Anjo, in What is a topic?

The most mysterious term that I encountered a lot recently is topic. I have no idea how to define it and, neither seem the weblog research proposals that suggest finding the topic of a post is something worth doing. Being on holiday currently, and given it was raining and snowing outside, I tried to apply the notion of “topic finding” to weblog conversations (see also: here, and here).

Anjo goes on, providing an example of “unique” terms extracted from three weblog conversations (more details in the post). Although those provide a good picture of what conversations are about, they do not really answer the question of what is a topic of each of them.

Which makes me thinking of my own experiences around the issue…

One of the things we planned to do this year, but didn’t get to do, was looking at personal categorisations. To be more specific the idea was to compare categories (~tags, ~topics) that a blogger assigns to her posts and the results of the text analysis of those posts to see if there is any correlation between the language used and conceptual categories. [I still think it's an experiment worth doing, but not sure I personally can devote serious time to it. Anyone interested?]

Thinking of my own weblog I can imagine that for some topics (I call them topics ;) that I use for my own weblog the correlation should be present (e.g. posts related to events are likely to be labelled with it and mention it in the text).

However there are others, those where I assign topic to organise my ideas on ill-structured themes (=I feel that those posts belong together, but I don’t know why yet, or I don’t have a good label for it). The examples of the second type are posts on life, knowledge mapping or transparency.

Which brings me to the reason I started to write this post. I think that topics are conceptual categories used to characterise a group of connected pieces (conversations with others, conversations with self, or something in between) and to give it a nametag. The common name makes sense – it makes it easier to remember those pieces belong together, to retrive, to communicate about.

The problem is that conceptual categories are subjective. They depend on a person, group or even groupthink (as with pressure to use certain tags to appear at right places in Technorati and not because they make more sense than others). So I suspect that once we define a topic of a conversation there will be someone who would say that it’s about something else (referring to Anjo’s examples – it could be “not about Skype, but about presence”).

That’s said I still think that defining a topic of a conversation makes sense. Personally, I’d prefer to have a Sigmund picture (~frequent terms and relations between them) for a conversation, as some kind of ontological fingerprint of what the conversation is about. Or there is a number of ways to select one of the terms from the “unique term list” for a conversation:

• by further selecting “least unique” from the subset (i.e. terms used by highest number of participants of the conversation)
• by selecting terms that match categories some of participants assign to posts
• by selecting terms that match predefined ontology/folksonomy/keyword list
• by selecting terms most of the participants are likely to agree (don’t ask me how to do that :)
• by selecting terms most closely resembling those of an external “customer” for the analysis or those that non-participant is likely to understand

Or we just have to find a way of matching personal caterogisations. Given there the tools are going this shouldn’t be that far…

Archived version of this entry is available at http://blog.mathemagenic.com/2005/12/20.html#a1717; comments are here.

• Me, speaking on researching blogs and blogging research, Amsterdam, 21 June
• What I want to do when I’m done with my PhD
• Weblog audience: how to you find your own?

Just a quote from danah boyd:

Our lives have started to escape categories.

The post is actually about experiences on topical vs. personal blogging. Jack Vinson has a follow-up on shifting interests, forgetting and categorisation.

I’m thinking of some examples from my study at Microsoft, but this will have to wait (mainly because work is hard as I’m less than 24 hours away from my long awaited vacation :)

Archived version of this entry is available at http://blog.mathemagenic.com/2005/10/21.html#a1700; comments are here.

• Why blogging 2
• Trip report (3): knowledge/community mapping
• Hypertext writing

Anjo documents the experiment of creating a cooking ontology by running smart tools through the content of Chocolate and Zucchini:

• From Weblogs to Ontologies
• Cooking: Creating a Corpus
• Cooking: Utensils – Anjo, did you try correlating terms with (explicit) categories of the weblog to define higher level concepts? E. Dehillerin and knifes have something to do with Kitchen Paraphernalia and Paris City Guide…

Curious to know what Sigmund would say :)

Archived version of this entry is available at http://blog.mathemagenic.com/2005/08/30.html#a1650; comments are here.

• Searching for knowledge as constructing personal learning experience
• NOAGGREGATE: what if I don’t want my digital bits to be connected at one place?
• Tag-based bookmarking in a browser and why people need bookmarking

Anjo gives a bit of insight into our internal discussions on uses of ontologies:

Andy Boyd came up with a wonderful new term: “ontological fingerprinting” and to illustrate how imaginative he is: zero hits on Google! Suppose one has an ontology (lexicon, thesaurus) and some software that can determine whether the terms in the ontology are present in a document. Applying the software, one gets a “fingerprint” of the concepts in the ontology for a given document. Comparing fingerprints for different documents, such is the assumption, provides a better metric of the similarity between these documents than comparing plain words. Ideas like this simply have to be tested in practice. Fortunately, Andy is making available a lot of real data to try it.

I like the term, but find it a bit misleading: usually documents do not have fingers :)

I’d associate the term with people – you may think of “ontological fingerprint” of a person, which could be something like conceptualisations produced by Sigmund based on analysis of weblog posts written by someone, set of personal categories someone uses to classify a document or mapping one’s documents to a shared ontology. Then you can look for others with similar “fingerprints” (this was one of uses I imagined for Sigmund, but didn’t have such a nice term to talk about it :).

May be we should rather talk about “ontological abstract” in case of documents…

Archived version of this entry is available at http://blog.mathemagenic.com/2005/01/28.html#a1493; comments are here.

• Knowledge workers time spent finding information
• Feed your blog to tOKo and see what comes out
• NOAGGREGATE: what if I don’t want my digital bits to be connected at one place?

As I still have work to do I’m not going to add relevant links for the on-going discussions on folksonomies, related tools and social implications, but just link to a few things.

Bits of research:

• Bookmark, Classify and Share: A mini-ethnography of social practices in a distributed classification community
• Folksonomies – Cooperative Classification and Communication Through Shared Metadata

Visualising del.icio.us:

• “related tag” network visualisation
• del.icio.us subscriptions as social networks

Regarding the last one:

• Wonder how many people actually look at their del.icio.us subscriptions (I don’t)
• Thought that it would be fun visualising social networks based on similar linking patterns, but discovered that that particular feature is not in del.icio.us anymore. Given how often I used it as an argument it’s pretty stupid not having screenshots.

Archived version of this entry is available at http://blog.mathemagenic.com/2005/01/26.html#a1489; comments are here.

• Combining data from various devices and privacy
• Lack of project data accessibility study
• Lessons learnt implementing expertise locator system

Had an interesting discussion with Anjo and Rogier on how and why people categorise things (documents, bookmarks, blogposts)… Thinking of all kinds of things that I’d like to know about my own categorisation:

• Why I categorise things the way I do? What are the criteria?
• When I categorise things for further retrieval, how often I actually go and retrieve them? Does categorisation helps in it?
• How categories evolve over time? In relation to: changes in thinking, changes in tasks, changes in tools I use?
  • e.g. did categorisation in my weblog changed since I started to use del.icio.us
• How categorisations in different spaces (e.g. file folders, email folders/tags, paper files, weblog topics)overlap? What explains overlaps? How you could connect them? Is there a need to connect them?

Just a brain dump…

See also: LiveTopics wishlist or topic-based blogging support on how I categorise things in my weblog and what features I miss.

Archived version of this entry is available at http://blog.mathemagenic.com/2004/12/09.html#a1451; comments are here.

• LiveTopics wishlist or topic-based blogging support
• Experimenting with creating an ontology based on weblog content
• Trees vs. webs

I shouldn’t be quoting that much, but I guess you can forgive someone with a trojan flu for lack of summarisation :)

Martin Dugage:

In a great post, Metadata for the masses (via Many-to-Many), Peter Merholz advocates free tagging of documents as opposed to choosing tags from inflexible top-down classification systems, which often force users to view the world in potentially unfamiliar ways. I tend to agree with that from my experience of taxonomies, which often become obsolete within two years. Then Peter writes about the limitations of free tagging:

Clearly, such tagging systems are not a panacea; they present many potential drawbacks. With no one controlling the vocabulary, users develop multiple terms for identical concepts. For example, if you want to find all references to New York City on Del.icio.us, you’ll have to look through “nyc,” “newyork,” and “newyorkcity.”

That’s were ontologies come in handy. They give a community of people the ability to develop a common meta-classification model that sits on top of existing ones and bridges them together. An ontology can define “nyc”, “newyork” and “newyorkcity” as synonyms, define “Time Square” as included in “nyc” etc. See for example Arisem, who are doing a good job there. In a sense, ontologies allow communities to build a common language from the ground up, which is essential in knowledge creating environment. Top-down norms can be introduced later, when language can be “industrialized” for larger communities.

I have mixed feelings about ontologies… From one side, they could be really useful. Next to connecting metadata from different people or communities ontologies could be used to connect data from different systems (e.g. your weblog and del.icio.us tags). And, of course, one could imagine all kinds of great things with inference rules (always wonder if I picked up the right meaning of the term from my colleagues ;).

From another side, ontology development could be complex and costly, so I’m always wondering if it’s worth it.

This doesn’t mean that we can’t think of ways to support converging metadata and emerging ontologies.

Converging metadata could be a result of social pressure.

It can be done as in k-collector where people can select from the list of existing topics (their own and others) or add their own. Or as in nutr.itio.us by choosing del.icio.us tags of others for a link you are about to add.

This approach can help cases like nyc/newyork/newyorkcity, where we deal with different ways of writing the same tag. For example, after finding out that majority of del.icio.us users used visualization and not visualisation I changed s into z in my own tags.

Next to it we can think of emergent ontologies.

Ages ago Liz Lawley described how relations you can do it with del.icio.us:

Add a site to your del.icio.us bookmarks, and then look to see who else has added it. What descriptive tags did they use for it? As an example, here are the current links to Metacrap in the del.icio.us system. I used the terms metadata and semweb. Other terms used include taxonomy, ontology, ia, humanFactor, and xml.

Funny enough, del.icio.us does it now automatically, but you can see it only for some posts, so I was surprised by the discovery.

So, check my links on aggregation, middle column, lower part – it shows other users who bookmarked same links and tags they used (rss, tool_rss, feed, meta, syndication, atom).

Of course, you need a critical mass of tagged links to do it and, of course, quality varies, but just think how much you can do with something like this. For example:

• have “raw” ontologies that could be good enough for some purposes and reworked manually for those where more precision is needed
• do “translations” of tags (e.g. my ‘aggregation’ is your ‘rss’)
• find like-minded people
• find areas of agreement (shared language/tagging) and disagreement (totally different tags) in a community

Side note: as an alternative to folksonomies, from Peter Merholz

Ethnoclassification, to the best of my knowledge, was coined by Susan Leigh Star for her Digital Libraries conference workshop “Slouching Toward Infrastructure.”

See also:

• Classifications for archiving, search and retrieval
• Distributed KM
• and other stuff under knowledge mapping or ontologies

Archived version of this entry is available at http://blog.mathemagenic.com/2004/10/26.html#a1401; comments are here.

• The Dark Side of Numbers: The Role of Population Data Systems in Human Rights Abuses
• The orange pill: Matt and Paolo are getting ready for BlogTalk
• You just don’t know what you’ll want to know down the road

Tom Coates on tag-based bookmarks in browsers (via Brian Dennis) – applying del.icio.us tag-bookmarking model to a browser:

To summarise the problems with current bookmarking systems then, we could say that (1) the process is slow and annoying (2) that it requires us to continually refine and redevelop our taxonomies if we’re going to keep track of everything, (3) that URLs can belong in a number of bins and that (4) we can be left with unmanageably large lists. An ideal system would therefore speed the process up of both bookmarking a site and retrieving it later. An ideal system would try to alleviate the problems of categorisation and would work as an a priori assumption that a URL might wish to be stored in multiple bins. An ideal system would not display all the links by default. An ideal system would, in fact, use tags…

Tom provides a whole scenario and mock-ups of tag-based bookmarks in a browsers… What I’d add is that we need to think broader than easy classification, bookmarking and finding. The study I blogged almost a year ago (Keeping found things found on the web) suggest that when deciding what to do with interesting web-pages people may think about:

• Portability – being able to take it with you
• Number of access points – being able to access information from different locations
• Preservation of information in its current state
• Currency of information – having updated version of information
• Context - remembering why it was saved
• Reminding – remembering that something has to be done with it
• Ease of integration into existing structures (e.g. e-mail with link can be easily archived with other e-mails, while bookmarks have their own structure)
• Communication and information sharing
• Ease of maintenance

From this perspective del.icio.us would be more powerful than tagged bookmarks in a browser :)

And, have no idea how I missed folksonomy as a new term for it :)

See also (re: more things to do with tags): LiveTopics wishlist or topic-based blogging support

Archived version of this entry is available at http://blog.mathemagenic.com/2004/10/18.html#a1392; comments are here.

• Implicit Structure and the Dynamics of Blogspace and more papers from HP Information Dynamics Lab
• The high cost of not finding information: Reinventing is more fun than reusing
• My del.icio.us tagroll

One of the directions that keep on popping up when I’m thinking about blogging in KM context is topic-based blogging. There are a couple of reasons behind it:

• personal - if blogs are used as a personal knowledge management tool than ability to tag posts is important to be able to organise, retrieve and share them
• corporate – once weblogs are used in a company one would want to be able to slice an aggregated stream of posts into topic-based streams to support knowledge sharing

liveTopics and k-collector are good examples of personal vs. corporate implementations (see also: liveTopics and k-collector compared) and del.icio.us is an example of connecting personal and shared views on topics together.

In this post I’d like to focus on personal side and describe what topic-based blogging functionalities one may want as a blogger. And because I’m very practical and selfish I’d describe it as my liveTopics wishlist :)

What liveTopics do now

• allow adding topics for every weblog posts – e.g. check this post via browser
• display a list of posts per topic – e.g. my posts about liveTopics
• display topics as a frequency list or recently updated list – e.g. my topic index
• provide an interface for managing topics (renaming/deleting + backup + some settings)

My liveTopics wishlist

• Printing
  • I’d like to be able to print posts for a topic or combination of topics (so far I can think of AND/OR combinations, but may be I’d want more once related topics are there ;)
• Aggregation
  • Topics indication in my RSS feed (e.g. in ENT format)
  • RSS feed for each topic (ideally for a combination of topics as well :)
• Related topics
  • Indication of a relation (below are different overlapping! dimentions)
    • Automatic vs. manual
      • Automatic: based on co-occurences in the same post – e.g. related tags for innovation at k-collector or del.icio.us
      • Manual: added by me to indicate how topics are related in my head – e.g. I’d like to be able to specify that blog communities is a narrower than communities, but also related to blog ecosystem
    • Internal vs. external
      • Showing that my posts on blog communities are related to my del.icio.us link collection on blogCommunities
    • Me vs. others
      • Showing that my posts on innovation are related to post on innovation at k-collector
  • Inference
    • Once relations are there one can do all nice things with inference rules.
      • For example automatically including narrower topics when broader topics are selected, e.g. when someone selects communities posts on blog communities are shown as well even if they are not explicitly marked that way
• Visualising
  • Dynamics
    • I’d love to see how my use of topics changes over time; something similar to temporal rhythms of interactions with others visualised from email arhives
  • Relations between topics
    • Just a nice image :) Something like delicious mind
• Small liveTopics/Radio specific things
  • Shortcuts for topics added automatically
  • Expanding of posts by topic in topic index pages

Of course, I wonder how many of those things are “nice to have”/”Lilia specific” and which features would be used by many blogger, but this is a “further research direction” as I’d write in a paper :)

Archived version of this entry is available at http://blog.mathemagenic.com/2004/10/13.html#a1384; comments are here.

• The high cost of not finding information: Reinventing is more fun than reusing
• Hard lessons learnt
• Topic-based RSS feeds from K-Collector

I knew that Anjo would get interested in Presidential Debate Analysis by Cameron Marlow (via Alex), but didn’t know that he would go this far given that today is Saturday :)

Sigmund on the US Presidential Debate:

I have applied Sigmund to the same transcript. Sigmund is a kind of electronic shrink that uses language technology to derive conceptualisations found in text. The tool is normally applied to collections of postings, for example weblogs or emails. For this exercise I considered each 90 second statement by the two debaters to be a single post.

Archived version of this entry is available at http://blog.mathemagenic.com/2004/10/02.html#a1371; comments are here.

• Experimenting with creating an ontology based on weblog content
• WBC04: day 2 morning
• Personal ways of doing things in public