[Up: web] [Robot Wisdom home page]

The Pragmatic Web: semantics as if surfers mattered

I was delighted, recently, to come across Steve Pepper's 'Ontopia' website, and especially the demos of his Omnigator browser for topic maps. The richest of the demos is about Italian opera [demo] and this sub-list of resource-types especially caught my fancy: article, A-Z Opera synopsis, bio, born, description, died, editorial guidelines, gallery, home page, illustration, libretto, Opera News synopsis, poster, premiere, revision, sound clip, synopsis, web site.

Steve seems to be one of the very few people in the semantic-web community who's really wrestling with the hardest problems of semantics (and posting partial results). [my rant] Most of his dozen demos include comparable lists, summarised below #

The opera-list obviously has quite a high 'Borges factor' [bkgd]-- 'born' and 'died' are not really the same kind of resources as 'sound clip' and 'home page'. And it's the latter kind that I want to explore, because I've been extending that set for the last few years in my own webpages, under the guise of 'text buttons'.

The idea of using text-buttons is catching on much more slowly than weblogs! Their purpose is two-fold: to minimise the extent of underlined blue text one has to tolerate (I hate reading underlined blue text!), and to minimise the ambiguity, when linking, about what sort of resource is being linked to. (I hate following a link expecting the etext of a book, and getting an Amazon ad-page instead!)

I usually limit text-buttons to one word, set off by square-brackets [dummy] to make it clear that they're outside the flow of the prose. (When they happen to fall at the end of a parenthetical phrase, I drop the brackets to avoid the unsightly '])' and use a colon to set them off: dummy)

(In my first paragraph, above, I reluctantly break this rule because the alternative would be so awkward:

I was delighted, recently, to come across Steve Pepper's 'Ontopia' website [website], and especially the demos [demos] of his Omnigator browser for topic maps.

But I predict there's a better compromise to be discovered.)

Winnowing Steve's full set of resource-types, I find two dozen that I've used as text-buttons, in some form:

article
bio
synopsis
tutorial

Of these, 'article' is the least straightforward, because on the Web its definition gets very blurry: some articles are written for offline magazines, some for online zines, some for encyclopedias, some are just self-published, etc. I often use [etext] to clarify that you'll get the real thing the author wrote, as opposed to [quote] or [extract] or [paraphrase] or [summary] or [translation] etc.

Along with [bio] I use [timeline] and [history]. Also related: [faq] and [overview].

And for any of these, if the resource is broken up into several pages it can be helpful to say [bio-4pg] or even [bio-multi] or [bio-short] or [bio-long]. The principle here is familiar to webloggers-- you want to tip off your readers to any 'caveats' that might make them regret having clicked. (My basic rule-of-thumb for navigation-design is: if the clicker is disappointed by what they find, then the linktext wasn't expressive enough.)

gallery
illustration
map reference
sound clip

I usually use [pic] or [pix] or [drawing] or [photo] or [map] or [RealAud]. Steve also has 'site using (foreign alphabet)' which would also be useful for pages that demo unfamiliar web-technologies.

web site
site about
informational site
home page
official site
main page

It would be tempting to collapse these into [site] but the subtle nuances are sometimes helpful. I often use [info] for informational pages.

discussion
discussion forum
download location
link collection

I use [bbs] and [chat] and [newsgroup] and [mailing list] and [archive] and [posting] and [thread], but there are many more subtleties that ought to be differentiated: is the discussion ongoing? can one join in? does one need to register to read the archives? etc.

Download-links should specify if it's free or demo or what, and which platform it's for.

I use [links] sometimes, but I think in the future linkpages should mostly be replaced by richly-linked informational pages. And most linkpages are so poorly annotated (and maintained) that it's a better policy to winnow out the best items and link them directly (rather than expecting each reader to re-do that for themselves).

definition
mentioned in
occurrence in text
kudo
opinion

These are sort-of-classic 'footnote material', but one needs to be cautious about handling them usefully. If most readers will need to look up the definition, it's much better policy to include the definition on your own page, and save your readers the extra click.

If the definition is complex, you can summarise it along with a [def] or [etym] link for more info. When annotating a book like 'Ulysses' I'm constantly doing Google-searches for names or obscure words, and sometimes instead of finding good info I just find a passing mention [passim] or a passing use (of a phrase). In the latter case I'll sometimes link the searchpattern itself so people can see how rare it is: [rare-searchpattern]. If a fact is mentioned in passing, and I want to attribute it, I use [cite].

I often use [review] or [critique] or [analysis] or [rave] or [rant] for expressions of opinion. When collating reviews of a particular book, you want to avoid: [review] [review] [review] [review] [review]. But the ideal approach is a lot of work-- you should read each one and summarise it in a word!

I sometimes use "[review] ditto ditto" for easier readability, and similarly sometimes "[info] more".

alternative
background
resource
source

These are meaningless except in context, which is fine in context but adds to the Borges-factor of Steve's lists. I often use [mirror] for an alternative etext-source. Sometimes I use [meta] to differentiate a meta-discussion.

date
description
identifier

These are normally so short that it's absurd to link them. (Just copy them onto your page.) But semantic-web folks aren't worried about such interface quibbles, yet.

Steve's other categories are too specialised to worry much about:

AZ Opera synopsis
Opera News synopsis
born
died
revision date
publication date
premiere
in use [range of dates]
character table
mapping table
writing direction
number of characters [in alphabet]
contract
description in 'free xml tools' index
editorial guidelines
flag [of country]
libretto
poster
proposal
proposed resolution
PSDS document
requirements document
resolution
revision
specification
specification text
status
version

I use [table] sometimes, and [guidelines] and [proposal].

But my main point here isn't about text-buttons, it's about this unrecognised subset of web-semantics that might be called 'web pragmatics'.

From the traditional AI-perspective [faq] all nodes on (any given level of) an ontology are of equal weight, but from the point-of-view of any real human navigating the Web, some nodes are much more important than others. And regardless of the particular (semantic) topics the surfer is interested in, there are (pragmatic) meta-topics like [map] that will be generally-- or even universally-- useful.

Some categories not yet mentioned: weblog, news, latest, help, search, buy, compare

I'm not advocating that these names be formally standardised-- search-engines might benefit from a less-nuanced subset, but pragmatics should always put the humans first.

And it will be a long time before the list begins to stabilise-- good designers should always be seeking to fuse different resources into new unities.

But I think we should start to map the relationships among all these, and start looking for web-design strategies that might leverage them (eg, what sorts of links does the ideal etext-page offer? author-bio, author-etexts, author-news...?) [more]

Opera: article, AZ Opera synopsis, bio, born, description, died, editorial guidelines, gallery, home page, illustration, libretto, Opera News synopsis, poster, premiere, revision, sound clip, synopsis, web site

Seville : article, description, editorial guidelines, home page, web site

Scripts: Article, Character table, Date, Definition, Discussion forum, In use, Mapping table, Mentioned in, Number of characters, Official site, Site about, Site using, Specification, Tutorial, Version, Writing direction

XMLtools: Description in 'Free XML tools' index, Download location, Home page, Informational site, Link collection, Resource, Specification

Ontopia: description, identifier, PSDS document, publication date, revision date, status, version

Published Subjects: description, identifier, PSDS document, publication date, revision date, status, version

TopicMaps: Alternative, Background, Date, Definition, Definition, Description, Discussion, Home page, Occurrence in text, Opinion, Proposal, Proposal, Proposed resolution, Requirements document, Resolution, Source, Specification text, Version

Atlas: Flag, Main page, Map reference

Resume?: contract, kudo, web site

You can submit a new URL or any other suggestion for this page by typing it into the box below. It will instantly become visible to anyone at this comments page. I should get around to checking it out and updating it above within a week or three, at which point I'll delete it from the comments page.

If you want credit, include your name and email (otherwise it's anonymous). You can use HTML but you don't have to.

[Up: web]
Web-design pages:
main : academia : info-design : adding value : resource-pages : lessons-learned : best-worst : plugging leaks

Special topics:
surfing-skills : url-hacking : open content : semantics : pagelength : linktext : startpages : bookmarklets : weblogging : colors : autobiographical pages : thumbnail-graphics : web-video : timeline of hypertext

Anti-XML/W3C/etc:
structure-myth : page-parsing : firstcut-parser : html-history : semantic web

Design prototypes:
topical portal : dense-content faq : annotated lit : random-access lit-summary : poetry sampler : gossipy history : author-resources : hyperlinked-timeline : horizontal-timeslice : web-dossier

Website-resource pages:
RobotWisdom.com : Altavista.com : 1911encyclopedia.com : Google.com : IMDb.com : Perseus.org : Salon.com : Yahoo.com

Older stuff:
design-lab : design-checklist : HyperTerrorist : design-theory : design cog-sci

Hosting provided by instinct.org. Content may be copied under Open Web Content License.