I was recently asked what my ideal research tool would be. Here are some of the high level ideas I jotted down.

• My ideal research tool would be platform independent. It could very easily switch from my Mac laptop to the Windows PC at the library to my mobile phone, camera or pen scanner.
• It would allow me to very easily highlight or scan the UID (isbn, issn, QR code, bar code - whatever) of the object or class of objects that are the source material. ie, I've got a book that I'm taking notes from, and rather than copying down all of the pertinent citation information, I can scan the bar code and have it added to my personal datastore for me.
• I would be able to clip text (again, highlight or scan) and scrape like data automatically. Adding the resource type would be as simple as selecting a button. eg, I'm researching a person, I highlight their name, and click an icon depicting a headshot, thereby adding the text string to my datastore with the type "person." Over time the system would learn and suggest the right types.
• I could do the same thing with images. Images though could also be tagged quite granularly - I could draw a shape around an object of interest in the image, and tag that.
• All of the system (creator, publisher, software, hardware) metadata would be transferred to my datastore along with the research I choose to add.
• Built in NLP and analytics would give me a leg up on entity extraction, coding and clustering - but only at a high degree of confidence. I'd rather the system ask for my input than guess and mis-categorize something. So that workflow needs to be quick and painless.
• Internet based resources would be scoured for me based on my trust scores of said resources (or my trust in people who trust those resources.)
• Related data from the internet would be returned and organized with a "follow up" indicator. This data would be organized along facets such as electronic resource, physical resource (ie - I'd have to go to it or have it sent to me), fee or free, and all usual descriptive, system and administrative metadata provided.
• Experts in the field would be identified, along with the "proof" of their expertise. My connections to these experts would be delineated to assist with making contact. Their identification should include contact information, web presence, bibliographies and the like.
• It would include image detection - akin to TinEye and/or SnapTell.
• Audio and video should be allowed in the system. Again with the same capture, tag, type, and detection capabilities (Shazam).
• Collaboration capabilities are required. I should be able to set granular permissions at the object level, and easily integrate with any contact network - address book, social network, closed network.
• I should be able to make notes in the system on the data I'm capturing. I should be able to make generic notes, or make notes on a specific set of captured data - a text string, or portion of an image.
• I should be able to make links between objects of any kind, and type those links, and flag them for follow up.
• The datastore should be backed up regularly and automatically.
• I should be able to import tags/schema/ontologies of any kind as well as create my own.
• I should be able to publish any set or subset of my research as linked data with appropriate styling for the human user.

In case you were wondering, I do love Zotero and Zotz, and hope development on those continue. :)

I recently had the opportunity to think about the future of the browser. It was pointed out that presuming it should be a browser was limiting, and so the term user-agent was substituted. What does that mean? It refers to the means by which humans interact with the data, tools and experiences available to them via the internet.

One of the problems with promoting linked open data on the web is the "what's in it for me" question. Most people understand the value of sharing, but most people are too impatient to wait for shared data - of interest to them - to show up to receive value-in-kind. This is not particularly motivating for putting in even the smallest effort at formatting data in a re-usable way. We are a society that wants instant gratification. Frustrating, but all too often it is the case.

As for myself, I want my user-agent to be a memory palace. For a couple of years now I've been poking at the idea of a semantic memory palace. Why? It benefits me and those I share my data with. It helps me to become a more organized, more efficient and more astute person. I can gather data of interest, as I find it, organize it in a way that I personally can use it, and share my tags, provenance information and mashups with whomever I wish. I can use it to learn and re-enforce the memory palace in my head. I can use it, as the Roman orators did, to help deliver more compelling presentations. I can use it in reports, in games, in visualizations. I can learn about the differences between how I think about the world, and how others do; improving my ability to communicate through understanding, opening my mind to new ideas and possibilities thanks to diversity of thought. Simply put, I can improve my memory; halt and hopefully reverse the decay that began when my kids came along!

What is a memory palace? It's a technique for remembering things better, by using visualizations and mnemonic devices. (If you want an excellent study on the topic, check out Frances Yates' book Art of Memory.) First, you choose a "palace" - a home for your memories. It can be a real or imagined, but should be a place you know well. You "place" things to remember in the rooms of your palace, and as you walk from room to room you can collect them in the order you want to remember them in. So, let's say you have a list of groceries - bread, butter, milk - and your palace is your childhood home. At your childhood home, you always entered the side door through a porch, and then into a kitchen, which led into a family room. You could place your grocery list like this: the bread in the porch, the butter in the kitchen, the milk in the family room. Rules for creating palaces can be found by searching for articles on the topic "method of loci." Next you can create a stable list of visualizations for yourself that represent letters and numbers. The image of an apple to represent "A," the image or a bird to represent "B," or a line, vesica piscis, triangle, square etc. to represent the numbers 1, 2, 3, 4. Unusual combinations are more memorable as they invoke our emotions, so to help us remember our grocery list we might place the image of a bird, carrying a vesica piscis with its feet and clutching a stick of butter in its beak in our kitchen. This way we can remember we need two sticks of butter, and it's the second thing on our list. This is a gross oversimplification of the technique; hopefully you get the idea!

So what do I want my user-agent to do? I want to be able to set my own images - folk tag data in my [browser/repository] to one of these categories. I want to be able to drag and drop the data into a room in my palace, and create as many instances of my palace as I need. I want to be able to recall the data at any time, by palace, by room, by image. I want it to capture provenance data automatically (source, date/time). I want it to let me add keyword tags if I so desire, or choose a schema to add other kinds of data - say Dublin Core or FOAF, as needed. I want to tag with sounds - imagine if Shazam could also identify bird calls, or my friend's voices. I want to be able to very easily state that for me a block of red color means that I am tagging something as being 'energetic.' Eventually I'd love to be able to tag with smells as well; for now simply representing with text strings will have to do. My palettes of tags could easily become large; so a method for organizing, scanning/finding and shortcuts for applying oft-used ones would be needed. And I'd want each tag to be represented persistently, so the tool would need to be able to manage a namespace or integrate with one. I see bits and pieces of these capabilities all over the technology landscape: Zotero & Zotz, Shazam, SnapNow & Flickr, countless Firefox plugins. I've captured some of the crazier ideas Kevin and I have had in the post on multi-sensory tagging.

Of course I want my agent to do other more practical things! This is the one that I'm most passionate about at the moment. Now, if only I had learned to code... No matter, I'm working on formalizing my use cases and requirements notes. Maybe someone will build it someday!

[Note: making a private thread public.]

we could create art projects out of this multi-sensory tagging

[Note: making a private thread public.]

I want to be able to have things appear on the frequency required to program them into memory without having to sit-down and rote memorize them. Also need to consider doing this for advertising - a system that can identify the concepts on a page/site that can then match the right advertiser is cool - but to do it across sites, using the same relevant ad to get more eyeballs and get it in the users memory quicker would be very cool.

[Making an old, private thread public.]

I need a browser tool. A user-agent. It needs to be able to allow me to create triples wherever I may be on the web. It should let me indicate the S-P-O. It should let me say if an entity is a person or software or a building. It should automatically link the entity type to a URI from schema that I set up. [Need Subject-Predicate-Object-Timestamp, actually.]

[Note: making a private thread public.]

This is fuzzy still, not sure where I'm going with it. Why do we store multiple copies of the same message in multiple places? Now we can store one copy, and pull the recipients to it instead.

I'm sure someone else has thought of this; I'm just reading something that made me think of it.

Not just for email; we need to change the UI/UX for data storage.

Sparked by this post I just got around to: http://kubasik.net/blog/2008/05/21/the-reality-of-semantic-desktops-deat...

[Note: making a private thread public.]

I want to be able to tag things by virtual representations of sensory inputs: colors, textures, sounds, tastes, smells, emotions. Need to be able to do on mobile as well to incorporate real-world experiences and tag them this way at the time of their happening to capture the full essence/context.

Would also be cool to add things like the lat/long (GPS data), weather-data for where/when a pic was taken.

Could be neat to add info on "what's playing" to a tagged item. (Shazam? Verizon or Blackberry song ID tools?)

Professional organizations ought to provide OpenID services. Who cares if LinkedIn, Yahoo, AOL or MyOpenID say I am who I say I am? If I have to get licensed somewhere, or if I need to join a certain group to have professional credibility, than that org ought to support me by saying, "Yes, this person has earned a measure of trust that we can confirm."

I am concerned about potential trends: 1) towards pay-for-verification services in the guise of focused subscription directories and 2) 'services to the community' provided by self-serving entrepreneurs. We need true sources of authority, not just technically consistent sources. I think my professional groups should use open identification tools/software with clear ethics policies/responsibilities on each side to help me establish my own brand.

[NB: I originally started this post here on my personal site, but ended up publishing it to my work blog over at Synaptica Central. I wanted to keep a copy though.]

...There seems to be some discussion about whether or not data is a commodity. I think there IS most definitely data that are a commodity.

Taxonomies are a valuable raw material in the management of information. A file that can be bought and sold and used to improve services. They can be generated by humans, machines, or even better: humans working with machines. Many taxonomies are a dime a dozen, with little to differentiate between versions of the same data. Some are like Kopi Luwak coffee - rare and extremely valuable. The word "taxonomy" is itself suffering from a kind of genericide. Classical definitions still apply: taxonomies have become commoditized.

The complexity of the controlled vocabulary will determine its value to a degree. A simple pick list should be easy and cheap to acquire - a list of countries, for example. Or colors, seasons, months - you get the idea. What is the value of a list of industries? Or companies? Maintenance is the primary cost factor - frequent changes require frequent updates, but an authority file in and of itself is not that complex. A broad and deep poly-hierarchical taxonomy I would expect to have more value. A poly-hierarchical taxonomy is one where a term in the taxonomy can have more than one parent term. Managing these relationships takes more time. An ontology - well, those aren't quite commodities yet, but they will get there. Why? Because they still require a great deal of thought and effort.

The source of the data will also help determine its value. Data from trusted sources - for whom integrity is paramount - should be valued higher. Is the data accurate? Is it maintained? Is it in a usable format? Does it have high availability? (Many quality vendors can be found at TaxonomyWarehouse.com.)

The uniqueness of the taxonomy will drive its value. Like our coffee example above, a taxonomy as ubiquitous as Starbucks will not be as valuable as say a pharmaceutical research vocabulary. Given the, uh, processes needed to produce Kopi Luwak, it is rare and therefore fetches a higher price, as would our R&D taxonomy.

The information security concerns also impact value. Our pharmaceutical company, or a financial services provider, is not about to release it's vocabulary into the wild. It is a significant intellectual asset that merits a substantial IT effort to protect.

I actually like the fact that taxonomies have become commoditized. Why? Competition drives improvement - in quality, in focus, in security and in usability. These are areas that the semantic web community needs to focus on - in my experience, security and usability need attention NOW. Good fences make good neighbors, and when we've got good fences, we can make more links and learn to trust. Icing on the cake!

Flickr image by INeedCoffee

I've been pondering this for a few days and wanted to put it out there for feedback. It's full of holes and fairy dust, but I'd love to hear what people think.

I was looking at an RDF graph visualization the other day - colorful linked clusters, like rigid versions of the pom poms I made as a kid. Clusters of related data, with occasional links to other clusters of related data, and other clusters nearby. Visual Complexity has a wonderful set of images representing data in this fashion. Reminded of why we librarians labor to determine the best shelving locations for books - so that similar topics can be scanned easily.