http://technology.timesonline.co.uk/tol/news/tech_and_web/article3532832.ece

Google could be superseded, says [Tim Berners-Lee]

"[...] Google may eventually be displaced as the pre-eminent brand on the internet by a company that harnesses the power of next-generation web technology, the inventor of the World Wide Web has said.

The search giant had developed an extremely effective way of searching for pages on the internet, [...] but that ability paled in comparison to what could be achieved on the "web of the future", which he said would allow any piece of information — such as a photo or a bank statement — to be linked to any other. [...]"

While I agree with Tim in principle that no technology, no matter how successful at some given time, is immune to being superceded by better tools, I see no reason for Google's demise because of that – for one, G. will probably lead, rather than follow, the coming paradigm shift (if it has to happen anyway).

That said, I'm far from convinced of the sanity of TBL's latter years' Semantic Web-mantra. It sounds, indeed it *assumes* a web in which everything is, hell, everything must be! perfect or else.... or the semantic layers(?) can go packing.

We live in a highly-imperfect, indeed chaotic, world (and getting worse day by sodding day), where it is our destiny to try to understand what makes it tick.... but assuming that we'll ever reach that Perfection Plateau, that the übercooperative Semantic Web will happen, is a bit much wishful thinking for me liking.

I thought TBL, when writing the specs[+] for by-definition-anarchistic WWW, has drawn the correct conclusions from Ted Nelson's visionary-but-ack-so-idealistic/ 100%-perfect/ ergo never realised/ Xanadu[tm], where every single granule, pixel, character, were to be tracked at all times, and exploited for profit with full accountability to owner/ publisher, and client/ reader alike.

Well, yes, Ted N. had that dream. Tim B-L had quite another. Guess which one captured our imagination.

[+] together with Robert Caillau

Berners-Lee is great but sooo utopist...

The full-blown semantic web needs everything to be thoroughly (if not perfectly) structured, but that doesn't mean that we can't reap vast benefits from moving as far towards the semantic web as is convenient and practical.

A latitude/longitude pair here, a factoid there, a statistic, a time of an event, a numerical ranking, a recipe ingredient, a song/composer or song/performer pair, a temperature/location/time datam, an "employee works for employer" assertion...

It all adds up. It's all data that can be mined in a more meaningful way than free text.

Perhaps from Tim Berners-Lee's point of view it looks like he almost single-handedly (or with a single organization) invented the current web with a detailed specification. And so that might make him believe you can pull these things off with a good, precise spec and some thinking; the Semantic Web!

But what if the world was just waiting for someone to come up the idea "pages containing a title and links", and every other part of his invention didn't matter that much? Anyone using <address>? <code>? <samp>? <dfn>? <fieldset>? <del>? <var>? Well, I know many people are using these (me included) but does that really affect the web or change its evolution? What it comes back to it seems are pages containing a title and links – and almost everything else that was put on top of that since then (like the image tag) was a gradual evolution of bad, spontaneous hacks, browser quirks, browser plug-ins, or proprietary browser enhancements.

Build anything more complex than <address> when <address> failed? Perhaps you can't design the web from top to bottom; perhaps all you can do is try to turn the noise into a signal by smarter, and perhaps one day truly learning and intelligent, parsing tools. Google's Q&A feature is a good example of that, showing how Google currently is not only not being superseded – they are in many ways leading the efforts towards the lower-case semantic web. Hand out only a simple lego brick, but then produce this brick many million times; make it a place for lazy people who don't often all agree on something.

> Build anything more complex than <address> when <address> failed?

It's not particularly interesting to know just that "this is an address". What we want is a tuple, i.e. a row of a database table. We want to know that "Pizza Hut is located at this address", or "This address is at 25 degrees North, 55 degrees East". The <address> tag doesn't give us that.

Microformats can address this. Microformats are additional structural hints inserted into standard HTML (e.g. as 'class' attributes).

Yahoo announced this month that they are going to "support semantic web standards" and crawl the data web:
http://www.ysearchblog.com/archives/000527.html

The semantic web won't be revolutionary; it will just kind of creep up on us.

> It's not particularly interesting to know just
> that "this is an address".

Perhaps the idea was that knowing this is an address, the browser can then do stuff like offer you an "add to address book" button when you visit the page; or a search engine crawler might show a map below the site in search results. Well, the latter is already done by Google with or without this tag – and I think with or without microformats too?

[This thread was automatically closed due to its age, but I opened it again to allow for Ianf's message here.]

[moved from "forks not taken"]

Philipp wrote a month+ ago:
[ http://blogoscoped.com/forum/127267.html#id127294]

»Perhaps from Tim Berners-Lee's point of view it looks like he almost single-handedly (or with a single organization) invented the current web with a detailed specification.«

Actually, Tim B-L did invent the web [= httpd + url + html], almost single-handedly – the original "people-tuple" included Robert Calliau of CERN; later the WWW-team that perfected the specifications was made up of at most 10 individuals. Before that there were only the Usenet, some hypertext authoring apps, and conference systems – walled gardens all; and Gopher, an early attempt at providing a uniform, single-line-address method of displaying no-mark-up directory contents. TBL's et al. genius was in transcending it all, and making HTML an easy to adopt "non-validatable" offshot of by-definition-always-kosher SGML. And then he happened to be in the right place, CERN, part of an existing global network that just required some hypertextualizing.

Philipp continues: »And so that might make him believe you can pull these things off with a good, precise spec and some thinking; the Semantic Web!«

Actually, the Semantic Web is not solely TBL's private bailwick, and methinks represents a natural progression of the Web @large (even if I don't believe it'll happen).

Indeed, had Tim &Co. been able to foresee the veritable explosion of the WWW in so short a time, I am sure they'd have added key elements of the Semantic into the Web 0.1 from the start. And set up a condition, say, of ALL non-valid HTML rendering automagically as mere plaintext, not richtext, to stimulate its proper use.

Other than that I personally decry the lack of any integral revenue-or-credit-sharing mechanisms for content providers (to date, the only ones to get paid for the Web are the distributors/ ISPs, while the authors/ content publishers are left out in the cold).

Well, hindsight is cheap. If you had a chance to redesign/ change the specs of the _original_ web, what would they be?

I will not argue that TBL was the original web's inventor and I think the world would be a better place if we had more geniuses like him :). The point I tried to make was differently. Perhaps a little fiction....

Here's a kid called Jimmy who invents 5 lego bricks and those bricks are terrific, and within a week, the whole kindergarten plays with only those lego bricks and builds great lego houses, roads, statues and so on. Now, Jimmy is convinced that his original thinking that led him to the first 5 lego bricks was just the beginning, and true enough, he notices that some of the advanced structures could use an extra lego piece here and there. For instance, someone built an antenna on one of his houses that does look like an antenna but it's not thin enough, so Jimmy starts going back to the drawing board, and he designs all kinds of special lego bricks emulating antennas, road blocks, trees, and so on. Next morning, he presents the 35 new bricks to the kids but he notices they simply prefer to build their own stuff based on those very simple base bricks, and that they reject most of his new designs. The kids tell Jimmy that he might have hit the kindergarten nerve with the 5 lego bricks, of which they actually only use 3 though – another kid in kindergarten, Susie, created 2 more bricks which are widely popular now as well – but that they're kind of too lazy to agree on all his special bricks and to read through those long brick explanation scribbles by Jimmy, and that they by and large don't really agree with him when he goes around telling people, "hey you can't use these two bricks #1 and #3 together in that way", no offense taken as they really love some of his original bricks. The kid Francesco even argues Susie would have come up with other bricks later on sooner or later too. Jimmy, rightfully insulted because Susie's brick inventions are by far not as well thought out and in many ways just copies of his, builds a large brick club and hits Francesco over the head, causing lots of tears around the kids and the kindergarten teacher to be very, very angry with Billy (admittedly, Billy had nothing to do with this beating, but teachers don't always pick the right kid).

Philipp, you'll have to do better than that.... I got lost somewhere around brick #12. Or was it #16?
Besides, trying to elaborate a moot point by way of fairy tales is not a good tactic. We all know the Web is not perfect, and, if the law of unidirectional growth of entropy applies also to the philosophic[k]al concept of the docuverse, then we're in for more of it, not less of it.

Let's drop the issue of TBL's sole or shared contributions, and concentrate instead on what I called this reborn thread: FORKS NOT TAKEN: how would you have gone about it differently?

PS. just remembered... prior to emergence of the WWW (which arrived first in the form of the line-mode TTY client installed at, and reachable via telnet "info.cern.ch:23" – hardly a world-shattering experience), there were other attempts to provide what is the essence of the WWW – one-click access to formatted, directly onscreen-rendered dataflow (we tend to forget, that TBL's greatest invention was probably the single-unbroken-word URL, or URI "scheme://server[:port]/path[/file[.suffix]]"). I recall distinctly several flavours of modified finger daemons – which allowed host admins to provide more than the usual "finger named-user" service. I believe I published 2 texts --in Australia, of all places-- which were accessible (preformated for display on an 80-charwidth TTY teminal) via such command-argument combo as (sample):

~/$ finger "biketales=danny"servername.edu.au

equivalent to later single-word url

http://servername.edu.au/danny/biketales