Choosing to drive everything from Ping.fm has many advantages and few disadvantages.  Ping has a wide selection of integrations and covers all the social networks I use, so I can use my tweets as status updates.  Ping lets me create “groups” so I can do both personal and professional tweets from a common interface, and share selectively.  I have a FaceBook page that I haven’t had time to do much with, so streaming status updates from Ping lets me reach my friends there when I’m busy.  There are social networks that I use passively like Plaxo and FriendFeed, and Ping is a great way for me to reach friends and colleagues there, too.

Ping also lets me submit via mobile technology, whether I use the browser, email, SMS, or chat client on my BlackBerry.  I like submitting through the BeeJive’s JiveTalk IM client, but have to count my keystrokes manually, which is a drag.  I’ve actually tried every method except SMS with Ping.  Email posting has truncated messages in the past so I avoid it (although that was the same day Twitter was melting down because of the Iranian elections so I’ll have to try again someday).  None of the browsers I’ve tried (including Bolt) will run the script to count my keystrokes as I type into the Ping submission form, and logging in to get to the posting pages is more of a hassle than using an IM client.

I’ve also tried originating messages on HootSuite and relaying through Ping, but I’ve had some issues (hence the Xin the schematic).  For example, I like to use [category tags] in square brackets at the lead of my tweets (more on why below), but if I post to HootSuite and relay through the Ping API the brackets and everything in between gets stripped.  I think the issue is at the Ping API since the same posts arrive intact on Twitter. 

The primary pathway for integrating status updates with the website itself is Twitter (hence the red arrow).  Twitter Tools uses the feed from @twitcontentguy to populate the sidebar on the main page, the headlines section on the Headlines page, and a Weekly Digest blog posting every Friday morning.  I decided not to automatically have Twitter Tools tweet every time my blog got updated for several reasons.  First, as Mike Axelrod and I discussed, there is a stylistic difference between a headline, a blog post title, and a tweet of either of these, so by manually tweeting the URL of the blog post I get the editorial control I need.  Second, Twitter Tools would only update Twitter, not the other social networks.  Third, there was a bit of an echo chamber effect, since the post would show up in the blog, the sidebar, and the headlines page with the same title. As it is, my blog posts have a notification stream of their own, via the RSS feed.  Blog posts show up in LinkedIn using the WordPress app, as well as on Plaxo and The Content Wrangler community site.  Of course the blog posts now include the Weekly Digests of my Twitter activity.

As for using Ping for micro-blogging, I can’t since I haven’t integrated the Ping custom URL on my blog due to technical issues.  I can do this now with my old WordPress.com blog (so I’ll do some testing there) but still need to get the custom URL working for my self-hosted blog.  Also, I haven’t found a good use case yet for micro-blogging; digesting the tweets weekly and in real-time with the Twitter Tools sidebar works great for getting the content to my blog.

2. How I Tweet. After four months I’ve settled into a better pattern of how I use and format my tweets.  There are several aspects I’ve thoughtfully considered: content, format, and frequency.

• Content. It’s pretty rare that I’ll share my personal status; when I do it’s usually because I’m participating in an event that others who follow the site care about, like a great webcast or speech.  In that event I’ll try to relay a quote or other tidbit from the event itself, not just my status and location.  More often I’m sharing content found elsewhere on the web.  This can be news or other relevant information.  Before I tweet about something, however, you can be sure I’ve actually read it and thought about it.
• Format.  Early on I adopted a format for my tweets : [category] @attribution: content link #hashtags
  • Category. I use [category tags] in square brackets at the start of most tweets, an idea that I borrowed from Scott Abel, the Content Wrangler.  These have the advantage of setting off my tweet from any prefix a site might attach to my post.  For example, LinkedIn affixes “Paul” in front of every status update, expecting us to write our status with a verb to follow our first name.  E.g. “Paul is out fishing.”  ”Paul just saw a UFO. Again.”  The square brackets buffer that, so it reads more like the character speaking in a script: “Paul [ECM] Big day of acquisitions in the ECM marketplace.“  They also make it easy for someone to scan the Weekly Digest for articles of interest. 
  • Attribution. Most of what I tweet is a found item of interest. In the attribution I’ll site the source and the individual quoted, whether it’s a blog, news channel, or a retweet.  E.g.
    [ECM] RT @dgschultz: EMC adds e-discovery with Kazeon buy http://bit.ly/6n1VV  or
    [ECM] via ReadWriteWeb: Gartner VP Toby Bell – Six ways to save money on ECM http://ping.fm/BPdx7
  • Content.  This would be the actual content portion of the tweet vs. the rest, which is meta.  Most of what makes it into my professional tweets is straight business (ECM news, events, etc.), but I do like to share the occasional funny bit or fishing success so that the tweets have some personality to them.  Stuff meant just for friends and family never hits @twitcontentguy; I have other personal twitter accounts for that, or I use Ping groups to selectively update my status.
  • Link.This is the shortened URL.  I’d guess that 90% or more of my tweets have a link.  Most often in the past I let Ping do the job of shortening the link.  Starting this week I’ll be using j.mp to shorten my URLs.  This is a new service from the great folks at Bit.ly. 
  • Hash tags.  I started tweeting shortly before hash tags took hold, so I was a bit slow to adopt.  I like putting them at the end of a post, as they fit.  Not everyone who reads my tweets is Twitterate, so the odd # in the middle of the post can be jarring.  E.g. compare these two posts for readability with mere mortals:
    [Webcast] DITA, Metadata Maturity & the Case for Taxonomy. Wed 9/2 1-2 EDT. http://ping.fm/m107W #dita #taxonomy #metadata
    [Webcast] #DITA, #Metadata Maturity & the Case for #Taxonomy. Wed 9/2 1-2 EDT. http://ping.fm/m107W
    Sure, the second one is more concise but I just plain don’t like the look of it. Update: Now letting Ping.fm filter out the hash tags from the services that don’t use them (like LinkedIn or FaceBook), so putting them in-line (judiciously).
• Frequency. If I’m doing well I tweet once or twice a day, giving me 5-10 items a week in the Digest.  I tend to tweet in spurts, because I tweet to share what I’m reading, and some days I’m in research mode and others not so much.  Using HootSuite has made reading those I follow more of a daily routine, so now RTs are more routine.  I still read the tweet streams I follow in batch mode, usually in the morning after email or over lunch.  Having HootSuite or TweetDeck open and watching it all day for me is the equivalent of reading the crawl on CNN – addicting at first, but eventually nauseating to watch.  I do keep it open but off to the side on my peripheral second monitor.

3. Getting my tweets to display inline. One goal of the Twitter integration was for an inline display of “headlines” on my blog.  I embedded the Twitter Flash Widget code directly on my Headlines page and that’s working fine.  I’m using that alongside the TextWise Gyzork feeds.  I’m also considering experimenting on the Headlines page with the HootSuite embed widget for columns of tweet streams of interest based upon my Twitter follows.

4. New TextWise feeds for the news page.  I haven’t done anything new with Gyzork yet, but I have to say the current feeds seeded with content from my first two blog postings are generating results that are very relevant.  This in turn is doing an excellent job of bringing my attention to articles of interest, which I very often tweet about.

5. Cleaning up the XSLT to better format the news feed. I worked with the XSLT for inlineRSS and got it to format the Gyzork posts in a similar way similar to my blog meta data.  I also hacked the PHP for Twitter Tools to apply some minor formatting to the sidebar and Headlines entries. 

6. Experiment with other semantic technologies to create automatic feeds of interest.  If any of you have other ideas for auto-generating news item listings for relevant content, please leave a comment or a pingback from your own blog posts.

I’d like to hear what you think about how this integration of Twitter with my site is working.  Drop me a line or better yet leave a comment below or on your own blog.  It would be great if you could share your best practices and favotire tools, tips, and tricks.

If you’re going to be at the 2009 Semantic Technology Conference in San Jose next week, please stop by on Wednesday afternoon and listen to Richard and me present on integrating taxonomy and folksonomy.  This presentation by Earley & Associates was developed by Stephanie Lemiuex over the last several years, who was originally scheduled to present it.  She has developed four categories describing how taxonomy and folksonomy can be used together, and has collected a wealth of illustrative examples.   Richard is an excellent presenter (he has a Ph.D. in Ontology!); I’m honored to share the dais with him and to have the opportunity to present Stephanie’s fantastic content.  

Here’s the official abstract:

Tagging isn’t new – it’s been around for a dog’s age in internet years.  But in the past few years some fresh ideas and tools have reinvigorated the social tagging world.  These new approaches include an attempt to improve findability through a bit of structure and control.  While the idea of adding control to folksonomy seems like going against the whole selling point of social tagging (flexibility, openness), it is bringing the tagging to a new level, making it more viable for practical use in enterprises.  This session will present hybrid approaches to formal taxonomies and social tagging.  How can they be used in the corporate environment?  What type of content is appropriate for social tagging?  What kind of software is available for the enterprise?  Learn how social tagging is not necessarily anathema to corporate taxonomy programs and how this hybrid approach can bring the best of both worlds: a fresh, up to date taxonomy with the structure needed to improve information findability.
Key Takeaways:

• Folksonomy and taxonomy defined
• Drawbacks of pure social tagging
• Social tagging in the enterprise
• Hybrid taxonomy & folksonomy approaches: Four models

Several things were important to me about how to do this:

1. I wanted to use this feature to notify about interesting content without a lot of authoring on my part.  Ideally posting an item to the digest would be automated with a semantic search bot pulling in the content.  Worst case I would have to drag and drop headlines and links in a Twitter-like fashion.
2. I didn’t want to add commentary – I wanted a simple re-blog capability.
3. I wanted attribution of the author to be very clear – this was not my content, just items of interest to share.

Inline RSS for a WordPress page

I concluded that what I really needed was a way to display RSS feeds in-line on a WordPress page to create my headlines or digest.  WordPress has loads of plug-ins that are great for integrating RSS feeds into sidebar widgets, but I found one plug-in – DGE inlineRSS – that provided all of the features I needed to put the RSS into the body of a WordPress posting or page. 

inlineRSS is pretty simple to use.  After install, there are three easy steps to getting a feed embedded into a page.  First, you need to configure inlineRSS to point to your feeds.  This is a simple matter of entering the RSS feed URL, and associating an XSLT file that will transform the feed into the HTML that displays on your site (inlineRSS provides a simple XSLT file that you can alter to meet your needs).  Next, you need to be sure that the configuration sets the path to the XSLT file on your site, and make any changes to the XSLT for your unique formatting.  Lastly, you need to enter the embed code for the inline RSS itself – this is simply:

!inlineRSS:myrssfeed

where myrssfeed is any of the feeds you configured in the inlineRSS options screen.  inlineRSS implements a WordPress filter to replace this code with the XSLT-formatted RSS feed.

Creating the Digest RSS Feed

The next step to getting a digest page up and running was to create the source of the RSS feed itself.  I considered three options:

• reFeed – a server for creating custom RSS feeds from items of interest
• Twitter or Ping.fm – generate a “tweetstream” of headlines and links (re-blogging vs. micro-blogging)
• TextWise – using a semantic search bot (specifically TextWise’s Gyzork demonstration app) to auto-generate a custom RSS feed that contains blog posts and news items that match specific semantic signatures

reFeed.  Last month, Mike Axelrod set up a reFeed server that he and I experimented with briefly.  This may hold some promise for the future for hand-selecting items of interest to reblog via an RSS feed.  However, Mike noted some technical issues, particularly with the quality of the RSS, so we tabled that project for the time being (he and I promise to blog more about reFeed in the near future).

Twitter and Ping.fm.  As Mike discusses in a related post, we’ve both been exploring Twitter and Ping.fm to stream reblog-type items to our websites.  Mike has the Ping.fm custom URL plug-in for WordPress working and has his “pingstream” going to headline items in his sidebar.  I’m doing the same, using Ping to drive Twitter, then putting the Twitter feed into my sidebar with the  Twitter Tools plug-in (look under Headlines in the sidebar to your right).  I’m still working on getting my Ping feed to display in-line with RSS or other techiques.

TextWise Gyzork.  You may be familiar with TextWise if only from my use of their technology here for finding related products, Wikipedia articles, and blogs for my blog posts.  TextWise uses semantic analysis to search the web for similar documents.  Each document (web page, blog post, etc.) gets a semantic signature that identifies the concepts in the document and their weight (i.e. relevance).  Gyzork lets me apply a semantic signature to any document, then create a saved semantic search that returns either blog posts or news items that match.  Gyzork also lets me create an RSS feed for that search.  If you check out my Headlines tab above, you’ll see a set of headlines (embedded using inlineRSS) that were generated by a Gyzork search seeded by one of my posts.  This search has been pretty good over the last six months at surfacing items I’m interested in.

Next Steps

I plan to continue to experiment with the automatic news digest.  Things I plan to try out:

• Propagating the news feed items to my status updates in LinkedIn and Twitter
• Getting my Ping.fm stream to display inline as a set of headlines
• Micro-blogging using Ping as the front end
• Working with TextWise technology to create more Gyzork feeds for the news page
• Cleaning up the XSLT to better format the news feed
• Experiment with other semantic technologies to create automatic feeds of interest

If you have any further ideas on this, please comment.

The Talis Connected Commons scheme is intended to directly support the publishing and reuse of Linked Data in the public domain by removing the costs associated with those activities.

The scheme is intended to support a wide range of different forms of data publishing. For example scientific researchers seeking to share their research data; dissemination of public domain data from a variety of different charitable, public sector or volunteer organizations; open data enthusiasts compiling data sets to be shared with the web community.

For qualifying data sets, Talis will provide, through the Talis Platform:

• Free hosting of up to 50 million RDF triples and 10Gb of content
• Access to data access services that operate on that data, including data retrieval and text search
• Free access to a public SPARQL endpoint for each dataset.

This means that data set providers will not incur any of the commercial costs normally associated with hosting data on the Talis Platform. In addition neither the data set provider or its users will incur any usage charges relating to the use of the Platform services made available on that data.

To qualify for entry into the scheme all data and content hosted in the Platform must be made available under one of the following public domain data licenses:

• Open Data Commons Public Domain Dedication and License
• Creative Commons CC0

As Mike Axelrod and I have been actively discussing, as more of these services become available through web APIs (e.g. Nova Spivak’s hosted ontlogy service, Amazon’s Public Data Sets, or text analysis services like TextWise, OpenCalais, or Amplify), developers can start mashing them up into useful virtual applications.  Marshall Kirkpatrick at ReadWriteWeb discussed the roadmap for this in a recent post. 

First, massive bodies of data are created or gathered, books are scanned, census data is collected, and patients donate their anonymous aggregate medical data to science. Next, the data is semantically analyzed and marked up (through any number of different semantic processing engines). Then, the data is stored and an API is made available (this is where the Talis Connected Commons comes in). Finally, developers build applications that leverage the smart data offered up through the platform, data visualizers find new stories to tell in images built from the marked up data and new relationships between people, organizations and concepts have the mist cleared away from them through systematic analysis of various permutations of previously unavailable structured data.

That last bit is what has Mike and me interested – finding new ways of making use of the relationships between data and content that all the various semantic tools unearth.

Nova Spivack’s semantic web company Twine is developing a free service to write and host semantic ontologies; the classification trees that enable machines to put concepts in topical context. Ready to play Aristotle and create an ontology of cheese, model airplanes, global anti-hunger organizations or any other topic?

What blogging was to publishing, a simple tool that made far more people able to participate, Twine’s new ontology writing and hosting service could be to the act of teaching machines about new topics.

The company wouldn’t let us publish the new service’s name but says it is aiming for a launch date this year, as soon as a go-to-market strategy and appropriate partnerships are lined up. The ontologies created won’t only work on Twine; they will be referenceable by semantic apps anywhere around the web.

A service for authoring and hosting ontologies certainly adds value.  My questions are:

• Who will write these public ontologies?
• When will Twine fix their dreadful user experience?

Mike Axelrod and I have spent a lot of time lately speculating what could be done by mashing-up various semantic web services like Twine’s.  It seem all of the technology providers have “one trick ponies” – you can get entity extraction, sentiment analysis, ontology management, tagging, related content – but what if you want to use multiple technologies?