Joe Lamantia.com

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

Gen­res
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.
Con­clu­sions
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…