thoughts'n'essays

chain of reaction

2008-05-01T06:08:00.000-07:00

One of the most important things in economy is to make predictions about the future. This does not only hold for speculating at the stock market, but also is day to day business for fiscal politics, company development, trend scouts, entrepreneurship, education plans, research funding, climate observation and hundreds of other fields. Some even state, that making predictions on the basis of experiences is the key factor for higher intelligence and complex cognitive abilities in the human brain.

The problem with guessing the future is, that there are too many factors to keep them all in mind. Some are forgotten easily, some are newly discovered, some are controversial. So in the world single entities emerged, busy with keeping track of all factors relevant to one topic at all times, thus creating highly specialized expert systems. Such entities (individuals, groups, etc.) are very important, but can easily fall out. If an expert of a special field dies, it takes a lot of non-experts to keep the network of factors alive. And maybe they won't succeed, and the knowledge/intuition is gone.

In order to not overload this entity, it would be good, to have a digitalized version of this relation network. Not only for causal chains, but also for positive and negative correlations between factors. A wiki would be a great way, to realize such a system. Every article is an event. This article lists all correlations with other events, and it lists all the events that lead to it, that were necessary for it and which events it is necessary for. All these relations have to be supported by research studies/articles. By not restricting the field of events, several networks will spring up at the same time, and maybe get connected after some time.

The gain would be, to not only have a good overview of complex chains of reaction, but also an index of all the research data supporting them. Remote effects could be much better understood or discovered, by revealing complex chains not seen before. And confusions between causal chains and correlations could be more easily uncovered. I think it would be worth a shot, because I have not seen anything like this so far, and a wiki would be a great tool for realizing this with a large base of users to contribute.

Online rating systems need an overhaul

2008-02-05T05:54:00.000-08:00

As Brad observed quite nicely in this post, ebay has a problem with its rating system. Well, actually not only does ebay, but all websites who use them do. The root lies in different behaviour of users, and different reasons for them to give ratings. And there's the question of what the parties that are subject to the ratings intent to achieve. In ebay it's mostly a question of finances - the better the score, the better the revenue. In IMDB or youtube it's all about attention, the more votes you get, the better.

In my opinion, it would be better not to hand out negative votes on a quorum-basis like brad suggested, but to overhaul the whole rating system. Ebay has a scale of 1-5, some websites have 1-10, 1-3, 0-10, etc. Most systems don't even have descriptions for the meaning of the scores. And users use these scales differently. There's the question of defaults: some have highest rating as default, some have average default, some have lowest as default. Some tend to rate extreme, not using the full range, some are very careful, not using the full range either. Until now, the companies assume that with large numbers, these effects will all even out into a nice bell shaped graph.

For example, to avoid bad results, IMDB has very special vote rating system for their Top250 list that goes like this:
"weighted rating (WR) = (v ÷ (v+m)) × R + (m ÷ (v+m)) × C" I don't think this is very understandable, and they also have to exclude non-regular voters.

Another problem is, that with the old rating system, many users are influenced. For example they have a look at a movies rating on IMDB with something like 60000 votes, averaged 7,6. and they think, "hey, this movie doesn't deserve 7,6, it should be lower". So what they do is, they don't won't rate the movie at the score, which they actually think the movie deserves. But they vote 1, in order to lower the aggregated result as much as possible towards the score they think it deserves. This is a problem that all rating systems have, where the user has information on other voters choice before they vote on their own. On Websites that is most of the time the case.

So this is my new system:

I think it would be best, if every vote is contextualized with all your other votes, and thus normalized and comparable and aggregateable in a combined rating. If on a scale from 1-5, one user only gives two votes 1 and 5, and another user only the votes 2 and 4, they actually express the same opinion. By normalizing the data, you would have as a result, that the first user has an average vote of 3, with an average variation of 2, which leads to a -1.000 and +1.000 from average vote. The second user has the same vote results, because he has an average vote of 3 and an average variation of 1.

If you ever had a course on variance analysis, this comes naturally to your mind. Apparently it doesn't for the creators of the voting systems, they just calculate the average vote.

What will it change? First of all, you won't be confused with interpreting different rating scales anymore. You'll have a good guess, what -0.612 means in contrast to +0.997. Second of all, the influence of prevoting result spotting won't affect the vote so much anymore. This is because, you can't overexpress your opinion anymore by voting more extreme. Extreme votes are only heavier weighted, if you have a lot more moderate votes, and that takes time and consideration.

So, the math is easy and already used heavily in analysing surveys and stuff like that. The voting wouldn't change from the users perspective, so they wouldn't notice much of a difference. Only the results would be much more interesting, because fake-votes are stripped of their power.

picture by Sara and Mike

New applications for wikis

2008-01-03T05:38:00.000-08:00

When you talk about wiki, everybody thinks mainly of wikipedia, wikimedia, wikibooks, wikinews and the other projects. They all focus on the article-part itself, and forget the power of the structure a wiki offers. For giving a wiki a more semantic touch, as is proposed for the web 3.0, you don't necessarily need to implement new software features. You can do it with what you have already. But a change in the way of thinking about wikis is needed to do that. I want to outline some ideas for new ideas of realizing a wiki, i think they are somewhat innovative, and would be quite useful.

The food chain
The chain is not so much a chain, but more of a web. The key for it aren't the animals and plants and other organisms themselves, but the combination of who eats what and which parts. Categories could explain groups with same eating habits like carnivores, herbivores, etc. In my opinion it would be nice to visualize small parts of the web. It would be helpful for understanding and creating ecosystems.

Product Parts
There are hundreds of thousands of companies creating products. All of them use products of other companies as product parts, materials, assembly machinery, etc. It would be really interesting to see, how this web is spun, how complex a product is, and what would happen if one element in the web would break down. It would help us to understand the meaning of natural ressources, the importancy of recycling trash, and open positions for further developments.

Water systems
Every river has a spring and ends in the ocean. No wait, they don't. Water is a self-organizing web of streams, from the smallest trickle to the widest stream. They are all in a hierarchy, which sometimes can change, and sometimes is forcefully changed by humans building canals.

Company shares
Everybody knows about it, stock market analysts are the gurus of it, but noone has a full picture. Many companies have shares of other companies, they have daughter companies and conglomerates etc. A wiki for companies could give that information to everybody, and show the problems of the system itself and the impacts of any disturbances in the market.

Can you think of other ideas? Write a comment on it!

pictures by Domenico Nardone and Tobin Fricke

Search Engines 3.0

2007-12-29T06:03:00.000-08:00

Ever been on Google Website for Searching something? Have you noticed how 1998ish it looks? Ever wondered why you have to type complicated terms with hyphens in a form field as if there was no 21st century? Are you one of the many who thinks that Google has grown a monopoly that needs some serious competitors?

I think the time has come for some new steps in search engine developments. If you can't improve the database and it's indexes anymore, you always can improve the communication between humans and machines. And when it comes to searching, we are 10 years behind, due to lack of competition.

My vision of a new search engine looks like this. On the start screen you see a tag cloud of the most important web content genre names, as well as the newest keywords in the news. All main keywords are spatially organized by dimensional analysis. By scrolling around and zooming in and out, you can explore less popular areas of the field. The tag cloud is in constant change, because new items grow out of nowhere when something new happens, and old topics shrink and disappear, when noone uses them. The more people search for something, the bigger it is. You can drag and drop any keyword and combinations onto the search field. When you click on a tagword, a new tag cloud opens up with the keyword in the center. But this time, the keywords have undergone a cluster analysis, thus simulating a semantic categorization. So the tag cloud is spatially grouped into clusters.

Search results aren't a list anymore. Noone likes these lists. Rather it is a new tag cloud, this time it is showing URLs of the top level domains. The more search results in a domain, the bigger it is. The URLs are also grouped by cluster analysis. By scrolling and zooming, you can navigate through large fields of results. This also would reduce SEO and Sub Level Domain Spamming a lot.

Similar concepts could be introduced for Image results, where not only keywords but also color, size, etc. could by taken into cluster analysis. News items could have timelines, geographical locations as hierarchies. I think it would be pretty easy to improve search engines, many bigger websites have amazing local solutions, so why doesn't Google crank it up a notch?

picture by Matt McGee

New Year

2007-12-27T06:10:00.001-08:00

As it goes for innovation, I want to bring my blog to a new level in the next year, too. So here is what I did: I opened a poll, for you to tell on which fields you want to read more. I'm pretty keen on thinking and developing new ideas, but the more feedback I get, the more productive I am. I also want you to comment on my posts not via e-mail, icq or in person, but right here in this blog, below the entry you just read itself. Don't be shy, I make a lot of mistakes, and only you can help me find them. I can learn a lot from what you're saying and thinking, after all you have been elected Time's Person of the Year!

Thank you so much!

picture by Laura Mundee

courageous city planning

2007-12-24T05:28:00.000-08:00

When you look at any larger city in the western world from above, most of the time you see the same pattern over and over again. You have blocks, streets, sidewalks and train tracks. The most important features are the roads and streets, they divide a city into a lot of blocks. Sidewalks are between the roads and the blocks, sharing space with the street. The wideness of the streets correlates with the size of the blocks it separates. Most of the time they have lines of parking spaces for cars. Blocks are a bunch of buildings in a square, that face the roads, forming squares and rectangles, with some larger yards in the middle. The yards are used for gardens, supermarkets, playgrounds, etc.. Towards the yards the houses show balconies, winter gardens, terraces and a lot of windows.

But does it have to be this way? The importance of cars in our society, has made them the main design-theme for our cities. This is why we have to combine streets and sidewalks in complicated intersection systems with traffic lights and long phases. This is why everyone who wants to move around in the city without using a car, has to participate in car traffic and bad air conditions. And this is why we consider cities to be rather ugly and smelly.

In my opinion there's the option for another design. As I tried to point out, we don't have to keep the system that historically replaced our roads for all kinds of relocation to become streets for motorized verhicles only. So here is what I am proposing as an alternative system. Imagine streets without sidewalk and pedestrians. All intersections would be much easier to design, and much more efficient in guiding traffic. But where would the pedestrians walk? There would be a second grid, intersecting the city, but shifted half a block to the southeast. So as a result, these pedestrian walkways would cut every block into half. One block is connected to the next block, not on the corners of the square, but right in the middle. The blocks would need to have at least one gap on each of the four sides, so that connections to other blocks can be established. Guiding pedestrians across a road is much easier and more efficient than at intersections. The buildings would have their entrances toward the center of the block, where it is more quiet and cleaner and nicer to look at. The pedestrians would experience a much greener and quiter city, where they would encounter more neighbours on the streets. the streets would be much easier to clean, and wouldn't need gardens to seperate living space from road noise and smells. Bigger streets could also crossed on a level below or above the ground, so that traffic lights wouldn't be needed at all anymore. It would be just like turning your back on the cars, and turn toward the neighbours and community.

I think such a coexistence would be a good solution between car-free city-centers and traffic jammed surrounding areas.

picture by Jonn 'Dorvak'

put usb flash drives into action!

2007-11-11T06:27:00.000-08:00

Nowadays everybody has them, small little devices accessible by USB, to store data on. Some have functionality for playing mp3s too, but that's it. So whenever I want to exchange data with a flash drive, I have to lend or exchange it. But who wants to give away a device he paid for, and wants to keep using without pausing?

My little idea is a device with two or more USB-Slots, that can transfer files from one to the other. So if you want to exchange files, you can just stick your and your friends usb flash memory into the device, select the files to be transferred, hit action, unplug the memory sticks, finished. This is less functionality than any mp3-player has, and should be easily created.

An enhanced version would be memory sticks, that have an USB plug on one side and an USB socket on the other. With the functionality for sending and receiving files implemented in the stick itself. So you could put an infinite number of sticks in a row. This wouldn't only come handy in a situation where you want to share files in a group, but also you could use multiple sticks in a single USB socket thus reducing the need for many slots by virtually increasing the size of the flash memory for other devices.

And now imagine what could develop out of this, keeping in mind that how small flash memories will become anytime soon, how fast transfer rates will become, and how USB could be replaced by Bluetooth or WiFi.

picture by Ian Hampton

multidimensional artist analysis

2007-11-06T07:06:00.000-08:00

Ever wondered what is the best way to categorize your music into different folder with no more than 15 artists per category in a strict hierarchy like your harddisk? Let me tell you, it is impossible to find a system that works, unless you go for alphabetic systems. I've tried several approaches. I sorted by genre, by time, by mood, by instruments but nothing ever works, and when i search for something, i always end up searching around not knowing in which folder I put this band in.

So what I have in mind is a website, where you can submit a list of artists. The artists are analysed with help of wikipedia and last.fm-stats etc. for certain multiple dimensions like time, mood, genre, and so on. Own dimensions can be added, but have to be tagged by hand for all the artists. Now every artist is being put in a multi-dimensional space, and a clusteranalysis will be run. The multidimensional space with the clusters will be visually presented to the user. The user can tag certain artists as prototypes, which ensures that they get their own cluster and they are in the middle of it. There are sliders for every dimension, which with the user can give weight to dimensions. He can also choose how many subcategory-levels are allowed in the hierarchy. When the user is satisfied with the cluster-boundaries, he can output a list with the new hierarchy.

This should not be confused with a "similar artist"-classification, because it is so much more powerful. But it is one way to put tag clouds into actual use.

picture by David Cohen

ipv6 meets rfid meets sensor networks

2007-11-01T13:35:00.000-07:00

Itaru Mimura from Hitachi Ltd. presents in this paper the problems that could be solved with introducing the current technologies of IPv6, RFID and sensor networks on a broad scale. And it shows the possibilities that are opened far beyond the problem-solving itself. It talks about our homes, about mobility, logistics, agriculture, industry, security, traffic, convenience and a lot more. But it also tells us, that this is not science-fiction, but could be introduced any day - because, look at the date, it's three years old already. All we need to do is some creative thinking, some ethic discussions on what is doable and what isn't, and some courage to explore new territory. Scientific progress goes 'boink'!

wikipedia-enhanced searching

2007-09-26T04:00:00.000-07:00

How often have we tried to search for something, and then stumbled across thousands of search results with the same word but a different meaning. The only chance to get rid off these is by using exclusion terms or more keywords of the same field of interest. But similar to my idea of wikipedia-enhanced translation support software, search engines could also make use of the huge knowledge database in the wikipedia. Not only could searched keywords be tagged with language, but also with a specific lemma or category. This way the search engine could have a better understanding of what the user is looking for, and throw away the rest of the search results.

For this to work search engines would need a way to analyze a websites content, regarding to what language and field of interest it is about. This sounds quite simple, but turned out to be much more complicated. Most websites still do not metatag the language correctly, and loan-words and productnames always mislead the first guesses. Specifying the field of interest has been tried with automatic translation software a lot, but still it is hard to disambiguize correctly. Using the wikipedia-article of a keyword for finding related keywords in the text or the category tree would help a lot with this problem. Also stochastic data on cooccurrences of other words for specific meanings would be helpful to accomplish this.

The feature of this search optimization lies in the further development of the semantic web idea. Wikipedia is a good basis to provide basic semantic knowledge about the world and how languages represent the world, so other applications that have to deal with semantic data should not ignore its existence.

picture by David Hand

Force the feedback!

2007-08-20T15:30:00.000-07:00

One very amazing thing I have seen as a kid, was when I visited a friend of mine who had very fance stereo, and everytime he uses the remote to alter volume, the dial button would turn itself up and down. Although I knew how easy this is to realize, and of how little use that is, it was the feature I liked the most on this upscale stereo. Having written some ideas about the future of feedback learning, it came to my mind how not only sensors can enhance devices, but also motorization. I'm not talking about turning PCs into robots, but rather enhancing what we already have by making it a bit smarter, to improve the men-machine-interface and thus open up new opportunities.

Computers can communicate with their operators only through audiovisual ways. This puts a lot of difficulties to work around in cases where operators seeing or hearing is impaired. Braille was a good invention for having haptic communication from the machine to the operator, but there could be more ways, which would help not only the visually impaired community. Think about a keyboard that can regulate the response of the keys. It could block keys completely, only a bit, so it would just be harder to press them, or even pull down the keys as if somebody was typing, like the self-playing piano-machines. This could come in handy for games, easier typing interfaces, orthographical live-checks, typewriter learning software and more. It could also implement force feedback communication with other humans. The same could be done for a mouse, thus guiding the hand of the operator on the mouse in difficult drawing situations, teaching mouse gestures, disallowing to leave certain areas, and so on. By having mouse buttons not only on or off, but with a dynamic range, you could also use it for smarter tools for something like image processing software, sound apps, and more games for sure.

I know that this is an idea that turns on the digital age and drags the digital to the analog world, like i did with asking for more sensors, but the more skills computer can get that are similar to human skills, the better the communication between humans and machines will work, and the easier it is for artificial intelligence to arise. Because we should never forget that when we think of intelligence, we talk about human cognitive skills, and not patterns that emerged by design. Honestly, we would not recognize intelligence in other forms even if it were shoved in our faces, as it happened before already many times. We are a bit ignorant when it comes to thinking outside of the picture, after all.

picture by Nicolas Nova

open the web, liberate connectivity!

2007-08-20T06:20:00.000-07:00

A few years ago the hype-word on the internet was P2P. Everything that was realised in a one-server-many-clients fashion before, suddenly started to introduce new protocols and applications to distribute the heavy load at the bottlenecks of the internet. So what we have now is distributed downloads and streaming techniques, but most of them still rely on certain servers distributing the basic information to get things going. So what has been distributed with P2P is not the access to the content, but the large data amounts of the content itself and the transmission capacities. But all the ISPs out there are still working in the old one-server-many-clients fashion.

On the other hand we have a major change of the infrastructure of the internet-users, because of the many broadband connections in the households, the mobility of the users and the new technologies of wireless network access. Instead of switching to IPv6, there is a whole new revolution in the very structure of the internet coming up. I am talking about the slow but steady decline of ISPs, and the raise of a true internet, where everybody can join in for free, if he has the technical devices. With whole cities starting to be covered with WLAN-Access, a whole new opportunity arises. Instead of sending all traffic to the ISP, who then uses the internet backbone to serve the users wishes, the city could create a local parcel of the internet. So that while still being a part of the internet, it is not dependent on the ISP or the backbone anymore, because communication happens P2P, within the cables and the air connecting the citizens of the city.

For this to happen P2P offered big solutions to us, and have been implemented with distributed web hosting solutions and Direct Connect and the like. The key players in this game are the ones that brought the internet to the masses in the first place - the universities. Most universities have an excellent LAN with all the buildings connected, with dial-up lines for scientists who work at home, with student dormitories networks included in cooperations, and with further cooperations with local government networks. The most expensive part of data transmission is getting long distance data. To reduce the ratio of long distance vs. short distance data, it is important to help build local parcels of the internet, where traffic is much "cheaper". OpenFON is dreaming of a wireless LAN-network, but they still rely on broadband-access to ISPs. The main reason for this is, that ISPs have no interest in providing any help for their downfall. Most of the implementations of a distributed Internet are fighted or ignored or critized. With every new version of WLAN Protocols the ISPs power is vanishing more and more, and will crumble to a nothing once the density of WLAN-Device has surpassed a critical level, and the internet protocols and the software have adapted to the idea of an internet that actually maps onto the real world.

If you talk to scientists on the futures of car traffic, they will all tell you that it will be controlled by a Wireless P2P-network. What the ants and other insects and animals have teached us, can now be put to actual use, by creating distributed internet, the mightiest communication network out there. Only then will the internet actually be invincible to attacks. Only then our devices can become a part of the internet, and not only a input/output station. And only then can the hierarchy of the internet be broken up, and the problems of IPv4 be solved.

By all the money we pay to the ISPs to let us take part in it, we gave them a lot of power, without restrictions on how to use it. Now they start tracking their users, handing over information to legal prosecution, and they still have policies where they can cut you off the net because they do not like what you do with it. On the other hand they did not take over responsibilities, they do not care about the content as long as it is legal, they do not take action in developing the internet with all its protocols and standards, they do not care about spam and computer fraud and identity theft and so on. ISPs as they exist now, as mindless money-making machines for technical services will not exist in the time to come. They will have to focus on either quality services by offering their customers high-speed or even realtime access to services on the internet, or on security by offering a safe zone, wherein content and legality is highly controlled.

My vision seems to be a bit strange, because it is not easily understood where the advantages lie. I think the biggest advantage is, that getting access to the internet will be a lot cheaper than it is now. By not paying monthly or annual fees, but only paying once for the device, is a huge change, and certainly a future for the extremely large cities in the poorer countries of the world. Second of all locality is supported, enhanced and put into focus. This will not only reduce redundant traffic a lot, thus saving energy and infrastructure, but also let local communities grow, which is very important in the globalization age. And furthermore new P2P projects in this world, like traffic control and media-broadcasting and personal communication will have a ground to grow on, and thus be relieved from the power of service providers. After the downfall of the ISPs maybe we will have a downfall of the telcos, tv-broadcasters, radio-transmitter-stations, publishers, schools, and so on. We will see...

first picture by Jose San Juan

job application support platform

2007-08-19T09:43:00.000-07:00

Everytime you want to apply for a job, you start at the same point. You spot some job ads, you do some research on the companies, you consider what you can do, what you want to do and what they are searching for, and you try to imagine if this all fits together in a single picture. If you see the picture, you start writing an initial contact e-mail, updating your years-old cv, asking friends for help, and then send it out, in hope you get an invitation to an interview. If you get the invitation and the job, you forget your cv completely. If you do not get an interview you try to modify and enhance the cv, to a point where the success rate is higher. If you get the interview but not the job, you will send out the cv to more ads, you review your ad selection process and you try to improve your interview skills.

Instead of starting from scratch everytime, it would be much better if you always have a cv ready that is up to date. And it would be much smarter, if not everyone would have to learn by making the same errors that are usually made. Figures are out that in a bunch of applications and cvs that large companies receive, a high percentage is dismissed instantly because of major flaws. One way to go would be to teach every single individual the whole concept of self presentation, formal aspects, etc. An easier way would be to provide a platform where professional aid is given to ensure the quality of your applications in regard of content, style, layout, etc.

I am thinking of a platform that serves as a permanent repository for CVs and application letters. Each individual who opens up an account, fills in his personal information, choses a premade layout or modifies one in the set of premades, choses a premade application letter or puts up his personal one for easy configuring with new data, and sets privacy setting on who can view his cv. Everytime he wants to respond to a job ad, he opens up a new item, links the ad, makes notes on details about the company and the contact persons. For a individualized cv and application letter he just ticks the items that should be included for this job ad response, reviews the final outcome, and sends it out. Afterwards he makes notes about the response he got, and tracks the further process if being invited to an interview.

The advantages for the user are manifold. First of all he can ensure that his application letter and cv are free of major flaws, professionally layouted and in a file-format that everybody can access and print out. Ever heard of Word-Documents that save the last edits to the document in the file itself, so the recipient can find out what has changed in prefinal versions? This will not happen anymore. It is much harder to forget some important information, when the user has a checklist or form that helps him with filling in the content. Standard application-recycling errors like copy'n'paste leftovers or wrong names in the text are in the past. Furthermore he can easily individualize each cv and application letter to the specific company and job ad he responds to without getting confused. By tracking the informations for his job application history, he can easily overview which letters he sent out to which company, which ones were replied to, which replies were positive and which negative. He can also get statistics on which cvs and application letters were successful and which were not, so he can further improve his turnover rate.

By using privacy settings he can also release some standardized cv for everyone to look at, so that employers can search in the database for potential recruiting material. When the turnover rate for interview invitations is good, but the job search still is not successful, further help in that area can be provided by giving general advice and links to professional coaches. Also a community can grow that helps each other with reviewing, developing new layouts, exchanging sources for job ads and ideas for new fields of profession to aim for.

The gain for employers is also visible, as the quality of applications is improved, without falling into standardized applications. They can scan the platform for potential workers even without putting up search ads. Also they can give advice to what should and what should not be in a cv or letter of first contact.

Further development could be a standardized incorporation of certificates from other services. So whenever a user gets new language skills certificated, he would not have to update all cvs by hand anymore. Multi-language support would also be a great advantage, as most of the time it is very hard to translate schools, courses, degrees and positions into another language. Without help this can end up in a big confusion, but with the job market going international, it is very important to be able to send out applications in a foreign language as well.

I don't know if such a platform already exists, but it surely would attract a lot of people who need help, because all their applications are turned down without an interview. Self presentation is not magic, it's just a skill you can acquire and with some help you can acquire it a lot faster than by trial and error.

Web reporting for everybody

2007-08-14T18:15:00.000-07:00

Every active online user has a lot of things going on at different places on the web. There's social networking accounts, forum accounts, onlineshop accounts, ebay accounts, webservices accounts, product registry accounts and so on. Most of these involve not only a personalized part of this special site, and information about the user to the site, but with the web going 2.0 it is very important to get information regularly to the user as well. In the beginning this was done mostly by emails. When spam came into the game, email became very unpopular. So most sites decided to introduce Dashboards and Profile pages, to keep the user informed. Although not being subject to spam, this concept has a strong disadvantage: the user has to check all his dashboards all the time to know what's going on. It was easy to check for new mails every minute, but you can't check 10 or more websites that often. So what the web needs is a user-centered reporting system.

A user-centered reporting system would look like this. The user uses a program or webservice, which checks all the sites he specifies for the data he specifies. Preferably the sites implement some push-technology when new data arrives, instead of relying on the user to poll for changes. The user choses the layout and processing for the incoming data, so know he needs to open only one page, to see what has changed since the last time he was on. He also gets information when changes have happened and backlinks to the specific dashboards or other links he wishes to see on certain items. He also can see what happens while he is watching the page.

The advantages of having one system for all kinds of sites and forums are quite numerous. A system like this would give the user more control about his online-activity and especially on the activity of other parts he wishes to know about and react to. Also by going live, communication could be enhanced by shortening time distance between actions and reactions. On the other side overall traffic would be reduced, because you don't have to check your regular sites when nothing new has happened. Subscription to RSS-feeds to just to know when changes happen, even if you don't want to see the content itself, would be reduced also. Furthermore notifications won't get lost in spam-filters, or site with low activity forgotten in the bookmark-list. The sites have an interest in doing this, because they want to get information to the user, in order to get the user to become active. And the users have an interest, because they want to know what happend as fast and as well presented as possible.

Getting information from already existing XML- and RSS-Feeds as well as open APIs would be a good start to get something like this going. In the beginning you would want to try some regular pull-checks by the system on the specified sites, until push-data is well known enough to be implemented on the sites itself. But I think most of us would gladly see the final downfall of email notifications, forum-thread subscriptions, login/logout-sessions, news headline scanning for new news, etc. I think the first one who actually succeeds in realizing such a system, will be the one with the most entry-pages to the web and thus the best opportunity for marketing and advertisement.

feedback enhanced learning

2007-08-09T09:41:00.000-07:00

If you want to learn something, you have to make errors. Although it's nice to learn from errors, that is only possible if you know what you want to learn and if you are able to recognize your errors. For most learning processes this is pretty easy, because the goal and the process and the errors are well known, because the learning process is an established one. This accounts for learning for exams, learning to ride a bike, learning how to be polite, and many other things that most of us have learned at some point in our lives. But there are also some areas that are very specialized, and only an exclusive groups tries to learn this. Most of the time it is a field of learning where is little known about the process and the standard errors. Why are some areas explored by few learners only? Either because it is not known you could actually achieve the goal, or because you don't see a use in it.

The crucial point to optimizing a learning process is not only to specify the goals with more and more details, but also to enhance perception of flaws and errors of actual performance of the learner. So while until now learning was focused on learning what to do, you know also learn how to learn it. It's a turn from a result-oriented approach to learning evolution towards a way that also includes procedural evolution. For example if you have a learner of running, and you want to optimize has step-frequency so he gets to the goal the fastest. A result-oriented way would try to find out which step-frequency would be the best to achieve the goal, and then breaking the step-frequency down into finer and finer time-spans, different frequencies for different phases of the run, etc. A procedural way would instead try to take a better look at what frequency the runner is actually performing, and getting the measuring more and more detailed.

So how can you achieve better perception of actual performance? One way would be to train the learner to better listen to and watch his own performance. A big step in enhancing this perception is, to not only take videos and present it to the learner afterwards, but to implement some instant feedback. Instant feedback is nowadays possible through digital equipment. In the past you had to rely on coaches yelling at you and mirrors for yourself to look at, but now it is possible for hardware to keep surveillance of your performance, process the data instantly and generate some computertranslated feedback into a signal that you can easily perceive and interpret. So if you have the same runner-learner, instead of a coach counting his steps of the race and calculating an average frequency for the overall run at the end, now you can have sensors in your shoes, that gather data with every step you take. This data can be processed while running, and generate some kind of sound that indicates if you're too slow, too fast or right on frequency.

By enhancing perception of your own performance in such a computertranslated way, your brain is able to learn abstract performance-features like step-frequency, so that after many training sessions with instant feedback, you can get rid of the feedback device and still now what frequency you're running on. Furthermore performance data is becoming more objective, and dependency on a coach is lowered. While instant feedback systems like this are already in use in certain areas like professional sports, it has a big potential to enter other areas. It has been shown that patients can lower pain or gain some control over bodily functions like heartbeat-rate and blood pulse by giving feedback. While this sounds like something like auto-suggestion or self-hypnosis, it is not, because in contrast to these you actually have input on the current performance, and are able to learn what kind of behaviour (motoric, mental, emotional, etc.) has what kind of influence.

Some other example, an overweight person wants to lose weight. Without feedback devices the person would just count the calories to take in, and trie to lower it to a level where it matches the calories an average person burns each day. And then the person realizes it doesn't work like this. This is because, there are other factors than calories that come into the process, where you don't have feedback on. Emotional states are important for gaining and losing weight, physical activity is important, and some other aspects, too. So a good dieting coach would advice to keep a diary of emotional issues, that can be compared to the according weight of the day, to see how it affects the process. Such a diary is a feedback, although not instant. But for physical activity you could give instant feedback, on how much calories are burned, thus giving the person instant reward feedback which would not only help the person to be happier and thus enhancing the performance,, but also make the person to try doing more physical activity right away.

You could also translate the data of a 3-axis-accelerometer sensor, like the iphone has, into a camera, and use it for helping the photographer to take pictures with steady hands, untilted view, panorama- or sequence-shots, and the like, by giving him enhanced audio or visual feedback. There are so many things you could learn so much faster, without relying on years of experience anymore, because digital sensors and processors can help to take a shortcut on the learning processes, because the time-stretch between performance and result is lowered to almost nothing. If you eat something bad, and you get sick afterwards, your body learns it, but the longer the time-distance between eating and sickness, the harder it is for the body to learn the connection. This lies in the nature of learning processes in the brain-cells, activity declines very fast and memory too, so for learning fast it is good to know what you do wrong fast.

A further effect is, that with a good system like this, you don't have to rely on coaches and teaches so much anymore, because well defined learning goals and errors can be learned just with the learner and the helping device. Like in school we discovered how much better it is to learn vocabulary alone with some cheap software with instant feedback, just because you don't want to pay a private teacher or your parents don't have the time. Now is the time, to transfer these benefits of the digital age to more and more learning processes of every day life.

Food shopping helping device

2007-08-08T05:53:00.000-07:00

When you walk into a supermarket to shop for food, you're confronted with thousands of different products and brands every time. Not only would it take too much time to take a look at all the products every time you go in there, but also it is impossible to read all the information about nutrition and ingredients for every single product. But not few customers need to know some of this information for their health, for easier decisions, or just for curiosity.

What I have in mind is a device, that stores all the information in a large database. It would use RFID and OCR-Scanners to identify a product, and instantly have all the proper information at hand. Well, this is nothing new, because what's the difference between reading the print on the package and reading the device screen. The big difference is, that you personalize your device with the information that is important to you. Many people have allergies, some have diabetes or other metabolic disorders, some are vegetarians, vegans, etc., some need to keep a strict diet for sport events or health problems, some want to boycott certain food products for other reasons. This device can be tuned to such special conditions, so that when you have diabetes and can't eat certain types of sugar, you don't need to go to the special area in your supermarket where only few and overpriced products are stored for this special need, but instead you just run around scanning products in a matter of seconds to know if you can eat it or not.

So all that we need is not new laws on what the companies should print on the packages anymore, but rather a law that ensures every company releasing the information in a standardized (computer-readable) format. Costs for such a device would be very low, as all the components are already available in mass production. The key is to have a good infrastructure on all the data coming in and being kept up to date. Maybe you could have something like a wiki in the beginning for gathering information through volunteers, and later on you could start automatizing it with the companies itself sending the information.

Where are the smart apps?

2007-07-24T16:51:00.000-07:00

When you get a new application on your computer, you're quite excited about it. You've heard a lot about what you may do with it, you have some expectations what you want to do with it, and it has some promises on the package what you could do with it. The problem is, once you've installed the thing, you're at a loss where to start. Some programs deliver a small introduction guide or video along, some have a handbook, a Help File or a Website with forums. The conventional guys go to the store and buy a book, some others get somebody who knows the application already as a personal tutor, and most of the people just try learning by doing, messing around with it until you get it to do what you want to do. But let's be frank, how much percent of the features an average modern application has do we actually know and use? Most of the time I popup some dropdowns, i only now 20% of the items, sticking to what I already know. Even with all the help introduced so far, you never even get close to the 50%. And just knowing what something is good for, doesn't mean at all you also know how to actually put it to use.

In my opinion the software industry needs to develop a new generation of software - one that knows and learns and keeps learning, what the user knows. Instead of flooding him everyday with options he only tries to ignore, the first day should be some first-contact session, where the user and the program get to know each other. On this basis the software should learn which features the user is already proficient with and which not. So in future sessions it will not show advanced features in the new smarter interface, unless the user asks for advancement. Such an advancement could be initiated by some learning session, where the user gets to learn how to use a yet component of the application, so he knows in the future what it is good for and when to use it. From this day on the application will include the feature in the smart interface. So this new interface will only show you what you can actually use. For sure you need some override mode, if you want to explore the application on your own, or you've learned something new elsewhere.

Another enhancement of applications is for the software to listen to what the user does with the software at which time. Regular processes could be shortened, use cases could be identified, and the interface could adopt to what is needed. This goes for dropdown menus, tool arrangements, handle presets, configuration states, etc. Some applications offer the feature to use differenct configurations by commandline-parameter, so you can run it in different predefined use cases. What I have in mind is a software that learns how the different configuration states and GUI layouts have to look like on it's own, offering the user on startup different usage options, and on-the-fly-switching between the different variants.

A feature like this could also help with anonymized reports to the developers, to gather information which features are used and which aren't, so the application can be further developed and designed to suit the users needs even better. The times where a software is only a peon that does the tasks the user tells it to are over, it needs to gain some intelligence due to the complexity, because no user has the time and patience to learn every application uses to its fullest extent. He should always be _able_ to do this, unlike with Microsoft Products where users just lose control, but he should also be able to voluntarily refrain from control.

So far only few products introduced features that somehow hint in that direction. Opera for example notices when you first use a mouse gesture, and asks if you want to use this feature in the future or not, so it can activate or deactivate it accordingly. Windows learns which menu-items you use regularly, and doesn't show the unused ones anymore after a while, unless you klick on a special arrow. The wand tool that some browsers offer for filling in userdata into forms on the world wide web uses a similar concept. There's much potential in that area, because the better you design learning mechanisms, the less companies need to spend on courses, instructors, books and tutorials for their workers who have to use it. Many big companies in the IT-industry already teach their workers with online interfaces, cisco and apple for example, and it's getting some good results. So if you design good learning mechanisms, people are eager to try them out, thus being more efficient workers in day-to-day tasks where the application is put to use.

Because if you don't know that your text-editor can sort your data easily, you'll just keep on doing it by hand until someone more proficient catches you in the act of doing so, and hinting out that a lot of time could be saved. But what are the chances of that to happen. Life long learning hasn't been applied to software-lives yet...

sociology of the virtual world

2007-07-22T12:18:00.000-07:00

According to the development of the internet according to the levels of needs stated by abraham maslow (which i explained in my first post), certain groups of users were attracted to the virtual world. Thus we can outline 5 user groups of onliners:

The first generation are the geeks, the boys, the black hats. These are the pioneers of the internet, the ones that made it all possible. They invented and used http, ftp, irc, newsgroups, gopher, telnet, they're thrilled by the possibilites that emerged with these technologies, and they wanted to use it just for the sake of using it. This community is the first one to join up in new online projects, searching for new possibilities, testing right away what just emerged. The key to understand them is, that they think of the virtual world as an alternative of the real world and not as an expansion. They didn't want the real world companies to enter the virtual realm in the first place, they don't want the virtual world subjected to real world country boundaries and legal systems and all the other projections of the real world. Their online identity is separate from their real world identity, with a well established nickname, separate money accounting, separate social circles, and most of the time a different personality. They want to be permanently online.

The second generation are the techs, the parents, the admins, the white hats. They are the ones preventing the virtual world to disrupt into chaos and anarchy. They develop the standards, they keep the systems running, they try to evolve the software and protocols to offer more security. They see the great potential of the virtual world, and want to help it to grow, bring order and rules. They also want the virtual world to act responsible, thus becoming more reliable. So they don't try to discover new possibilities on their own, but rather watch what the geeks discover and clean up the mess afterwards. Thus providing a fairly good damage control, in extreme cases they also don't shy away from ratting out, reporting to the police or even sueing those who don't stick to the rules. As they don't see any legal system in the virtual world, they just transfer the real worlds systems.

The geeks have problems with the techs, because they think of the techs as the fun-spoilers, the self-proclaimed virtual policemen, lawyer and judges all in one, forgetting about the stability they bring. The techs think of the geeks as the anarchic, childish, irresponsible disruptors, forgetting that they are the ones that get it all going. The geeks would like to see that the techs wouldn't apply real world systems, but rather regulating the virtual world soleley by technical and virtual means. Good examples for these are standardizations, provider policies and security blogs. Or for punishments G-lines, account suspension and port blocking. Bad examples are termination of provider account, report to the police, sues and fines. The techs would like the geeks to act more responsible, to not exploit every vulnerability just because they can. They would like to see black hats turn into white hats, kids into grown ups, anonymous ghosts into adressable persons. Their common interest is to keep real world criminals out of the virtual world, so when it comes to identity theft, credit card fraud, and the likes, they are eager to work together.

The third generation are the mods, the girls, the chatters, the smiley-inventors. This group takes the possibilities and securities for granted, and now want to heavily use it for communicative means. They are the ones that build communities, moderate chats and forums, and that use emails and IMs for personal talk. It is the first group that includes larger numbers of females, thus bringing emotions and empathy to the virtual world. They want the virtual world to be a place open for everyone to join, communicate and express themselves. They join and use only the established possibilities and technologies. The greatest potential they see in the internet is that shortening of distances for communications, as well as new ways of communicating.

The mods can't distinguish the geeks from the techs. They consider them all to be boring and lacking emotions, ignorant of feelings, but they're thankful that they created the virtual world the way it is. The geeks think of the mods as the pointless people, they don't understand what they're wasting their time with smalltalk and gossiping. They like it to have them in the virtual world, because they create communities with interesting content, and of course they like to have more people around, but especially people to make fun of, mess around with and exploit the lack of technical knowledge. The techs think of the mods as the tools. They like to see them come, as an indicator of their personal success. These are the ones they secured the virtual world for, and so they think they are dependent on them. They like to promote and demote mods by technical means, always trying to get them to understand the importance of their rules. The place where they all work together is the development of netiquettes, forum guidelines and new communication protocols.

The fourth generations are the networkers, the grandparents, the businessmen. They only join because they feel left behind if they wouldn't. They expect to find the same modalities as in the real world, they don't really understand the advantage of having a virtual world. They use the virtual world for replacing real world technologies they already now. They have a hard time enjoying the virtual world, and they'd prefer if it wouldn't exist at all. It is the first group that includes larger numbers of old people, thus bringing tradition and history to the virtual world. They expect to be treated with respect and politeness, and they always search for people they already know. For learning new possibilities in the virtual world, they mostly rely on personal instruction from members of the other groups.

The networkers can't distinguish geeks, techs and mods. They consider them all to be young, fast, always on the latest trends, and ignorant of the slow, the ones that need help. The geeks don't really notice the networkers, except that they keep bugging them with questions, which answers they won't understand anyway. They prefer to ignore their existence. The techs feel challenged by this group, because they really want to help them, but they have the hardest time making any progress. On the other hand they see the networks as a group who finally recognizes all the work they put into the safety they try to provide. The mods have the feeling that their space is being invaded by a group, who they actually tried to avoid having around, but they admit that they can learn something from the networkers. Where they all come together is, when it comes to transfering good concepts of the real world into the virtual world.

The fifth generation are the creatives, the graduates, the performers, the mac-users. They feel completely independent, and think of the virtual world not as a world, but just one more place they can use for showing what they create, or not. They see the advantages the virtual world offers, but they don't feel like they can be a permanent part of it. If what they find isn't what they've been looking for, they keep looking for it somewhere else. It is the first group that includes larger numbers of artists, thus bringing art and creativity to the virtual world. They won't admit that they don't know the newest developments in the virtual world, but they see the fault for that with the other groups not providing the newest informations on a silver plate. They want the virtual world to be a perfect world, without all the restrictions and problems of the real world.

The creatives don't care about distinguishing the other groups. They think of them as the audience, the ones that will applaud their work. If they don't applaud, they consider them to be ignorant of art. The geeks like to have the creatives around, because they feel entertained by what they do. The techs feel at a loss with the creatives, because they randomly stick to the rules or not, whether it suits them for their purposes or not. Security breaches suddenly can become a statement, to which the techs don't have a good answer. The mods are happy that the virtual world gained whole new user groups and content to gossip about. The networkers are happy that they are not the last ones to join new technologies after all, and are proud to help the creatives as often as possible, thus trying to gain respect from them. The place where all five groups come together, are large projects that extend to the virtual and the real world equally.

This is my view, on how the online community can be categorized and understood. These categories aren't fixed, everybody can change his position over time. They are also stereotypes, noone will be 100% of one category, but more a mixture of all of em, with different emphasis.

Some other perspective on internet sociology can be viewed in an article by Chris Short in this PDF.

maslow and the big five

2007-07-19T08:22:00.000-07:00

As i can't get the hierarchy of needs from the first post of my blog out of my mind, i gave some more thoughts to the meaning of it.

Psychologist have tried to factorize human personality since the beginning. Over decades of surveying and factoring and theorizing, a system called the Big Five has evolved as the most accurate to describe human personality. The factors are Neuroticism, Extraversion, Agreeableness, Conscientiousness, and Openness to Experience. Well, five personality factors, five levels of need, you do the math, i do the explanation:

The first level of need "physiological needs" corresponds to the personality dimension of Extraversion. Extraversion is the need to show that you exist, who you are, what you do. This might be the hardest of the five correspondences to understand, but if you look at the definition of personality, you might understand:
"A set of qualities that make a person distinct from another." --Wiktionary
Personality is, what makes you different from others within a society, but you only can be part of a society, if you act within the society, that means interacting with other human beings, that means showing that you exist, who you are and what you do, that means extraversion.
So if you have a very high level of physiological needs, you're trying to be a media star, you need the audience, you love to consume, you want to be spoiled, you're a Paris Hilton.
On the contrary if you have a very low level of physiological needs, you avoid the public attention, you're self-centered, consume means nothing to you, you take care of yourself, you're ascetic, you're a Benedict XVI..

The second level of need "safety" corresponds to the dimension of Conscientiousness. Conscientiousness is all about rules, control, plans, details, perfection. You achieve to satisfy the need for safety, by being planning your life, following rules, pay attention to details, by being conscientious.
A very high level of need for safety creates personalities, that are bureaucrats, control freaks, police state dictators, they are Adolf Hitlers.
A very low level of safety needs creates personalities, that are anarchists, nomads, inordinates, they are Paul Erdőss.

The next correspondence is between the need for love and belonging and the personality dimension of Agreeableness. Agreeableness is about taking interest in the other human beings of your society, to help them, to build a community, to symphatize and be concerned. This is what family life, friendships and partnerships are based on, and this is where we find love and belonging.
So if you have a very high level of need for love and belonging, you're a helper, an altruist, the heart of the community, the housewife, you got a dog personality, you're a Mother Therese.
If you have a very low level of need for love and belonging, you're individualistic, independent, you value the world above society, you have a cat personality, you're a Socrates.

The key personality dimension for the need of esteem would be Neuroticism. The key for this connection is the anxiety of neurotics. If you are anxious, you fear that you're not good enough in comparison to others. Being good enough in the eyes of others is esteem.
A high level of need of esteem is present in persons that lack confidence, that are easily disturbed, that get upset, like Michael Jacksons.
A low level of need of esteem is part of personalities that have no doubt on what they do, stand firm as a rock, don't see problems but challenges, are quite narcisstic, they are Diego Maradonas.

The fifth and highest level of needs, the need for self-actualization, is corresponding to the personality dimension of Openness to Experience. This dimension is all about ideas, creativity, imagination, beauty, etc. while self-actualizatio is about making the best of your abilities. These two fit just perfectly.
If you have a high level of need for self-actualization, you are constantly creating something, ideas, art, inventions, you're a Leonardo da Vinci.
If your level of need for self-actualization is very low, you're very conventional, you prefer the concrete over the abstract, the straightforward over the subtle, continuity over change. There is no role model for this, because the highest level of need isn't a deficiency need, but a growth need. If you don't need to self-actualize, you will just go unnoticed, as millions of others have before.

So by applying this correspondency, if you find out what your levels of needs are, you know what kind of personality you have, and the other way round. Don't confuse the levels of needs with the amount to which the needs are fulfilled. They are probably never fulfilled all at the same time, that's the drive that keeps us moving every day. The bigger the discrepancy between the level of need and the actual level of fulfillment, the more problems you have, and the unhappier you are. Large gaps can lead to suicide, or fundamentally life-changing decisions. The levels of needs can be lowered or heightened over time, but it's mostly a slow process. Pharma-products also have a temporary influence on these levels.

P.S.:While writing this article I stumbled across a paper by Francis Heylighen, who stated in 1992 a very similar system as I just did: pespmc1.vub.ac.be/papers/Maslow.pdf

wikipedia-enhanced translation aids

2007-07-16T17:12:00.000-07:00

Everybody asks me, what you can do with general linguistics after finishing degree. The first thing that always comes to my mind, is working at a company for translation software. But the problem is, there is no good software out there yet, due to the fact that no software has the world knowledge a human being has. So whatever these programs produce, we've all seen it at babelfish - it looks like some underpaid chinese worker without any knowledge of the involved languages did it, just with the help of dictionaries. Nevertheless, if you ask anyone in the software industry or many other industries as well, if machine translation would achieve level where it could actually be useful, it would be a major success, and have influence on millions of jobs and companies.

But how do we teach a software, what we as human beings had to learn for years and decades? Do we raise an AI in a family, so it has the same experiences as we have? Or do we hire some more cheap workers to put everything into the software it needs to know? Well, for me it's quite obvious that both approaches will fail. But why don't we use the digital ressources we already got? I'm not talking about translation dictionaries, but rather about wikipedia, with all it's facts, language data, categories and interwiki-links, or wiktionary, a major source for word categories, phonology and morphology. I'm not saying this is all you need to learn a language, but these ressources could help with some of the biggest problems todays translation software faces - like ambigous words, figures of speech, analogies, etc. In combination with state of the art grammar theories, you could at least build some excellent translation aid software, but even if you try the full step to independent translation, it probably would produce some notably better results than babelfish or other attempts of the last years.

In my opinion the main reason why no good translation software has sprung up so far, is based in the nature of english language. Most successful software we know has been created in english-speaking countries. But english is a language that is very creative in using a single word in hundreds of contexts and meanings. For humans it is very easy to grasp, because even with a small dictionary you can express many things, by using transfers and analogies of the words you already know. For software it is very hard, because it can't _understand_ the words. Many try to counter this difficulty by using statistics about contexts, thus generating a statistical substitute for understanding, but then you rely on a corpus. Due to these difficulties, english and other languages that behave like this shouldn't be the first choice for attempts at creating translation software, but what you rather want to have is languages with a large dictionary base, few ambiguities and few changes over time.

The most important development for a translation aid software project is the semantic wiki. Because if you have a multilingual encyclopedia of our world with computer-readable content, it would be a big help for translating even complicated constructions. But for this more than a decade will pass, unless a large bunch of new general linguists show up soon and get some serious work going.