutorials like this for almost any programming language available. After you've wrapped your head around the basic concepts (relations and ID's), you can then learn about more advanced ideas like normalization or joins.

If you had to read about the conceptual underpinnings of databases in order to get your head around them, they would never have taken off the way they did.

There are no hackable applications that I know of you can use to learn topicmaps. This means that there are no beginners tutorials that let you hack a useful application together based on topicmaps. All the tutorials I've seen teach you topicmap concepts and how to create a topicmap. They don't help you create an application to organize your CD's.

In a recent post to the topicmaps list, I asked to come up with something that was easily hackable, useful and would help people understand topicmaps. There were some ideas, but we couldn't come up with an acceptable example. (This is not to say that it can't be done. We should think about what exactly is missing here.)

So when you're interested in topicmaps, you read the specs or one of these tutorials, and unless you're familiar with data models and a bunch of advanced metadata concepts (reification anyone?), it's gonna blow your head. Many people give up right there.

The distance between the topicmap model and a real life application seems larger than the distance between the relational database model and a real life application.

Similarly:

Topicmap tools are at the topicmap level, not at the real-life usefulness level. In other words, topicmap tools let you manage, create, and merge topicmaps. But they don't let you do anything specifically useful outside of the topicmap realm (like create a simple application for managing your CD's, like Access lets you do).

The level of abstraction of topicmaps is higher (which provides great Power and Flexibility) than that of databases. But that means that, to get from a topicmap to a useful application, you need more stuff in between. RDF seems to be resolving this problem: the stuff in between are useful standards like FOAF. Topicmaps need standards like that, that let you do something useful with topicmaps, and thus make the tools useful and topicmaps understandable.

Shirky at his best describes patterns that seem to fly against common sense. Situated Software: "The curious thing to me about Teachers on the Run was that it worked where the Web School version failed. RateMyProfessors.com has been available for years, with a feature set that put the simplistic write/read/vote capabilities of Teachers on the Run to shame. Yet no one at ITP had ever bothered to use RateMyProfessors.com, though the weekend's orgy of rating and voting demonstrated untapped demand."

Jon Udell on Macromedia Flex: a developers environment for developing applications in Flash. Competing with XUL (open source) and XAML (microsoft).

So Apple is patenting a hierarchical menu on the iPod? United States Patent Application: 0040055446: In a portable multimedia device, a method, apparatus, and system for providing user supplied configuration data are described."

Someone seems to have messed up, the link to the American express Seinfeld/Superman ad shown on TV doesn't work: HTTP 404 Not Found. (It's fixed now)

An answer in reply to my recent XML question. More examples would be good, I'm trying to understand this...

Mena's Corner: "Nine days ago was the one year anniversary of our incorporation of Six Apart (prior to that we were a LLC). In this last year, we went from two people working out of our apartment to a company of twenty-four including a Japanese subsidiary (announcement) and a team in Europe working as our exclusive agent.
[...]
When we knew that Movable Type Pro had to be postponed because the direction of the product hadn't been clearly outlined in terms of development resources, we should have told our users."

Six Apart, eating their own dogfood, started blogging from within their own corporation.

Andrew points to A new kind of comments spam?.

Looking at this a bit more, it is clearly spam. Someone is creating a circle of fake "personal" websites using cookie cutter text (the 'about me' is the same for all the sites), various free hosts and varying images. Warning: the sites pop up ads of the most annoying kind.

It really wouldn't take a hacker much work to create thousands of fake "personal" websites that are a lot better and varied than this one, and then comment spam the hell out of the weblog world to build pagerank. I don't want to give anyone ideas but I'm pretty sure this is inevitable.

XML, trees and laticces.

I've been wondering about XML and highlighting text. If I have some text, and I highlight a piece and tag it with some subject, I can easily express that in XML.

But what if then I highlight another, partly overlapping part of the sentence and want to tag that with another tag? How am I supposed to identify that with XML? I don't want to do what HTML does with overlapping Italics Bold tags, because I want to remember what I highlighted as 1 section, not cut it up in different sections that are tagged the same way. (Is this making sense).

So what is the solution to this? Is there any? I think the fundamental problem is that, at a data modeling level, XML lets you build trees very easily, but not laticces (overlapping structures).

Help.

Annotated Aggregator client HTTP tests: UI standards to complement Aggregator HTTP standards.

Does something like CleverCactus exist for Windows? (invitation-only p2p)

I didn't catch the second part of this conversation - sometimes you just want out. Every day, I take the train to work. There are conductors who check your ticket, and there is a train driver in the driver's seat. I got on today, and passed by a kid being trained as driver. The angry teacher was puffing: "I will only tell you this once!"

RAP - RDF API for PHP. Now who does an XTM API for PHP? Call it XAP :)

I started rewriting the EasyTopicMaps wiki, starting with WhatAreTopicmaps.

Lost Boy: Scattered Self: "Let's synthesise the above rambling into a series of recommendations for building user-centric services:"

Read this if you're building web tools/services.

Structured Procrastination: "I have been intending to write this essay for months. Why am I finally doing it? Because I finally found some uncommitted time? Wrong. I have papers to grade, textbook orders to fill out, an NSF proposal to referee, dissertation drafts to read. I am working on this essay as a way of not doing all of those things. This is the essence of what I call structured procrastination, an amazing strategy I have discovered that converts procrastinators into effective human beings, respected and admired for all that they can accomplish and the good use they make of time."

A wiki that let's you do faceted browsing! Very cool.

XhtmlInterWikiMarkupStandard - InterWiki: "After some discussion on interwiki-discuss, we may as well start collecting ideas for a possible XHTML InterWiki markup standard."

AIfIA - Español published a whole new set of translations and some original articles in Spanish. Thanks to Javier Velasco, Jorge Vergara, Fernando Siles, Malisa Gutierrez, and Jorge Arango.

Via Simon, The Passivator (Ftrain.com). It highlights passive verbs and adverbs, calling your attention to them and helping you improve your writing. (passive verbs and adverbs make often not the best writing).

We're getting at a point where it becomes too much work (really) to follow all the going ons with the semantic web and such.

Deep Web Research: A Subject Tracer Information Blog.

I reinstated the EasyTopicMaps wiki. I ported all content. The formatting needs some work.

WebChoir Products - TermChoir: "It is a completely web-based editorial support system for managing and maintaining controlled vocabularies, classifications and thesauri, whether imported or created in-house."

Anyone used this?

Simon Willison's Blogroll says I updated my blog "-1 years, 12 months ago". Is that 2 years into the future? And if so, would my recent fooling around with futureme.org have something to do with that?

VoteLinks - Technorati Developers Site: "Indexing and tracking applications treat all links as endorsements, or expressions of support. This is a problem, as we need to link to those we disagree with as well, to discuss why."

Peter Bowyer's weblog: Template engines:- PHP's biggest joke: "PHP needs a recommended template system." It sure does.

Hey, Amazon now lets you look inside my book :)

A Brief Introduction to the Goals of Frassle: trust based taxonomy aggregation?

I like Google's "Don't be evil" motto. I think a huge problem with large companies is that their core values are distorted. People who work there then make decisions reflecting those core values (provide vallue to shareholders, for example).

If I ever start a company, my 4 rules for the company will be something like: (inspired by Asimov's robot rules)

Rule 1. Don't be evil.

Rule 2. Make a positive difference in the world, unless it means violating rule 1.

Rule 3. Survive and grow as a company, unless it means violating rule 1 or 2.

Rule 4. Treat employees well, unless it means violating rule 1, 2 or 3.

These rules say: If you can make a positive difference by being evil, don't do it. If treating employees well endangers survival of the company, and you're making a positive difference and not being evil, don't treat them so well. If you can survive as a company and you're not being evil, but you're not making a positive difference in the world, don't bother. Find something else to do.