Stephen Arnold Blows a Gasket

Stephen Arnold, a provider of “news and information… about search and content processing,” has his hatchet out in Temis, Spammy PR, and Quite Silly Assertions.

Stephen Arnold, “Thy wit is a very bitter sweeting; it is a most sharp sauce,” well served in to the goose (as Arnold characterizes himself) himself. Steve, if you want to be paid for your work, ask up front. The payback you offer Temis, a text-mining solution vendor, says far more about you than it does about the company you target.

In the words of a marketing exec at a semantic-analysis company, not Temis, “What a super childish article.”

Sectumsempra -- For Enemies Arnold’s article is no dissection of the fine points of competitive market positioning. It deals broad cutting strokes (“Sectumsempra — For Enemies”?) in an attempt to hack away, well, much more than just the credibility of a vendor’s public relations. Arnold brackets his article introduction of Temis (in a lede buried four paragraphs in, as deeply as his hatchet blade) around a Jean Genet quotation,

“I recognize in thieves, traitors, and murderers, in the ruthless and the cunning, a deep beauty – a sunken beauty.” (The Thief’s Journal)

WTF? Temis is (likened to) thieves, traitors, and murders!?

What spurred this attack, which also devotes several paragraphs to an attempted take-down of Temis’s client, the American Society for Microbiology?

Arnold received a press release “attributed to an individual identified as Martine Fallon [sic]” that he characterizes as spam.

Arnold writes, “I considered that Martine Fallon may be a ruse like Betty Crocker.”

WTF redux!

For the record, it’s Martine Falhon. She is marketing and communications manager at Temis. Her full rap sheet is publicly accessible on LinkedIn. She’s a quite pleasant person, actually. I met her when I keynoted Temis’s 2009 user conference (which, by way of disclosure, was a paid gig, and Temis was one of seven sponsors of my Text Analytics 2009 market study).

Oy. A vendor sends a release to an analyst in its space — with an unsubscribe option that Arnold was able to use! — and Stephen Arnold blows a gasket. Yes, Arnold states he “previously asked the firm’s public relations expert, who seems to be more inclined to spam than research, to cease sending me meaningless spammy news releases. My request was ignored.”

“Bad, bad, bad public-relations expert,” I say. “Get back to your research. They need your help in the labs.”

Everything copacetic now? Well, not quite. There’s more to Arnold’s complaint. His complaint’s true core is not spam PR. It is here:

“What fascinated me is that Temis asked me to facilitate an introduction for them to a $1.2 billion company’s president. I did this and moved on. I assumed in the manner of French cultural norms that I would be rewarded with entrecote. Wrong. My reward has been spam.”

It’s about money. Temis didn’t pay Stephen Arnold for freely offered advice. Temis didn’t pay Stephen Arnold for freely offered advice so Temis gets slammed, for supposed spam, for claims Arnold questions, for being French. Yes, for being French. Arnold writes,

“I quite like Expert System SA in Bologna, Italy, and Bitext in Madrid, Spain. Great food, interesting culture, and -– nota bene –- no spam. One has to get the semantics correct. No spam from Italy. No spam from Spain. Hmmm. There’s a cultural message perhaps?”

There is some, slim attempt at substance in Arnold’s article. Arnold questions Temis’s claim to being “the leading provider of Semantic Content Enrichment solutions for the Enterprise.” Then he writes, “Leading? Semantic content enrichment. What’s that?” and asks, “What about outfits like Access Innovations, Concept Searching, Expert System SA, Smartlogic, and more than 75 other firms in the semantic space.”

Of course, it’s legit to question, but where’s the logic in offering specific other providers as contenders when you claim not to know what they’re contending for?

In case you doubted: Arnold’s list of contenders shows that he’s not completely clueless about “semantic content enrichment.” That said, of the companies he listed, only Expert System is squarely in that biz.

Myself, I describe the term functionally in a sponsored newsletter article I wrote for OpenText (a former consulting client): “Semantic content enrichment adds value to online information by tagging topics and providing context-sensitive links.” Content enrichment is a significant capability for digital-publishing platforms, for instance, MarkLogic’s. MarkLogic’s Open Enrichment Framework came out in 2008, integrating text-analysis products from companies that include both Temis and Access Innovations.

I (and OpenText, MarkLogic, and Temis) don’t stand alone is seeing enterprise value in semantic content enrichment. I’ll further point you to Barry Graubart, who cites a number of user organizations in a blog entry about a Software & Information Industry Association seminar last June (which I attended),
Semantic Technology Driving Real Revenue for Publishers.

In the end, my only quibble, as an industry observer, with Temis’s leadership claim is that I see the company as “a” rather than “the” leading semantic-content-enrichment provider. My issue with Stephen Arnold’s Temis slur, while I don’t have a dog in the fight, is far more significant. I find Arnold’s vindictive payback quite distasteful and even destructive. It calls out for a response. I hope I’ve adequately answered the challenge.

13 thoughts on “Stephen Arnold Blows a Gasket

  1. I haven’t heard of Arnold before, but I thought I’d give his article a shot (if only because Temis has been on my radar for a while). Wow, what an obtuse, self indulgent article! I only got through three paragraphs before the dense diversion into German philosophy reminded me that I was in the break room, and even hollow CES articles would be a better read.
    That being said, I’m curious as to what advantages Temis brings over other text analysis engines. Their claim of having licensed Xerox technology around Y2K sounds promising – do you know if it’s the XFST (and are you at liberty to admit it)? Finite-state technology is pretty powerful, and might explain Arnold’s appraisal of Temis as a legit contender. Still – your scathing review is spot-on.

      1. Oh, cool! Thanks for the heads-up. I work with a couple members of the old InXight team here at SAP, including the guy that wrote the book on XFST, and had no idea anyone else was using the technology. I’ll point the rest of my team to your article, they’ll definitely be interested.

    1. Hello bpiche,

      I am not going to comment on what Temis does or doesn’t do by way of semantic enrichment, however I did want to share my perspective on the difference between semantic enrichment and text analytics.

      When I think about semantic enrichment, I see it as transforming a piece of content into a linked data source. In order to do this you do indeed need text analytics for entity and relationship extraction, but you need more than that. Let me give you a quick example.

      A text analytics engine might recognize that [Marie Wallace] is a person, [Ireland] is a place, and Marie comes from Ireland and annotate the entities/relationships found. However when doing semantic enrichment, I would want to convert those annotations to openly addressible URIs that contribute to the linked data cloud. So I would need to disambiguate which Marie Wallace this is and then either find/reuse an existing URI (lots of folks use dbpedia) or else create a new URI in my own namespace. So, if I find Marie Wallace the actress it will be one URI, whereas Marie Wallace the IBMer would be another one. In addition, I would then want to be able to use these unique URIs to improve my own text analytics

      This is just one difference I see with semantic enrichment that I wanted to share.

      1. Yep. Semantic enrichment is cool stuff, and allows awesome technologies like Facebook and Google Search to connect entities to each other. It’s something we discuss every month at the Cambridge Semantic Web meetup, although I don’t work with linked data technologies professionally (I’m text analysis). Definitely the bleeding edge of web tech.
        Do you know if anyone is combining the categorization functionality of text analysis with linked data markup? Like, automatically tagging subjects, predicates, and objects with RDF based on the entity tags they’ve been given from categorization markup? Or like… doing category search on the data stores returned by SPARQL queries? To my knowledge, you can’t search ontologies based on categories, just relationships.

      2. bpiche,

        To respond to your question (for some reason I couldn’t reply to your reply – strange) … I also come from a text analytics backround and therefore everything I have done with semtech has been very much from a text analytics perspective. So yes, you can use categorization (or any text analytics) techniques for populating an ontology it just may be different link types.

        I also don’t ontologies and taxonomies (categories) as being different, the ontology is just a much more generalized model whereas with taxonomies they are parent-child relationships. In fact during some of our earlier semtech work we ran into the standard problem that its not easy to navigate ontologies so when we built a development tool we represented the ontologies as essentially faceted taxonomies – a simplification I know, but it worked OK.

        What I have liked about semtech from a text analytics perspective is that its given me so much freedom to add other data sources, semantics, inferencing, etc. to help me out in places where text analytics gets stuck.

        But I had better stop, because I could go on about this stuff forever and I don’t want to hijack Seth’s blog.

        [ Sorry Seth 🙂 ]

  2. Thanks for sharing, Seth.

    First off love the title “blowing your gasket” is one of my favorite expressions. Wonderfully colorful.

    No idea who this guy us, but heck… and I thought I was a drama queen. Temis is a solid text analytics company and I wouldn’t imagine they deserved that attack. I would however agree that their claim to be THE leaders in “semantic content enrichment” is somewhat presumptuous. However that’s what marketing folks say and not sure anyone ever takes the “best in market” expression at face value.

    As an aside, it’s great to see this term (semantic content enrichment) generating discussion. Maybe semtech going mainstream at last? I live in hope 🙂 Although I suspect that is me being overly optimistic and I suspect that most people still don’t differentiate it from just text analytics.

  3. Seth, you have more than adequately answered the challenge.
    Quel Pitre! (pardon my French).
    I can only reinforce Marie’s comment about text analytics and semantic enrichment.
    Temis is a great company with great and creative people… they definitely don’t deserve such a vile attack.
    If aggressing start-up claiming that they are THE leaders in a specific domain is Stephen’s new mission… de Gaulle would have answered “Vaste Programme”.

  4. Is it possible that he is attempting to use his method of search engine malipulation that he discussed at the Text Analytics Summit 2011? I would have assumed he’d always use harder to detect stimulus. For those who missed his talk, he discussed methods, without detail, that use particular stimulus to trigger reactions in the known algorithms of the search engines. Certainly a good idea given how they are known to have biases. However, the proof is in seeing the results. Searching on Bing and Google don’t seem to show any negative associations of Temis. So I’m left with just speculation as to why the article might be so odd…

    Thanks for the heads up on Temis, as I had never heard of them before.


  5. I can’t believe there are people interested in this drivel, in which I include the Arnold article and the equally pretentious “rebuttal”.

    Hard to see the interminable waffle spouted by Arnold as a “vile attack”. It all reads like an outburst from Comic Book Store Guy.

    Tragic – I advise everyone involved to get a life (each).

Leave a Reply