Tag: semantics

Tag Clouds: "A New User Interface?"

May 3rd, 2006 — 10:58pm

In Piv­ot­ing on tags to cre­ate bet­ter nav­i­ga­tion UI Matt McAl­lis­ter offers the idea that we’re see­ing “a new user inter­face evolv­ing out of tag data,” and uses Wikio as an exam­ple. For con­text, he places tag clouds within a con­tin­uüm of the evo­lu­tion of web nav­i­ga­tion, from list views to the new tag-based nav­i­ga­tion emerg­ing now.
It’s an insight­ful post, and it allows me to build on strong ground­work to talk more about why and how tag clouds dif­fer from ear­lier forms of nav­i­ga­tion, and will become [part of] a new user inter­face.
Matt iden­ti­fies five ‘leaps’ in web nav­i­ga­tion inter­faces that I’ll summarize:

  1. List view; a list of links
  2. Left-hand col­umn; a stan­dard loca­tion for lists of links used to navigate
  3. Search boxes and results pages; mak­ing very large lists manageable
  4. Tab nav­i­ga­tion; a list of other nav­i­ga­tion lists
  5. Tag nav­i­ga­tion; tag clouds

A Les­son in ‘Lis­tory’
As Matt men­tions, all four pre­de­ces­sors to tag based nav­i­ga­tion are really vari­a­tions on the under­ly­ing form of the list. There’s use­ful his­tory in the evo­lu­tion of lists as web nav­i­ga­tion tools. Early lists used for nav­i­ga­tion were sta­tic, cho­sen by a site owner, ordered, and flat: recall the lists of favorite sites that appeared at the bot­tom of so many early per­sonal home pages.
These basic nav­i­ga­tion lists evolved a vari­ety of order­ing schemes, (alpha­bet­i­cal, numeric), began to incor­po­rate hier­ar­chy (shown as sub-menus in nav­i­ga­tion sys­tems, or as indent­ing in the left-column Matt men­tions), and allowed users to change their order­ing, for exam­ple by sort­ing on a vari­ety of fields or columns in search results.
From sta­tic lists whose con­tents do not change rapidly and reflect a sin­gle point of view, the lists employed for web nav­i­ga­tion and search results then became dynamic, per­son­al­ized, and reflec­tive of mul­ti­ple points of view. Ama­zon and other e-commerce des­ti­na­tions offered recently viewed items (yours or oth­ers), things most requested, sets bounded by date (pub­lished last year), sets dri­ven by vary­ing para­me­ters (related arti­cles), and lists deter­mined by the nav­i­ga­tion choices of oth­ers who fol­lowed sim­i­lar paths.)
But they remained fun­da­men­tally lists. They item­ized or enu­mer­ated choices of one kind or another, indi­cated implicit or explicit prece­dence through order­ing or the absence of order­ing, and were designed for lin­ear inter­ac­tion pat­terns: start at the begin­ning (or the end, if you pre­ferred an alter­na­tive per­spec­tive — I still habit­u­ally read mag­a­zines from back to front…) and work your way through.
Tag clouds are dif­fer­ent from lists, often by con­tents and pre­sen­ta­tion, and more impor­tantly by basic assump­tion about the kind of inter­ac­tion they encour­age. On tag-based nav­i­ga­tion Matt says, “This is a new layer that pre­empts the search box in a way. The visual rep­re­sen­ta­tion of it is a tag cloud, but the inter­ac­tion is more like a pivot.” Matt’s men­tion of the inter­ac­tion hits on an impor­tant aspect that’s key to under­stand­ing the dif­fer­ences between clouds and lists: clouds are not lin­ear, and are not designed for lin­ear con­sump­tion in the fash­ion of lists.
I’m not say­ing that no one will read clouds left to right (with Roman alpha­bets), or right to left if they’re in Hebrew, or in any other way. I’m say­ing that tag clouds are not meant for ‘read­ing’ in the same way that lists are. As they’re com­monly visu­al­ized today, clouds sup­port mul­ti­ple entry points using visual dif­fer­en­tia­tors such as color and size.
Start­ing in the mid­dle of a list and wan­der­ing around just increases the amount of visual and cog­ni­tive work involved, since you need to remem­ber where you started to com­plete your sur­vey. Start­ing in the “mid­dle” of a tag cloud — if there is such a loca­tion — with a brightly col­ored and big juicy visual morsel is *exactly* what you’re sup­posed to do. It could save you quite a lot of time and effort, if the cloud is well designed and prop­erly ren­dered.
Kunal Anand cre­ated a visu­al­iza­tion of the inter­sec­tions of his del.icio.us tags that shows the dif­fer­ences between a cloud and a list nicely. This is at heart a pic­ture, and accord­ingly you can start look­ing at it any­where / any­way you pre­fer.
Visu­al­iz­ing My Del.icio.us Tags

We all know what a list looks like…
iTunes Play Lists

What’s In a Name?
Describ­ing a tag cloud as a weighted list (I did until I’d thought about it fur­ther) misses this impor­tant qual­i­ta­tive dif­fer­ence, and reflects our early stages of under­stand­ing of tag clouds. The term “weighted list” is a list-centered view of tag clouds that comes from the pre­ced­ing frame of ref­er­ence. It’s akin to describ­ing a com­puter as an “arith­metic engine”, or the print­ing press as “mov­able type”.
[Aside: The label for tag clouds will prob­a­bly change, as we develop con­cepts and lan­guage to frame new the user expe­ri­ences and infor­ma­tion envi­ron­ments that include clouds. For exam­ple, the lan­guage Matt uses — the word ‘pivot’ when he talks about the expe­ri­ence of nav­i­gat­ing via the tag cloud in Wikio, not the word ‘fol­low’ which is a default for describ­ing nav­i­ga­tion — in the post­ing and his screen­cast reflects a pos­si­ble shift in fram­ing.]
A Cam­era Obscura For the Seman­tic Land­scape
I’ve come to think of a tag cloud as some­thing like the image pro­duced by a cam­era obscura.
Cam­era Obscura
Where the cam­era obscura ren­ders a real-world land­scape, a tag cloud shows a seman­tic land­scape like those cre­ated by Amber Frid-Jimenez at MIT.
Seman­tic Land­scape

Seman­tic Land­scape

Like a cam­era obscura image, a tag cloud is a fil­tered visu­al­iza­tion of a mul­ti­di­men­sional world. Unlike a cam­era obscura image, a tag cloud allows move­ment within the land­scape. And unlike a list, tag clouds can and do show rela­tion­ships more com­plex than one-dimensional lin­ear­ity (expe­ri­enced as prece­dence). This abil­ity to show more than one dimen­sion allows clouds to reflect the struc­ture of the envi­ron­ment they visu­al­ize, as well as the con­tents of that envi­ron­ment. This frees tag clouds from the lim­i­ta­tion of sim­ply item­iz­ing or enu­mer­at­ing the con­tents of a set, which is the fun­da­men­tal achieve­ment of a list.
Ear­lier, I shared some obser­va­tions on the struc­tural evo­lu­tion — from sta­tic and flat to hier­ar­chi­cal and dynamic — of the lists used as web nav­i­ga­tion mech­a­nisms. As I’ve ven­tured else­where, we may see a sim­i­lar evo­lu­tion in tag clouds.
It is already clear that we’re wit­ness­ing evo­lu­tion in the pre­sen­ta­tion of tag clouds in step with their greater visu­al­izatin capa­bil­i­ties. Clouds now rely on an expand­ing vari­ety of visual cues to show an increas­ingly detailed view of the under­ly­ing seman­tic land­scape: prox­im­ity, depth, bright­ness, inten­sity, color of item, color of field around item. I expect clouds will develop other cues to help depict the many con­nec­tions (per­ma­nent or tem­po­rary) link­ing the labels in a tag cloud. It’s pos­si­ble that tag clouds will offer a user expe­ri­ence sim­i­lar to some of the ontol­ogy man­age­ment tools avail­able now.
Is this “a new user inter­face”? That depends on how you define new. In Shap­ing Things, author and futur­ist Bruce Ster­ling sug­gests, “the future com­posts the past” — mean­ing that new ele­ments are sub­sumed into the accu­mu­la­tion of lay­ers past and present. In the con­text of nav­i­ga­tion sys­tems and tag clouds, that implies that we’ll see mix­tures of lists from the four pre­vi­ous stages of nav­i­ga­tion inter­face, and clouds from the lat­est leap; a fusion of old and new.
Exam­ples of this com­post­ing abound, from 30daytags.com to Wikio that Matt McAl­lis­ter exam­ined.
30DayTags.com Tag Clouds

Wikio Tag Cloud

As lists encour­aged lin­ear inter­ac­tions as a result of their struc­ture, it’s pos­si­ble that new user inter­faces rely­ing on tag clouds will encour­age dif­fer­ent kinds of seek­ing or find­ing behav­iors within infor­ma­tion expe­ri­ences. In “The endan­gered joy of serendip­ity” William McK­een bemoans the decrease of serendip­ity as a result of pre­cisely directed and tar­geted media, search­ing, and inter­ac­tions. Tag clouds — by offer­ing many con­nec­tions and mul­ti­ple entry paths simul­ta­ne­ously — may help reju­ve­nate serendip­ity in dan­ger in a world of closely focused lists.

2 comments » | Ideas, Tag Clouds

NYTimes.com Redesign Includes Tag Clouds

April 11th, 2006 — 9:58pm

Though you may not have noticed it at first (I didn’t — they’re located a few steps off the front page), the recently launched design of NYTimes.com includes tag clouds. After a quick review, I think their ver­sion is a good exam­ple of a cloud that offers some increased capa­bil­i­ties and con­tex­tual infor­ma­tion that together fall in line with the likely direc­tions of tag cloud evo­lu­tion we’ve con­sid­ered before.
Specif­i­cally, the New York Times tag cloud:

  1. allows users to change the cloud’s con­text — and thus its con­tent — with a set of con­trols (vis­i­ble as tabs, run­ning across the top)
  2. lets cloud con­sumers change the dis­play behav­ior of the cloud by switch­ing modes from list to cloud in-line, not out­side the user’s area of activity
  3. sup­ports the chain of under­stand­ing for cloud con­sumers by pro­vid­ing clear indi­ca­tion of the time period cov­ered (the note about update frequency)
  4. offers [lim­ited] capa­bil­i­ties to work with / share tag cloud con­tent out­side the cloud via email — though the mes­sage con­tains only a link to the cloud page, and not a full rendering

NYtimes.com Tag Cloud

The NYTimes.com tag cloud shows the most pop­u­lar search terms used by read­ers within three time frames: the last 24 hours, the last 7 days, and the last 30 days. Choos­ing search terms as the makeup for a cloud is a bit curi­ous — but it may be as close to socially gen­er­ated meta­data as seemed rea­son­able for a first explo­ration (one that doesn’t require a sub­stan­tial change in the busi­ness or pub­lish­ing model).
Given the way that clouds lend them­selves to show­ing mul­ti­ple dimen­sions of mean­ing, such as change over time, I think the Times tag cloud would be more valu­able if it offered the option to see all three time frames at once. I put together a quick cut and paste of a con­cept screen that shows this sort of lay­out:
Screen Con­cept: 3 Clouds for Dif­fer­ent Time Frames

In an exam­ple of the rapid mor­ph­ing of memes and def­i­n­i­tions to fit shift­ing usage con­texts (as in Thomas Vanderwal’s obser­va­tions on the shift­ing usage of folk­son­omy) the NYTimes.com kept the label tag cloud, while this is more prop­erly a weighted list: the tags shown are in fact search terms, and not labels applied to a focus of some kind by tag­gers.
It’s plain from the lim­ited pres­ence and vis­i­bil­ity of clouds within the over­all site that the staff at NYTimes.com are still explor­ing the value of tag clouds for their spe­cific needs (which I think is a mature approach), oth­er­wise I imag­ine the new design con­cept and nav­i­ga­tion model would uti­lize and empha­sized tag clouds to a greater degree. So far, the Times uses tag clouds only in the new “Most Pop­u­lar” sec­tion, and they are offered as an alter­na­tive to the default list style pre­sen­ta­tion of pop­u­lar search terms. This posi­tion within the site struc­ture places them a few steps in, and off the stan­dard front page-to-an-article user flow that must be one of the core sce­nar­ios sup­ported by the site’s infor­ma­tion archi­tec­ture.
NYTimes.com User Flow to Tag Cloud

Still, I do think it’s a clear sign of increas­ing aware­ness of the poten­tial strength of tag clouds as a way of visu­al­iz­ing seman­tic infor­ma­tion. The Times is an estab­lished entity (occa­sion­ally serv­ing as the def­i­n­i­tion of ‘the estab­lish­ment’), and so is less likely to endan­ger estab­lished rela­tion­ships with cus­tomers by chang­ing its core prod­uct across any of the many chan­nels used for deliv­ery.
Ques­tions of risk aside, tag clouds (here I mean any visu­al­iza­tion of seman­tic meta­data) couLd be a very effec­tive way to scan the head­lines for a sense of what’s hap­pen­ing at the moment, and the shift­ing impor­tance of top­ics in rela­tion to on another. With a tag cloud high­light­ing “immi­gra­tion”, “duke”, and “judas”, vis­i­tors can imme­di­ately begin to under­stand what is news­wor­thy — at least in the minds of NYTimes.com read­ers.
At first glance, low­er­ing the amount of time spent read­ing the news could seem like a strong busi­ness dis­in­cen­tive for using tag clouds to stream­line nav­i­ga­tion and user flow. With more con­sid­er­a­tion, I think it points to a new poten­tial appli­ca­tion of tag clouds to enhance com­pre­hen­sion and find­abil­ity by giv­ing busy cus­tomers pow­er­ful tools to increase the speed and qual­ity of their judg­ments about what to devote their atten­tion to in order to acheive under­stand­ing greater depth. In the case of pub­li­ca­tions like the NYTimes.com, tag clouds may be well suited for con­vey­ing snap­shots or sum­maries of com­plex and deep domains that change quickly (what’s the news?), and offer­ing rapid nav­i­ga­tion to spe­cific areas or top­ics.
A new user expe­ri­ence that offers a vari­ety of tag clouds in more places might allow dif­fer­ent kinds of move­ment or flow through the larger envi­ron­ment, enabling new behav­iors and sup­port­ing dif­fer­ing goals than the cur­rent infor­ma­tion archi­tec­ture and user expe­ri­ence.
Pos­si­ble Screen Flow Incor­po­rat­ing Clouds

Step­ping back from the specifics of the design, a broader ques­tion is “Why tag clouds now?” They’re cer­tainly timely, but that’s not a busi­ness model. This is just spec­u­la­tion, but I recall job post­ings for an Infor­ma­tion Archi­tect posi­tion within the NYTimes.com group on that appeared on sev­eral recruit­ing web­sites a few months ago — maybe the new team mem­bers wanted or were directed to include tag clouds in this design? If any of those involved are allowed to share insights, I’d very much like to hear the thoughts of the IAs / design­ers / prod­uct man­agers or other team mem­bers respon­si­ble for includ­ing tag clouds in the new design and struc­ture.
And in light of Mathew Patterson’s com­ments here about cus­tomer accep­tance of mul­ti­ple clouds in other set­tings and con­texts (price­line europe), I’m curi­ous about any usabil­ity test­ing or other user research that might have been done around the new design, and any the find­ings related to tag clouds.

Comment » | Ideas

Tag Clouds: Navigation For Landscapes of Meaning

March 14th, 2006 — 4:53pm

I believe the value of sec­ond gen­er­a­tion clouds will be to offer ready nav­i­ga­tion and access to deep, com­plex land­scapes of mean­ing built up from the cumu­la­tive seman­tic infor­ma­tion con­tained in many inter­con­nected tag clouds. I’d like share some thoughts on this idea; I’ll split the dis­cus­sion into two posts, because there’s a fair amount of mate­r­ial.
In a pre­vi­ous post on tag clouds, I sug­gested that the great value of first gen­er­a­tion tag clouds is their abil­ity to make con­cepts and meta­data — seman­tic fields — broadly acces­si­ble and easy to under­stand and work with through visu­al­iza­tion. I believe the shift in the bal­ance of roles and value from first to sec­ond gen­er­a­tion reflects nat­ural growth in cloud usage and aware­ness, and builds on the two major trends of tag cloud evo­lu­tion: enhanced visu­al­iza­tion and func­tion­al­ity for work­ing with clouds, and pro­vi­sion of exten­sive con­tex­tual infor­ma­tion to accom­pany tag clouds.
Together, these two growth paths allow cloud con­sumers to fol­low the indi­vid­ual chains of under­stand­ing that inter­sect at con­nected clouds, and bet­ter achieve their goals within the infor­ma­tion envi­ron­ment and out­side. Fun­da­men­tally, I believe the key dis­tinc­tions between first and sec­ond gen­er­a­tion clouds will come from the way that clouds func­tion simul­ta­ne­ously as visu­al­iza­tions and nav­i­ga­tion mech­a­nisms, and what they allow nav­i­ga­tion of — land­scapes of mean­ing that are rich in seman­tic con­tent of high value.
For exam­ples of both direc­tions of tag cloud evo­lu­tion com­ing together to sup­port nav­i­ga­tion of seman­tic land­scapes, we can look at some of the new fea­tures del.icio.us has released in the past few months. I’ve col­lected three ver­sions of the infor­ma­tion archi­tec­ture of the stan­dard del.icio.us URL details page from the past seven months as an exam­ple of evo­lu­tion hap­pen­ing right now.
The first ver­sion (screen­shot and break­down in Fig­ure 1) shows the URL details page some­time before August 15th, 2005, when it appeared on Matt McAlister’s blog.
Fig­ure 1: Del.icio.us URL Page — August 2005

The lay­out or infor­ma­tion archi­tec­ture is fairly sim­ple, offer­ing a list of the com­mon tags for the url / focus, a sum­mary of the post­ing his­tory, and a more detailed list­ing of the post­ing his­tory that lists the dates and tag­gers who book­marked the item, as well as the tags used for book­mark­ing. There’s no cloud style visu­al­iza­tion of the tags attached to this sin­gle focus avail­able: at this time, del.icio.us offered a ren­dered tag cloud visu­al­iza­tion at the aggre­gate level for the whole envi­ron­ment.
Envi­ron­ment and sys­tem design­ers know very well that as the scope and com­plex­ity of an envi­ron­ment increase — in this case, the num­ber of tag­gers, focuses, and tags, plus their cumu­la­tive his­to­ries — it becomes more impor­tant for peo­ple to be explic­itly aware of the con­text of any item in order to under­stand it prop­erly. Explicit con­text becomes more impor­tant because they can rely less and less on implicit con­text or assump­tions about con­text based on the uni­ver­sal aspects of the envi­ron­ment. This is how cloud con­sumers’ needs for clearly vis­i­ble and acces­si­ble chains of under­stand­ing dri­ves the fea­tures and capa­bil­i­ties of tag clouds. Later ver­sions of this page addresses these needs in dif­fer­ing ways, with dif­fer­ing lev­els of suc­cess.
Fig­ure 2 shows a more recent ver­sion of the del.licio.us his­tory for the Ma.gnolia.com ser­vice. This screen­shot taken about ten days ago in early March, while I was work­ing on a draft of this post.
Fig­ure 2: Del.icio.us URL Page — Early March 2006

Key changes from the first ver­sion in August to this sec­ond ver­sion include:

  1. Chang­ing visu­al­iza­tion of the Com­mon Tags block to a cloud style rendering
  2. Remov­ing the indi­vid­ual tags cho­sen by each tag­ger from the Post­ing His­tory block
  3. The addi­tion of a large and promi­nent block of space devoted to “User Notes”
  4. Mov­ing the Post­ing His­tory block to the right column
  5. Chang­ing visu­al­iza­tion of the Post­ing His­tory block to a proto-cloud style rendering

The most impor­tant change in this sec­ond ver­sion is the removal of the indi­vid­ual sets of tags from the Post­ing His­tory. Sep­a­rat­ing the tags applied to the focus from asso­ci­a­ton with the indi­vid­ual tag­gers that chose them strips them of an impor­tant layer of con­text. Remov­ing the nec­es­sary con­text for the tag cloud breaks the chain of under­stand­ing (Fig­ure 3) link­ing tag­gers and cloud con­sumers, and obscures or increases the costs of the social con­cep­tual exchange that is the basic value of del.icio.us to its many users. In this ver­sion, cloud con­sumers con­sumers read­ing the URL details page can only find spe­cific tag­gers based on the con­cepts they’ve matched with this focus by vis­it­ing or nav­i­gat­ing to each indi­vid­ual tag­gers’ area within the larger del.icio.us envi­ron­ment one at a time.
Fig­ure 3: Chain of Under­stand­ing
The switch to ren­der­ing the Com­mon Tags block as a tag cloud is also impor­tant, as an indi­ca­tor of the con­sis­tent spread of clouds to visu­al­ize seman­tic fields, and their grow­ing role as nav­i­ga­tion tools within the larger land­scape.
The User Notes are a good exam­ple of an attempt to pro­vide addi­tional con­tex­tual infor­ma­tion with (poten­tially) high value. User Notes are cre­ated by users exclu­sively for the pur­pose of pro­vid­ing con­text. The other forms of con­text shown in the new lay­out — the Post­ing His­tory, Related Items — serve a con­tex­tual func­tion, but are not cre­ated directly by users with this goal in mind. The dif­fer­ence between the two pur­poses for these items undoubt­edly influ­ences the way that peo­ple cre­ate them, and what they cre­ate: it’s a ques­tion that more detailed inves­ti­ga­tions of tag­ging prac­tices will surely exam­ine.
The third ver­sion of the same URL his­tory page, shown in Fig­ure 4, was released very shortly after the sec­ond, prov­ing tag cloud evo­lu­tion is hap­pen­ing so quickly as to be dif­fi­cult to track delib­er­ately on a broad scale.
Fig­ure 4: Del.icio.us URL Page — March 2006 #2

This ver­sion changes the con­tent and lay­out of the Post­ing His­tory block, restor­ing the com­bined dis­play of indi­vid­ual tag­gers who tagged the URL, with the tags they applied to it, in the order in which they tagged the URL for the first time.
The third ver­sion makes two marked improve­ments over the first and sec­ond versions:

  1. Pre­sen­ta­tion of the indi­vid­ual chains of under­stand­ing that inter­sect with this focus / cloud in nav­i­ga­ble form, to increase aware­ness of the con­text for this item and allow users to retrace these paths to their origins
  2. Pre­sen­ta­tion of indi­vid­ual tag­gers’ flat­tened clouds that inter­sect this focus as nav­i­ga­tion mech­a­nisms for mov­ing from the cur­rent focus to else­where within the larger landscape

These three dif­fer­ent ver­sions of the del.icio.us URL details page show that the amount and type of con­tex­tual infor­ma­tion accom­pa­ny­ing a sin­gle focus is increas­ing, and that the num­ber of con­crete nav­i­ga­ble con­nec­tions to the larger seman­tic land­scape of which the focus is one ele­ment also increas­ing
Over­all, it’s clear that clouds are quickly emerg­ing as nav­i­ga­tion tools for com­plex land­scapes of mean­ing, and that cloud con­text has and will con­tinue to become more impor­tant for cloud cre­ation and use.
And so before dis­cussing the con­text nece­sary for clouds and the role of clouds as nav­i­ga­tion aids in more detail, it will be help­ful to get an overview of land­scapes of mean­ing, and how they arise.
Land­scapes of Mean­ing
A land­scape of mean­ing is a densely inter­con­nected, highly valu­able, exten­sive infor­ma­tion envi­ron­ment rich in seman­tic con­tent that is cre­ated by com­mu­ni­ties of tag­gers who build con­nected tag clouds. In the early land­scapes of mean­ing emerg­ing now, a con­nec­tion between clouds can be a com­mon tag, tag­ger, or focus: any one of the three legs of the Tag­ging Tri­an­gle required for a tag cloud (more on this below). Because tag clouds visu­al­ize seman­tic fields, con­nected tag clouds visu­al­ize and offer access to con­nected seman­tic fields, serv­ing as bridges between the indi­vid­ual accu­mu­la­tions of mean­ing each cloud con­tains.
Con­nect­ing hun­dreds of thou­sands of indi­vid­u­ally cre­ated clouds and fields, as del.icio.us has enabled social book­mark­ers to do by pro­vid­ing nec­es­sary tools and infra­struc­ture, cre­ates a very large infor­ma­tion envi­ron­ment whose ter­rain or geog­ra­phy is com­posed of seman­tic infor­ma­tion. Such a seman­tic land­scape is a land­scape con­structed or made up of mean­ing. It is an infor­ma­tion envi­ron­ment that allows peo­ple to share con­cepts or for social pur­poses of all kinds, while sup­ported with visu­al­iza­tion, con­tex­tual infor­ma­tion, func­tion­al­ity, and far-ranging nav­i­ga­tion capa­bil­i­ties.
The flickr Land­scape
flickr is a good exam­ple of a land­scape of mean­ing that we can under­stand as a seman­tic land­scape. In a pre­vi­ous post on tag clouds, I con­sid­ered the flickr all time most pop­u­lar tags cloud (shown in Fig­ure 5) in light of the basic struc­ture of clouds:
“The flickr style tag cloud is …a visu­al­iza­tion of many tag sep­a­rate clouds aggre­gated together. …the flickr tag cloud is the visu­al­iza­tion of the cumu­la­tive seman­tic field accreted around many dif­fer­ent focuses, by many peo­ple. …the flickr tag cloud func­tions as a visu­al­iza­tion of a seman­tic land­scape built up from all asso­ci­ated con­cepts cho­sen from the com­bined per­spec­tives of many sep­a­rate tag­gers.”
Fig­ure 5: The flickr All Time Most Pop­u­lar Tags Cloud

From our ear­lier look at the struc­ture of first gen­er­a­tion tag clouds we know a tag cloud visu­al­izes a seman­tic field made up of con­cepts referred to by labels which are applied as tags to a focus of some sort by tag­gers.
Based on our under­stand­ing of the struc­ture of a tag cloud as hav­ing a sin­gle focus, the flickr cloud shows some­thing dif­fer­ent because it includes many focuses. The flickr all time most pop­u­lar tags cloud com­bines all the indi­vid­ual tag clouds around all the indi­vid­ual pho­tos in flickr into a sin­gle visu­al­iza­tion, as Fig­ure 6 shows.
Fig­ure 6: The flickr Land­scape of Mean­ing

This means the flickr all time most pop­u­lar tags cloud is in fact a visu­al­iza­tion of the com­bined seman­tic fields behind each of those indi­vid­ual clouds. It’s quite a bit big­ger in scope than a tra­di­tional sin­gle focus cloud. Because the scope is so large, the amount of mean­ing it sum­ma­rizes and con­veys is tremen­dous. The all time most pop­u­lar tags cloud is in fact a his­toric win­dow on the cur­rent and his­tor­i­cal state of the seman­tic land­scape of flickr as a whole.
This is where con­text becomes crit­i­cal to the proper under­stand­ing of a tag cloud. The cloud title “All time most pop­u­lar tags” sets the con­text for this tag cloud, within the bound­aries of the larger land­scape envi­ron­ment defined and com­mu­ni­cated by flickr’s user epx­e­ri­ence. With­out this title, the cloud is mean­ing­less despite the large and com­plex seman­tic land­scape — all of the infor­ma­tion envi­ron­ment of flickr — it visu­al­izes so effec­tively, because cloud con­sumers can­not retrace a com­plete chain of under­stand­ing to cor­rectly iden­tify the cloud’s ori­gin.
flickr — 1st Gen­er­a­tion Land­scape Nav­i­ga­tion
The flickr cloud is a pow­er­ful nav­i­ga­tion mech­a­nism for quickly and eas­ily mov­ing about within the land­scape of mean­ing built up by all those thou­sands and thou­sands of indi­vid­ual clouds. Still, because it is a first gen­er­a­tion cloud, we can­not directly fol­low any of the many indi­vid­ual chains of under­stand­ing con­nect­ing this cloud’s tags back to spe­cific tag­gers, or the con­cepts they asso­ciate with spe­cific pho­tos or focuses. In this visu­al­iza­tion, the group’s under­stand­ing of mean­ing is more impor­tant than any individual’s under­stand­ing. And so the flickr cloud does not yet allow us com­pre­hen­sive nav­i­ga­tion of the under­ly­ing seman­tic land­scape illus­trated in Fig­ure 6 (chains of under­stand­ing sug­gested in light green). The flickr cloud also remains a first gen­er­a­tion tag cloud because users can­not con­trol its con­text.
Fig­ure 7: A Seman­tic Land­scape

Even so, these nav­i­ga­tional and con­tex­tual needs will help iden­tify the way that users rely on clouds to work in land­scapes of mean­ing.
Growth of Land­scapes
Land­scapes of mean­ing like flickr, del.icio.us, or the bur­geon­ing num­ber of social seman­tic busi­ness ven­tures debut­ing as I write — typ­i­cally grow from the bot­tom up, emerg­ing as dozens or thou­sands of indi­vid­ual tag clouds cre­ated for dif­fer­ent rea­sons by dif­fer­ent tag­gers coin­ci­den­tally or delib­er­ately inter­con­nect and over­lap, all of this hap­pen­ing through a vari­ety of social mech­a­nisms. Tag­gers typ­i­cally cre­ate con­nected or over­lap­ping tag clouds one at a time, adding tags, focuses, and tag­gers (by cre­at­ing new accounts) in the ad hoc fash­ion of open net­works and archi­tec­tures. But first we should look at the Tag­ging Tri­an­gle to under­stand the most basic ele­ments of a tag cloud.
The Tag­ging Tri­an­gle
To make a tag cloud, you have to have three ele­ments: a focus, a tag­ger, and a(t least one) tag. I call this the Tag­ging Tri­an­gle, illus­trated in Fig­ure 8. In the most com­mon ren­der­ings of famil­iar tag clouds, one or two of these ele­ments are often implied but not shown: yet all three are always present.
This illus­tra­tion shows a cloud of labels, not tags, because a ren­dered cloud is really a list of labels. The labels shown in most first gen­er­a­tion clouds are often tags, but struc­turally they could also be a set of names for tag­gers, as in the del.icio.us post­ing his­tory block proto-cloud we saw above, or a set of focuses as in the ‘Inverted Cloud’ I sug­gested.
Fig­ure 8: The Tag­ging Tri­an­gle
An Exam­ple Land­scape
A sim­ple exam­ple of the growth of seman­tic land­scapes leads nat­u­rally to the dis­cus­sion of spe­cific ways that tag clouds will enable nav­i­ga­tion within large land­scapes of mean­ing.
Fig­ure 9 shows the tag cloud accreted around a sin­gle focus. This cloud includes some of the tags that Tag­ger 1 has used in total across all the tag clouds she’s cre­ated (those other clouds aren’t shown). We’ll assume that she’s cre­ated other clouds for other focuses.
Fig­ure 9: A Sin­gle Tag Cloud

When a sec­ond per­son, Tag­ger 2, tags that same focus (again with a sub­set of the total set of all his tags), and some of those tags are the same as those used for this focus by Tag­ger 1, their indi­vid­ual tag clouds for this focus (shown by the dashed line in the cumu­la­tive tag cloud) con­nect via the com­mon tags, and the cumu­la­tive cloud grows. If any of the tags from their total sets are the same, but are not used for this focus, they form another con­nec­tion between the two tag­gers. Fig­ure 10 shows two indi­vid­ual clouds con­nected in both these ways.
Fig­ure 10: Two Con­nected Clouds

When a third tag­ger adds a third cloud with com­mon tags and unique tags around the same focus, the cumu­la­tive cloud grows, and the num­ber of both kinds of con­nec­tions between tags and tag­gers grows. Fig­ure 11 shows three con­nected clouds.
Fig­ure 11: Con­nected Clouds

Every tag cloud visu­al­izes a seman­tic field, and so the result of this bot­tom up growth is a series of inter­linked seman­tic fields cen­tered around a com­mon focus, as Fig­ure 12 shows. Since seman­tic fields are made of con­cepts, linked fields result in linked con­cepts.
Fig­ure 12: Con­nected Seman­tic Fields

The total num­ber and the vari­ety of kinds of inter­con­nec­tions amongst these three tag­gers, their tags, and a sin­gle focus is remark­able. As this sim­ple exam­ple shows, the total num­ber and den­sity of con­nec­tions link­ing even a mod­er­ate size pop­u­la­tion of tag­gers, tags, and focuses could quickly become very large. This increased scale dri­ves qual­i­ta­tive and quan­ti­ta­tive topol­ogy changes in the net­work that per­mit a land­scape of mean­ing to emerge from con­nected seman­tic fields.
Land­scapes And Depth
The accu­mu­la­tion of con­nec­tions and con­cepts cre­ates a land­scape of mean­ing with real depth; but it’s the depth of a land­scape that dri­ves its value. For this dis­cus­sion, I’m defin­ing depth loosely as the amount of seman­tic infor­ma­tion or the den­sity of the seman­tic field either across the whole land­scape, or at a cho­sen point.
Value of course is a very sub­jec­tive judge­ment. In par­tic­i­pa­tory economies like that of del.icio.us, the value to indi­vid­ual users is pre­dom­i­nantly one of loosely struc­tured seman­tic exchange based on accu­mu­la­tion of col­lec­tive value through shared indi­vid­ual efforts. From a busi­ness view­point, a group of investors and yahoo as a buyer saw con­sid­er­able value in the emer­gent land­scape and / or other kinds of assets
To make the idea of depth a bit clearer, Fig­ure 13 illus­trates two views of a seman­tic land­scape built up by the over­lap of tag clouds. The aer­ial view shows the con­tents, dis­tri­b­u­tion, and over­lap of a num­ber of tag clouds around a set of focuses. The hori­zon view shows the depth of the seman­tic field for each focus, based on the amount of over­lap or con­nec­tion between the cloud around that focus and all the other clouds.
Fig­ure 13: Seman­tic Land­scape Depth Views

Of course this is only a con­cep­tual way of show­ing the cumu­la­tive seman­tic infor­ma­tion that makes up a land­scape of mean­ing, so it does not address the rel­a­tive value of this infor­ma­tion. Plainly some indi­ca­tion of the qual­ity of the seman­tic infor­ma­tion in a land­scape is crit­i­cal impor­tant to mea­sure­ments of both depth and value. Met­rics for qual­ity could come from a com­bi­na­tion of assess­ment of the diver­sity and gran­u­lar­ity of the tag pop­u­la­tion for the focus, bench­marks for the domain of the focus and tag­gers (health­care indus­try), and an esti­mate on the matu­rity of the domain, the focus, and the tag clouds in the seman­tic land­scape.
Look­ing ahead, it’s likely that accepted met­rics for defin­ing and describ­ing the depth, value, and char­ac­ter­is­tics of seman­tic fields and land­scapes will emerge as new com­bi­na­tions of some of the mea­sure­ments used now in the realms of cog­ni­tive lin­guis­tics, set the­ory, sys­tem the­ory, topol­ogy, infor­ma­tion the­ory, and quite a few other dis­ci­plines besides.
In Part Two
The sec­ond post in this series of two will fol­low sev­eral of the top­ics intro­duced here to con­clu­sion, as well as cover some new top­ics, including:

  • How chains of under­stand­ing shape needs for cloud con­text and nav­i­ga­tion paths
  • How the tag­ging tri­an­gle will define nav­i­ga­tion within land­scapes of meaning
  • The emer­gence of strat­i­fi­ca­tion in land­scapes of meaning
  • The idea that clouds and land­scapes have a shape which con­veys mean­ing and value
  • The kinds of con­tex­tual infor­ma­tion and con­trols nec­es­sary for nav­i­ga­tion and social exchanges

Watch­ing Nav­i­ga­tion Fol­low Chains of Under­stand­ing
I’ll close with a screen­cast put together by Jon Udell that cap­tures a wide rang­ing nav­i­ga­tion path through the del.icio.us landscape.

Comment » | Ideas, Tag Clouds

Second Generation Tag Clouds

February 23rd, 2006 — 5:34pm

Lets build on the analy­sis of tag clouds from Tag Clouds Evolve: Under­stand­ing Tag Clouds, and look ahead at what the near future may hold for sec­ond gen­er­a­tion tag clouds (per­haps over the next 12 to 18 months). As you read these pre­dic­tions for struc­tural and usage changes, keep two con­clu­sions from the pre­vi­ous post in mind: first, ade­quate con­text is crit­i­cal to sus­tain­ing the chain of under­stand­ing nec­es­sary for suc­cess­ful tag clouds; sec­ond, one of the most valu­able aspects of tag clouds is as visu­al­iza­tions of seman­tic fields.
Based on this under­stand­ing, expect to see two broad trends sec­ond in gen­er­a­tion tag clouds.
In the first instance, tag clouds will con­tinue to become rec­og­niz­able and com­pre­hen­si­ble to a greater share of users as they move down the nov­elty curve from nou­veau to known. In step with this grow­ing aware­ness and famil­iar­ity, tag cloud usage will become:
1. More fre­quent
2. More com­mon
3. More spe­cial­ized
4. More sophis­ti­cated
In the sec­ond instance, tag cloud struc­tures and inter­ac­tions will become more com­plex. Expect to see:
1. More sup­port for cloud con­sumers to meet their needs for con­text
2. Refined pre­sen­ta­tion of the seman­tic fields under­ly­ing clouds
3. Attached con­trols or fea­tures and func­tion­al­ity that allow cloud con­sumers to directly change the con­text, con­tent, and pre­sen­ta­tion of clouds
Together, these broad trends mean we can expect to see a sec­ond gen­er­a­tion of numer­ous and diverse tag clouds val­ued for con­tent and capa­bil­ity over form. Sec­ond gen­er­a­tion clouds will be eas­ier to under­stand (when designed cor­rectly…) and open to manip­u­la­tion by users via increased func­tion­al­ity. In this way, clouds will visu­al­ize seman­tic fields for a greater range of sit­u­a­tions and needs, across a greater range of speci­ficity, in a greater diver­sity of infor­ma­tion envi­ron­ments, for a greater num­ber of more var­ied cloud con­sumers.
Usage Trends
To date, tag clouds have been applied to just a few kinds of focuses (links, pho­tos, albums, blog posts are the more rec­og­niz­able). In the future, expect to see spe­cial­ized tag cloud imple­men­ta­tions emerge for a tremen­dous vari­ety of seman­tic fields and focuses: celebri­ties, cars, prop­er­ties or homes for sale, hotels and travel des­ti­na­tions, prod­ucts, sports teams, media of all types, polit­i­cal cam­paigns, finan­cial mar­kets, brands, etc.
From a busi­ness view­point, these tag cloud imple­men­ta­tions will aim to advance busi­ness ven­tures explor­ing the poten­tial value of aggre­gat­ing and expos­ing seman­tic fields for a vari­ety of strate­gic pur­poses:
1. Cre­at­ing new mar­kets
2. Under­stand­ing or chang­ing exist­ing mar­kets
3. Pro­vid­ing value-added ser­vices
4. Estab­lish­ing com­mu­ni­ties of inter­est / need / activ­ity
5. Aid­ing over­sight and reg­u­la­tory imper­a­tives for trans­parency and account­abil­ity.
Mea­sure­ment and Insight
I think tag clouds will con­tinue to develop as an impor­tant poten­tial mea­sure­ment and assess­ment vehi­cle for a wide vari­ety of pur­poses; clouda­li­cious is a good exam­ple of an early use of tag clouds for insight. Other appli­ca­tions could include using tag clouds to present meta­data in geospa­tial or spa­tiose­man­tic set­tings that com­bine GPS / GIS and RDF con­cept / knowl­edge struc­tures.
Within the realm of user expe­ri­ence, expect to see new user research and cus­tomer insight tech­niques emerge that employ tag clouds as visu­al­iza­tions and instan­ti­a­tions of seman­tic fields. Maybe even cloud sort­ing?
Clouds As Nav­i­ga­tion
Turn­ing from the strate­gic to the tac­ti­cal realm of expe­ri­ence design and infor­ma­tion archi­tec­ture, I expect tag clouds to play a grow­ing role in the nav­i­ga­tion of infor­ma­tion envi­ron­ments as they become more com­mon. Nav­i­ga­tional appli­ca­tions com­prise one of the first areas of tag cloud appli­ca­tion. Though nav­i­ga­tion rep­re­sents a fairly nar­row usage of tag clouds, in light of their con­sid­er­able poten­tial in reify­ing seman­tic fields to ren­der them action­able, I expect nav­i­ga­tional set­tings will con­tinue to serve as a pri­mary exper­i­men­tal and evo­lu­tion­ary venue for learn­ing how clouds can enhance larger goals for infor­ma­tion envi­ron­ments such as enhanced find­abil­ity.
For new infor­ma­tion envi­ron­ments, the rules for tag clouds as nav­i­ga­tion com­po­nents are largely unwrit­ten. But many infor­ma­tion envi­ron­ments already have mature nav­i­ga­tion sys­tems. In these set­tings, tag clouds will be one new type of nav­i­ga­tion mech­a­nism that infor­ma­tion archi­tects and user expe­ri­ence design­ers inte­grate with exist­ing nav­i­ga­tion mech­a­nisms. David Fiorito’s and Richard Dalton’s pre­sen­ta­tion Cre­at­ing a Con­sis­tent Enter­prise Web Nav­i­ga­tion Solu­tion is a good frame­work / intro­duc­tion for ques­tions about how tag clouds might inte­grate into mature or exist­ing nav­i­ga­tion sys­tems. Within their matrix of struc­tural, asso­cia­tive and util­ity nav­i­ga­tion modes that are invoked at vary­ing lev­els of prox­im­ity to con­tent, tag clouds have obvi­ous strengths in the asso­cia­tive mode, at all lev­els of prox­im­ity to con­tent, and poten­tial strength in the struc­tural mode. Fig­ure 1 shows two tag clouds play­ing asso­cia­tive roles in a sim­ple hypo­thet­i­cal nav­i­ga­tion sys­tem.
Fig­ure 1: Asso­cia­tive Clouds

I also expect nav­i­ga­tion sys­tems will fea­ture mul­ti­ple instances of dif­fer­ent types of tag clouds. Nav­i­ga­tion sys­tems employ­ing mul­ti­ple clouds will use com­bi­na­tions of clouds from vary­ing con­texts (as flickr and tech­no­rati already do) or domains within a larger infor­ma­tion envi­ron­ment to sup­port a wide vari­ety of pur­poses, includ­ing implicit and explicit com­par­i­son, or views of the envi­ron­ment at mul­ti­ple lev­els of gran­u­lar­ity or res­o­lu­tion (high level / low level). Fig­ure 2 illus­trates mul­ti­ple clouds, Fig­ure 3 shows clouds used to com­pare the seman­tic fields of a one focus cho­sen from a list, and Fig­ure 4 shows a hier­ar­chi­cal lay­out of nav­i­ga­tional tag clouds.
Fig­ure 2: Mul­ti­ple Clouds

Fig­ure 3: Cloud Com­par­i­son Lay­out

Fig­ure 4: Pri­mary / Sec­ondary Lay­out

Struc­tural and Behav­ioral Trends
Let’s move on to con­sider struc­tural and behav­ioral trends in the sec­ond gen­er­a­tion of tag clouds.
Given the suc­cess of the sim­ple yet flex­i­ble struc­ture of first gen­er­a­tion tag clouds, I expect that sec­ond gen­er­a­tion clouds will not sub­stan­tially change their basic struc­ture. For exam­ple, tag clouds will not have to change to make use of chang­ing tag­ging prac­tices that enhance the seman­tic depth and qual­ity of tags applied to a focus, such as faceted tag­ging, use of qual­i­fiers, hier­ar­chi­cal tag­ging, and other forms. James Melzer iden­ti­fies some best prac­tices on del.icio.us that make con­sid­er­able sense when the focus of a seman­tic field is a link. His rec­om­men­da­tions include:

  • Source your infor­ma­tion with via:source_name or cite:source_name
  • Cre­ate a par­ent cat­e­gories, and thus a rudi­men­tary hier­ar­chy, with parent_tag/subject_tag
  • Men­tion pub­li­ca­tions names with in:publication_name
  • Flag the type of resource with .exten­sion or =resource_type
  • Use a com­bi­na­tion of gen­eral and spe­cific tags on every book­mark to pro­vide both clus­ter­ing and differentiation
  • Use syn­onyms or alter­nate forms of tags
  • Use unique or dis­tinc­tive terms from doc­u­ments as tags (don’t just use major sub­ject terms)

The two ele­ment struc­ture of first gen­er­a­tion tag clouds can accom­mo­date these tag­ging prac­tices. How­ever, with a seman­tic field of greater depth and rich­ness avail­able, the inter­ac­tions, behav­iors, and pre­sen­ta­tion of tag clouds will evolve beyond a sta­tic set of hyper­links.
Cloud con­sumers’ need for bet­ter con­text will drive the addi­tion of fea­tures and func­tion­al­ity that iden­tify the con­text of a tag cloud explic­itly and in detail. For exam­ple, clouds cre­ated by a defined audi­ence will iden­tify that audi­ence, whether it be sys­tem admin­is­tra­tors, free­lance web design­ers, DJ’s, or pas­try chefs rat­ing recipes and cook­ing equip­ment and pro­vide indi­ca­tion of the scope and time peri­ods that bound the set of tags pre­sented in the cloud. Flickr and oth­ers do this already, offer­ing clouds of tags cov­er­ing dif­fer­ent inter­vals of time to account for the chang­ing pop­u­lar­ity of tags over their lifes­pan.
Mov­ing from pas­sive to inter­ac­tive, tag clouds will allow users to change the cloud’s seman­tic focus or con­text with con­trols, fil­ters, or other para­me­ters (did some­one say ‘slid­ers’ — or is that too 5 min­utes ago…?). I’ve seen sev­eral pub­lic requests for these sorts of fea­tures, like this one: “It would be great if I could set pref­er­ences for items such as time frame or for tags that are rel­e­vant to a par­tic­u­lar area etc or even colour the most recent tags a fiery red or remove the most recent tags.” Fig­ure 5 shows a tag cloud with con­text con­trols attached.
Fig­ure 5: Con­text Con­trols
Fig­ure 6: Behav­ior Con­trols
Diver­si­fy­ing con­sumer needs and goals for way find­ing, ori­en­ta­tion, infor­ma­tion retrieval, task sup­port, prod­uct pro­mo­tion, etc., will bring about inverted tag clouds. Inverted tag clouds will cen­ter on a tag and depict all focuses car­ry­ing that tag.
Fig­ure 7: Inverted Clouds Show Con­cep­tu­ally Related Focuses
In the vein of con­tin­ued exper­i­ment, tag clouds will take increased advan­tage with RIA / AJAX and other user expe­ri­ence con­struc­tion meth­ods. Fol­low­ing this, tag clouds may take on some of the func­tions of known nav­i­ga­tion ele­ments, appear­ing as sub-menus / drop down menus offer­ing sec­ondary nav­i­ga­tion choices.
Fig­ure 8: Clouds As Drop Menus

Along the same lines, tag clouds will demon­strate more com­plex inter­ac­tions, such as spawn­ing other tag clouds that act like mag­ni­fy­ing lenses. These over­lap­ping tag clouds may offer: mul­ti­ple lev­els of gran­u­lar­ity (a gen­eral view and zoom view) of a seman­tic field; the­saurus style views of related con­cepts; para­me­ter dri­ven term expan­sion; com­mon types of rela­tion­ship (other peo­ple bought, by the same author, syn­onyms, pre­vi­ously known as, etc.)
Fig­ure 9: Mag­ni­fy­ing Clouds
Look­ing at the inter­sec­tion of usage and behav­ior trends, I expect tag clouds will evolve, dif­fer­en­ti­ate, and develop into stan­dard gen­res. Gen­res will con­sist of a sta­ble com­bi­na­tion of tag cloud con­tent, con­text, usage, func­tion­al­ity, and behav­ior within dif­fer­ent envi­ron­ments. The same busi­ness and user goals that sup­port gen­res in other media and modes of visu­al­iza­tion will drive the devel­op­ment of these tag cloud gen­res. One genre I expect to see emerge shortly is the search result.
Read­ing over the list, I see this is an aggres­sive set of pre­dic­tions. It’s fair to ask if I really have such high expec­ta­tions for tag clouds? I can’t say tag clouds will take over the world, or even the Inter­net. But I do believe that they fill a gap in our col­lec­tive visu­al­iza­tion toolset. The quan­tity, qual­ity, and rel­e­vance of seman­tic infor­ma­tion in both real and vir­tual envi­ron­ments is con­stantly increas­ing. (In fact, the rate of increase is itself increas­ing, though that is a tem­po­rary phe­nom­e­non.) I think tag clouds offer a poten­tial to quickly and eas­ily sup­port the chain of under­stand­ing that’s nec­es­sary for seman­tic fields across diverse kinds of focuses. There’s need for that in many quar­ters, and I expect that need to con­tinue to grow.
For the moment, it seems obvi­ous that tag clouds will spend a while in an early exper­i­men­tal phase, and then move into an awk­ward ado­les­cent phase, as fea­tures, appli­ca­tions and gen­res sta­bi­lize in line with grow­ing aware­ness and com­fort with clouds in var­i­ous set­tings.
I expect these pre­dic­tions to be tested by exper­i­ments will play out quickly and in semi or fully pub­lic set­tings, as in the exam­ple of the dia­log sur­round­ing 83 degrees usage of a tag cloud as the sole nav­i­ga­tion mech­a­nism on their site that Rashmi Sinha’s post The tag-cloud replaces the basic menu — Is this a good idea? kicked off recently.
My answer to this ques­tion is that replac­ing all nav­i­ga­tion menus with a tag cloud is only a good idea under very lim­ited cir­cum­stances. It’s pos­si­ble that 83 Degrees may be one of these lim­ited instances. Star­tups can ben­e­fit con­sid­er­ably from any pos­i­tive atten­tion from the Web’s early adopter com­mu­nity (wit­ness Don’t Blow Your Beta by Michael Arring­ton of Techcrunch). The page’s designer said, “In this case it was done as a design/marketing effort and not at all for UI”. Since attract­ing atten­tion was the spe­cific pur­pose, I think the result is a suc­cess. But it’s still an exper­i­men­tal usage, and that’s con­sis­tent with the early stage of evo­lu­tion / devel­op­ment of tag clouds in gen­eral.
I’m look­ing for­ward to what hap­pens next…

10 comments » | Ideas, Tag Clouds

Tag Clouds Evolve: Understanding Tag Clouds

February 22nd, 2006 — 1:18pm

Zeld­man jok­ingly called tag clouds “the new mul­lets” last year. At the time, I think he was taken a bit by sur­prise by the rapid spread of the tag cloud (as many peo­ple were). A big year later, it looks like this ver­sion of the world’s favorite dou­ble duty hair­cut will stay in fash­ion for a while. Zeld­man was dis­cussing the first gen­er­a­tion of tag clouds. I have some ideas on what the sec­ond gen­er­a­tion of tag clouds may look like that will con­clude this series of two essays. These two pieces com­bine ideas brew­ing since the tag­ging break­out began in earnest this time last year, with some pre­dic­tions based on recent exam­ples of tag clouds in prac­tice.
Update: Part two of this essay, Sec­ond Gen­er­a­tion Tag Clouds, is avail­able.
This first post lays ground­work for pre­dic­tions about the sec­ond gen­er­a­tion of tag clouds by look­ing at what’s behind a tag cloud. I’ll look at first gen­er­a­tion tag clouds in terms of their reliance on a “chain of under­stand­ing” that seman­ti­cally links groups of peo­ple tag­ging and con­sum­ing tags, and thus under­lies tag­ging and social meta­data efforts in gen­eral. I’ll begin with struc­ture of first gen­er­a­tion tag clouds, and move quickly to the very impor­tant way that tag clouds serve as visu­al­iza­tions of seman­tic fields.
Anatomy of a Tag Cloud
Let’s begin with the famil­iar first gen­er­a­tion tag cloud. Tag clouds (here we’re talk­ing about the user expe­ri­ence, and not the pro­gram­matic aspects) com­monly con­sist of two ele­ments: a col­lec­tion of linked tags shown in vary­ing fonts and col­ors to indi­cate fre­quency of use or impor­tance, and a title to indi­cate the con­text of the col­lec­tion of tags. Flickr’s tags page is the iconic exam­ple of the first gen­er­a­tion tag cloud. Screen shots of sev­eral other well known tag cloud imple­men­ta­tions show this pat­tern hold­ing steady in first gen­er­a­tion tag­ging imple­men­ta­tions such as del.icio.us and tech­no­rati, and in newer efforts such as last.fm and ma.gnolia.
Wikipedia’s entry for tag cloud is quite sim­i­lar, read­ing, “A tag cloud (more tra­di­tion­ally known as a weighted list in the field of visual design) is a visual depic­tion of con­tent tags used on a web­site. Often, more fre­quently used tags are depicted in a larger font or oth­er­wise empha­sized, while the dis­played order is gen­er­ally alpha­bet­i­cal… Select­ing a sin­gle tag within a tag cloud will gen­er­ally lead to a col­lec­tion of items that are asso­ci­ated with that tag.“
In terms of infor­ma­tion ele­ments and struc­ture, first gen­er­a­tion tag clouds are low com­plex­ity. Fig­ure 1 shows a schematic view of a first gen­er­a­tion tag cloud. Fig­ures 2 through 5 are screen­shots of well-known first gen­er­a­tion tag clouds.
Fig­ure 1: Tag Cloud Struc­ture
Fig­ure 2: last.fm
Fig­ure 3: tech­no­rati
Fig­ure 4: del.icio.us
Fig­ure 5: Ma.gnolia
Tag Clouds: Visu­al­iza­tions of Seman­tic Fields
The sim­ple struc­ture of first gen­er­a­tion tag clouds allows them to per­form a very valu­able func­tion with­out undue com­plex­ity. That func­tion is to visu­al­ize seman­tic fields or land­scapes that are them­selves part of a chain of under­stand­ing link­ing tag­gers and tag con­sumers. This is a good moment to describe the “chain of under­stand­ing”. The “chain of under­stand­ing” is an approach I use to help iden­tify and under­stand all the dif­fer­ent kinds of peo­ple and mean­ing, and the trans­for­ma­tions and steps involved in pass­ing that mean­ing on, that must work and con­nect prop­erly in order for some­thing to hap­pen, or an end state to occur. The chain of under­stand­ing is my own vari­a­tion / com­bi­na­tion of com­mon cog­ni­tive and infor­ma­tion flow map­ping using a sce­nario style for­mat. I’ve found the term res­onates well with clients and other audi­ences out­side the realm of IA.
How does the chain of under­stand­ing relate to tag clouds? The tags in tag clouds orig­i­nate directly from the per­spec­tive and under­stand­ing of the peo­ple tag­ging, but undergo changes while becom­ing a tag cloud. (For related read­ing, see Rashmi Sinha’s A social analy­sis of tag­ging which exam­ines some of the social mech­a­nisms under­ly­ing the activ­ity of tag­ging.) Tag clouds accrete over time when one per­son or a group of peo­ple asso­ciate a set of terms with a focus of some sort; a photo on flickr, a URL / link in the case of del.icio.us, an album or song for last.fm. As this list shows, a focus can be any­thing that can carry mean­ing or under­stand­ing. The terms or tags serve as car­ri­ers and ref­er­ences for the con­cepts each tag­ger asso­ciates with the focus. Con­cepts can include ideas of about­ness, ori­gin, or pur­pose, descrip­tive labels, etc. While the con­cepts may change, the focus remains sta­ble. What’s key is that the tag is a ref­er­ence and con­nec­tion to the con­cept the tag­ger had in mind. This con­nec­tion requires an ini­tial under­stand­ing of the focus itself (per­haps incor­rect, but still some sort of under­stand­ing), and the con­cepts that the tag­ger may or may not choose to asso­ciate with the focus. And this is the first step in the chain of under­stand­ing behind tag clouds, as shown in Fig­ure 6.
Fig­ure 6: Ori­gin: Focus and Con­cepts
As a result, tag clouds are more than col­lec­tion of descrip­tive or admin­is­tra­tive terms attached to a link, or other sort of focus. The tag is a sort of label that ref­er­ences a con­cept or set of con­cepts. A cloud of tags is then a col­lec­tion of labels refer­ring to a clus­ter of aggre­gated con­cepts. The com­bi­na­tion of tags that refer to con­cepts, with the orig­i­nal focus, cre­ates a ‘seman­tic field’. A seman­tic field is the set of con­cepts con­nected to a focus, but in a form that is now inde­pen­dent of the orig­i­nat­ing tag­gers, and avail­able to other peo­ple for under­stand­ing. In this sense, a seman­tic field serves as a form of rei­fied under­stand­ing that the tag­gers them­selves — as well as oth­ers out­side the group that cre­ated the seman­tic field — can now under­stand, act on, etc. (This speaks to the idea that infor­ma­tion archi­tec­ture is a dis­ci­pline strongly aimed at reifi­ca­tion, but that’s a dif­fer­ent dis­cus­sion…). Fig­ure 7 shows this sec­ond step in the chain of under­stand­ing; with­out it, there is no seman­tic field, and no tag cloud can form. And now because this post is writ­ten from the view­point of prac­ti­cal impli­ca­tions for tag cloud evo­lu­tion, I’m going to hold the def­i­n­i­tion and dis­cus­sion of a seman­tic field and focus, before I wan­der off track into semi­otics, lin­guis­tics, or other ter­ri­to­ries. The most impor­tant thing to under­stand is that *tag clouds com­prise visu­al­iza­tions of a seman­tic field*, as we’ve seen from the chain of under­stand­ing.
Fig­ure 7: Seman­tic Field
I believe tag clouds are rev­o­lu­tion­ary in their abil­ity to trans­late the con­cepts asso­ci­ated with nearly any­thing you can think of into a col­lec­tively vis­i­ble and action­able infor­ma­tion envi­ron­ment, an envi­ron­ment that car­ries con­sid­er­able evi­dence of the orig­i­nal under­stand­ings that pre­cede and inform it. In a prac­ti­cal infor­ma­tion archi­tec­ture sense, tag clouds can make meta­data — one of the more dif­fi­cult and abstract of the fun­da­men­tal con­cepts of the dig­i­tal uni­verse for the prover­bial per­son on the street — vis­i­ble in an eas­ily under­stood fash­ion. The genius of tag clouds is to make seman­tic con­cepts, the frames of under­stand­ing behind those con­cepts, and their man­i­fes­ta­tion as applied meta­data tan­gi­ble for many, many peo­ple.
Fig­ure 8: Seman­tic Field As Tag Cloud
With this notion of a tag cloud as a visu­al­iza­tion of a seman­tic field in mind, let’s look again at an exam­ple of a tag cloud in prac­tice. The flickr style tag cloud (what I call a first gen­er­a­tion tag cloud) is in fact a visu­al­iza­tion of many tag sep­a­rate clouds aggre­gated together. Seman­ti­cally then, the flickr tag cloud is the visu­al­iza­tion of the cumu­la­tive seman­tic field accreted around many dif­fer­ent focuses, by many peo­ple. In this usage, the flickr tag cloud func­tions as a visu­al­iza­tion of a seman­tic land­scape built up from all asso­ci­ated con­cepts cho­sen from the com­bined per­spec­tives of many sep­a­rate tag­gers.
To sum­ma­rize, cre­at­ing a tag cloud requires com­ple­tion of the first three steps of the chain of under­stand­ing that sup­ports social meta­data. Those steps are:
1. Under­stand­ing a focus and the con­cepts that could apply that focus
2. Accu­mu­lat­ing and cap­tur­ing a seman­tic field around the focus
3. Visu­al­iz­ing the seman­tic field as a tag cloud via trans­for­ma­tion
The fourth step in this chain involves users’ attempts to under­stand the tag cloud. For this we must intro­duce the idea of con­text, which addresses the ques­tion of which orig­i­nal per­spec­tives under­lie the seman­tic field visu­al­ized in a tag cloud, and how those con­cepts have changed before or dur­ing pre­sen­ta­tion.
How Cloud Con­sumers Under­stand Tag Clouds
Users need to put a given tag cloud in proper con­text in order to under­stand the cloud effec­tively. Their end may goals may be find­ing related items, sur­vey­ing the think­ing within a knowl­edge domain, iden­ti­fy­ing and con­tact­ing col­lab­o­ra­tors, or some other pur­pose, but it’s essen­tial for them to under­stand the tags in the cloud to achieve those goals. Thus when­ever a user encoun­ters a tag cloud, they ask and answer a series of ques­tions intended to estab­lish the cloud’s con­text and fur­ther their under­stand­ing. Con­text related ques­tions often include “Where did these tags come from? Who applied them? Why did they choose these tags, and not oth­ers? What time span does this tag cloud cover?” Con­text in this case means know­ing enough about the con­di­tions and envi­ron­ment from which the cloud was cre­ated, and the deci­sions made about what tags to present and how to present them. Fig­ure 9 sum­ma­rizes the idea of con­text.
Fig­ure 9: Cloud Con­text

Once the user or con­sumer places the tag cloud in con­text, the chain of under­stand­ing is com­plete, and they can being to use or work with the tag cloud. Fig­ure 10 shows the com­plete chain of under­stand­ing we’ve exam­ined.
Fig­ure 10 Chain of Under­stand­ing
In part two, titled “Sec­ond Gen­er­a­tion Tag Clouds”, I’ll share some thoughts on likely ways that the sec­ond gen­er­a­tion of tag clouds will evolve in struc­ture and usage in the near future, based on how they sup­port a chain of under­stand­ing that seman­ti­cally links tag­gers and tag cloud con­sumers. Con­text is the key for tag cloud con­sumers, and we’ll see how it affects the likely evo­lu­tion of the tag cloud as a visu­al­iza­tion tool.
Update: Part two Sec­ond Gen­er­a­tion Tag Clouds is available

10 comments » | Ideas, Tag Clouds

Of Madeleines And Metadata

January 23rd, 2006 — 2:08pm

A few months ago, I put up a posted called Tag­ging Comes To Star­bucks, in which I attempted to make the point that it’s bizarre when a product’s meta­data *over­whelms the expe­ri­ence of the prod­uct itself in it’s cus­tom­ary real world set­ting*.
My exam­ple was the meta­data encrusted pack­ag­ing of madeleines — “petite french cakes…” — at Star­bucks. Like the famous tooth­pick instruc­tions Dou­glas Adams immor­tal­ized in So Long and Thanks For All The Fish, this is a strong dis­con­ti­nu­ity of expe­ri­ence (though not nec­es­sar­ily one indi­cat­ing things gone awry at the core of civ­i­liza­tion) that implies new cog­ni­tive / per­cep­tual phe­nom­e­non.
New expe­ri­ences and frames of ref­er­ence usu­ally lack descrip­tive vocab­u­lary, which explains why I can’t pin this down neatly in words. But this is surely some­thing we can expect to encounter more in a future pop­u­lated with find­able things called spimes.
The bal­ance hasn’t shifted so far that we’re liv­ing inside Baudrillard’s ‘desert of the real’, but we are get­ting closer with each addi­tional layer of sim­u­la­tion, abstrac­tion, and meta­data applied to real sit­u­a­tions and objects.
After all it is impos­si­ble to inter­act (smell, touch, taste…) directly with these very ordi­nary pas­tries with­out expe­ri­enc­ing the inter­ven­ing lay­ers of meta­data pack­ag­ing.
Madeleines in situ:
The label­ing:
From SLATFATF: “It seemed to me that any civ­i­liza­tion that had so far lost its head as to need to include a set of detailed instruc­tions for use in a pack­age of tooth­picks, was no longer a civ­i­liza­tion I could live in and stay sane.” ~ Wonko the Sane

4 comments » | User Experience (UX)

The Aargh Page: Visualizing Pirate Argot

January 10th, 2006 — 1:13pm

What hap­pens when this clas­sic ver­nac­u­lar inter­jec­tion meets lin­guis­tics, data visu­al­iza­tion, and the Web?
The Aargh page, of course. (It should really be The Aargh! Page, but this is so fan­tas­tic that I can’t com­plain…)
Here’s a screen­shot of the graph that shows fre­quency of vari­ant spellings for aargh in Google, along two axes:
Note the snazzy mouseover effect, which I’ll zoom here:
Look­ing into the ori­gins aargh inevitably brings up Robert New­ton, the actor who played Long John Sil­ver in sev­eral Dis­ney pro­duc­tions based on the writ­ings of Robert Louis Steven­son. I remem­ber see­ing the movies as a child, with­out know­ing that they were the first live action Dis­ney movies broad­cast on tele­vi­sion. So do plenty of other peo­ple who’ve cre­ated trib­ute pages</>.
Aargh may have many spelling vari­a­tions, but at least three of them bear a stamp of legit­i­macy, as the edi­to­r­ial review of
The Offi­cial Scrab­ble Play­ers Dic­tio­nary (Paper­back) at Amazon.com explains, “If you’re using the 1991 edi­tion or the 1978 orig­i­nal, you’re woe­fully behind the Scrabble-playing times. With more than 100,000 2– to 8-letter words, there are some inter­est­ing addi­tions (“aargh,” “aar­rgh,” and “aar­rghh” are all legit­i­mate now), while words they con­sider offen­sive are no longer kosher. “
There’s even Inter­na­tional Talk Like A Pirate Day, cel­e­brated on Sep­tem­ber 19th every year. The orga­niz­ers’ site offers a nifty English-to-Pirate-Translator.
Most ran­dom per­haps is the Wikipedia link for Aargh the videogame, from the 80’s, with­out pirates.

Comment » | The Media Environment

Mental Models and the Semantics of Disaster

November 4th, 2005 — 3:47pm

A few months ago, I put up a post­ing on Men­tal Mod­els Lotus Notes, and Resililence. It focused on my chronic inabil­ity to learn how not to send email with Lous Notes. I posted about Notes, but what led me to explore resilience in the con­text of men­tal mod­els was the sur­pris­ing lack of acknowl­edge­ment of the scale of hur­ri­cane Kat­rina I came across at the time. For exam­ple, the day the lev­ees failed, the front page of the New York Times dig­i­tal edi­tion car­ried a gigan­tic head­line say­ing ‘Lev­ees Fail! New Orleans floods!’. And yet no one in the office at the time even men­tioned what hap­pened.
My con­clu­sion was that peo­ple were sim­ply unable to accept the idea that a major met­ro­pol­i­tan area in the U.S. could pos­si­bly be the set­ting for such a tragedy, and so they refused to absorb it — because it didn’t fit in with their men­tal mod­els for how the world works. Today, I came across a Resilience Sci­ence post­ing titled New Orleans and Dis­as­ter Soci­ol­ogy that sup­ports this line of think­ing, while it dis­cusses some of the inter­est­ing ways that seman­tics and men­tal mod­els come into play in rela­tion to dis­as­ters.
Quot­ing exten­sively from an arti­cle in The Chron­i­cle of Higher Edu­ca­tion titled Dis­as­ter Soci­ol­o­gists Study What Went Wrong in the Response to the Hur­ri­canes, but Will Pol­icy Mak­ers Lis­ten? the post­ing calls out how nar­row slices of media cov­er­age dri­ven by blurred seman­tic and con­tex­tual under­stand­ings, inac­cu­rately frame social responses to dis­as­ter sit­u­a­tions in terms of group panic and the implied break­down of order and soci­ety.
“The false idea of post­dis­as­ter panic grows partly from sim­ple seman­tic con­fu­sion, said Michael K. Lin­dell, a psy­chol­o­gist who directs the Haz­ard Reduc­tion and Recov­ery Cen­ter at Texas A&M Uni­ver­sity at Col­lege Sta­tion. ‘A reporter will stick a micro­phone in someone’s face and ask, ‘Well, what did you do when the explo­sion went off?’ And the per­son will answer, ‘I pan­icked.’ And then they’ll pro­ceed to describe a very log­i­cal, ratio­nal action in which they pro­tected them­selves and looked out for peo­ple around them. What they mean by ‘panic’ is just ‘I got very fright­ened.’ But when you say ‘I pan­icked,’ it rein­forces this idea that there’s a thin veneer of civ­i­liza­tion, which van­ishes after a dis­as­ter, and that you need out­side author­i­ties and the mil­i­tary to restore order. But really, peo­ple usu­ally do very well for them­selves, thank you.‘
Men­tal mod­els come into play when the arti­cle goes on to talk about the ways that the emer­gency man­age­ment agen­cies are orga­nized and struc­tured, and how they approach and under­stand sit­u­a­tions by default. With the new Home­land Secu­rity par­a­digm, all inci­dents require com­mand and con­trol approaches that assume a ded­i­cated and intel­li­gent enemy — obvi­ously not the way to man­age a hur­ri­cane response.
“Mr. Lin­dell, of Texas A&M, agreed, say­ing he feared that pol­icy mak­ers in Wash­ing­ton had taken the wrong lessons from Kat­rina. The employ­ees of the Depart­ment of Home­land Secu­rity, he said, ‘are mostly drawn from the Depart­ment of Defense, the Depart­ment of Jus­tice, and from police depart­ments. They’re firmly com­mit­ted to a command-and-control model.’ (Just a few days ago, Pres­i­dent Bush may have pushed the process one step fur­ther: He sug­gested that the Depart­ment of Defense take con­trol of relief efforts after major nat­ural dis­as­ters.)
“The habits of mind cul­ti­vated by mil­i­tary and law-enforcement per­son­nel have their virtues, Mr. Lin­dell said, but they don’t always fit dis­as­ter sit­u­a­tions. ‘They come from orga­ni­za­tions where they’re deal­ing with an intel­li­gent adver­sary. So they want to keep infor­ma­tion secret; ‘it’s only shared on a need-to-know basis. But emer­gency man­agers and med­ical per­son­nel want infor­ma­tion shared as widely as pos­si­ble because they have to rely on per­sua­sion to get peo­ple to coöper­ate. The prob­lem with putting FEMA into the Depart­ment of Home­land Secu­rity is that it’s like an organ trans­plant. What we’ve seen over the past four years is basi­cally organ rejec­tion.‘
If I read this cor­rectly, mis­aligned orga­ni­za­tional cul­tures lie at the bot­tom of the whole prob­lem. I’m still curi­ous about the con­nec­tions between an organization’s cul­ture, and the men­tal mod­els that indi­vid­u­als use. Can a group have a col­lec­tive men­tal model?
Accoridng to Col­lec­tive Men­tal State and Indi­vid­ual Agency: Qual­i­ta­tive Fac­tors in Social Sci­ence Expla­na­tion it’s pos­si­ble, and in fact the whole idea of this col­lec­tive men­tal state is a black hole as far as qual­i­ta­tive social research and under­stand­ing are concerned.

2 comments » | Modeling, The Media Environment

Defining Enterprise Semantics

September 15th, 2005 — 8:31am

JP Mor­gen­thal of DMReview.com offers a snap­shot of the process for defin­ing enter­prise seman­tics in Enter­prise Archi­tec­ture: The Holis­tic View: The Role of Seman­tics in Busi­ness.
Mor­gen­thal says, “When you under­stand the terms that your busi­ness uses to con­duct busi­ness and you under­stand how those terms impact your busi­ness, you can see clearly how to sup­port and main­tain the processes that use those terms with min­i­mal effort.“
Not a sur­prise, but how to make it hap­pen, and how to explain that to the business?

  1. Cap­ture — In this phase of the process a rep­re­sen­ta­tive for each busi­ness process estab­lishes the vocab­u­lary and their mean­ings required to sup­port that process. For exam­ple, in a sup­ply chain process the rep­re­sen­ta­tive might cap­ture words, such as buyer, trans­port or pay­ment method. In addi­tion to these words the rep­re­sen­ta­tive would explain what these terms mean in rela­tion to the process.
  2. Cat­e­go­riza­tion — In this phase, the vocab­u­lar­ies across all processes are orga­nized into a sys­tem… The sys­tem can be a sim­ple tax­o­nomic struc­ture that sim­ply relates process to vocab­u­lary, or it can be a more com­plex onto­log­i­cal struc­ture that cap­tures the rela­tion­ships of words across processes.
  3. Lever­age — This is phase where the tech­ni­cal staff imple­ments the vocab­u­lar­ies in the form of a dic­tio­nary or reg­istry. This dic­tio­nary can rep­re­sent a sim­ple lookup facil­ity or it can become an active part of the infra­struc­ture feed­ing the busi­ness rules and busi­ness process engines.

1 comment » | Information Architecture

On Semantics At The Enterprise Level

September 14th, 2005 — 5:39pm

In the same way that infor­ma­tion archi­tec­ture helps take users’ under­stand­ings of the struc­ture, mean­ing, and orga­ni­za­tion of infor­ma­tion into account at the level of domain-specific user expe­ri­ences, infor­ma­tion spaces, and sys­tems, the com­plex seman­tic bound­aries and rela­tion­ships that define and link enterprise-level domains is a nat­ural area of activ­ity for enter­prise infor­ma­tion archi­tec­ture.
Look­ing for some tech­ni­cally ori­ented mate­ri­als related to this level of IA — what I call enter­prise seman­tic frame­works — I came across a solid arti­cle titled Enter­prise Seman­tics: Align­ing Service-Oriented Archi­tec­ture with the Busi­ness in the Web Ser­vices Jour­nal.
The authors — Joram Boren­stein and Joshua Fox — take a web-services per­spec­tive on the busi­ness ben­e­fits of enterprise-level seman­tic efforts, but they do a good job of lay­ing out the case for the impor­tance of seman­tic con­cepts, under­stand­ing, and align­ment at the enter­prise level.
From the arti­cle abtract:
“Enter­prises need trans­parency, a clear view of what is hap­pen­ing in the orga­ni­za­tion. They also need agility, which is the abil­ity to respond quickly to changes in the inter­nal and exter­nal envi­ron­ments. Finally, orga­ni­za­tions require inte­gra­tion: the smooth inter­op­er­a­tion of appli­ca­tions across orga­ni­za­tional bound­aries. Encod­ing busi­ness con­cepts in a for­mal seman­tic model helps to achieve these goals and also results in addi­tional corol­lary ben­e­fits. This seman­tic model serves as a focal point and enables auto­mated dis­cov­ery and trans­for­ma­tion ser­vices in an orga­ni­za­tion.“
They also offer some ref­er­ences at the con­clu­sion of the article:

  • Boren­stein, J. and , J. (2003). “Seman­tic Dis­cov­ery for Web Ser­vices.” Web Ser­vices Jour­nal. SYS-CON Pub­li­ca­tions, Inc. Vol. 3, issue 4. www.sys-con.com/webservices/articleprint.cfm?id=507
  • Cowles, P. (2005). “Web Ser­vice API and the Seman­tic Web.” Web Ser­vices Jour­nal. SYS-CON Pub­li­ca­tions, Inc. Vol. 5, issue 2. www.sys-con.com/story/?storyid=39631&DE=1
  • Gen­ovese, Y., Hay­word, S., and Com­port, J. (2004). “SOA Will Demand Re-engineering of Busi­ness Appli­ca­tions.” Gart­ner. Octo­ber 8.
  • Linthicum, D. (2005). “When Build­ing Your SOA…Service Descrip­tions Are Key.” WebServices.Org. March 2005. www.webservices.org/ws/content/view/full/56944
  • Schulte, R.W., Valdes, R., and Andrews, W. (2004). “SOA and Web Ser­vices Offer Lit­tle Ven­dor Inde­pen­dence.” Gart­ner. April 8.
  • W3C Web Ser­vices Archi­tec­ture Work­ing Group: www.w3.org/2002/ws/arch/

Comment » | Information Architecture

Back to top